From friess@adacore.com  Thu Aug 31 10:46:13 2006
From: =?iso-8859-1?Q?Micha=EBl_Friess?= <friess@adacore.com>
To: <report@adacore.com>
Date: Thu, 31 Aug 2006 16:45:50 +0200
Message-ID: <00a701c6cd0c$276f8740$ca01a8c0@munich>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: [F831-014] - #999 including GNAT Pro license and customer information in GNAT Pro
Status: O
Content-Length: 1133
Lines: 28

A result the discussion that happened during the summer meeting breakout
session "Timed Evals" is a proposal for a general mechanism that would check
in the compiler for the license issued to customers or evaluators.

The license file included in standard GNAT Pro distributions would include
the GPL/GMGPL license + the name/ACT# of the customer GNAT Pro is licensed
to.

The license file include in European GNAT Pro evaluations (such a scheme is
not needed in the US) would include a GPL license + the name/ACT# of the
prospect + an optional field indicating the expiration date of the
evaluation. After that date the compiler would stop working.

Note that this will not impact current GNAT Pro customers.

As per a discussion with Robert, Franco, Thomas and I, this TN is created
for implementation of this scheme, and should of course first include a
design for this implementation. Thomas is in charge of the design &
implementation work. He will synchronize with the QA team to make sure the
final scheme is compatible with the QA requirements.

The summary of the breakout session will be filled under this TN.

Michal





From report-internal@adacore.com  Thu Aug 31 10:46:16 2006
To: friess@adacore.com
From: report@adacore.com
Cc: <report@adacore.com>
Message-Id: <20060831144614.DB7CF48CDA8@nile.gnat.com>
Date: Thu, 31 Aug 2006 10:46:14 -0400 (EDT)
Subject: Re: [F831-014] #999 including GNAT Pro license and customer information in GNAT Pro
Status: O
Content-Length: 350
Lines: 15

[Automatic message from GNAT report ticket mechanism]

Dear Customer,

Your report has been automatically assigned ticket number [F831-014].
Please be sure to put this number in the subject line of any
subsequent followup message. Messages from us on this report will
quote this ticket number.

Thank you for using GNAT.

-- The GNAT support team




From friess@adacore.com  Thu Aug 31 10:47:39 2006
From: =?iso-8859-1?Q?Micha=EBl_Friess?= <friess@adacore.com>
To: "'Automatic Filer'" <file@gnat.com>
Date: Thu, 31 Aug 2006 16:47:17 +0200
Message-ID: <00ab01c6cd0c$5ace9ef0$ca01a8c0@munich>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: [F831-014] Summary of the "timed GNAT Pro eval" breakout session
Status: O
Content-Length: 1745
Lines: 41

The reason for this breakout session is the strong input from Europe who
would like a better control on GNAT Pro evaluations. We would like to setup
a mechanism to help with this in order to apply pressure to people so that
they do not use GNAT Pro evals longer than they are supposed to.

Note that the name of GNAT Pro for evaluations should be "GNAT Pro Eval"
instead of "GNAT Pro".

This requirement is for Europe only. The US do not see such a need so the
timing mechanism should be optional.

We can think of 2 mechanisms:

- a mechanism only based on a GPL license issued to evaluators of GNAT Pro

- a mechanism that adds a timing constraint on the license, i.e. that issues
a warning after a certain deadline or that prevents the compiler from
working after this deadline.

We have then moved beyond these mechanisms and thought that it would be good
if everyone could have a license (GNAT Pro & non GNAT Pro users). The idea
is that all users should have a license included in their installation and
we could check for this license to be present.

A more general mechanism can then be put in place. In addition to the
license file, we can provide the name of the company this license is given
to, and we could (optional feature) include an expiration date that could be
used in the GNAT Pro eval context. After the expiration date, the compiler
stops working. Gnatls would display all this information (license, company
name).

This license file will be midly checksumed. It may provide additional legal
protection as md5 is considered cryptographic secured. As this license file
would contain customer specific information it should obviously be
regenerated for each customer; the CRM/GNAT Tracker may help in this task.

Michal





From friess@gnat.com  Thu Aug 31 10:48:21 2006
From: friess@gnat.com
To: file@gnat.com
Message-Id: <20060831144820.EA35C48CDA8@nile.gnat.com>
Date: Thu, 31 Aug 2006 10:48:20 -0400 (EDT)
Subject: [bugtool:F831-014] New TN status: 5 quinot "including GNAT Pro license & customer info in GNAT Pro"
Status: O
Content-Length: 4
Lines: 4






From quinot@gnat.com  Thu Aug 31 12:26:11 2006
From: quinot@gnat.com
To: file@gnat.com
Message-Id: <20060831162610.F1AE348CE02@nile.gnat.com>
Date: Thu, 31 Aug 2006 12:26:10 -0400 (EDT)
Subject: [bugtool:F831-014] Description: "including GNAT Pro license & customer info in GNAT"
Status: O
Content-Length: 4
Lines: 4






From quinot@adacore.com  Fri Sep  1 12:43:26 2006
Date: Fri, 1 Sep 2006 18:43:22 +0200
From: Thomas Quinot <quinot@adacore.com>
To: =?iso-8859-1?Q?Micha=EBl?= Friess <friess@adacore.com>
Cc: report@adacore.com
Message-ID: <20060901164322.GA92617@melamine.cuivre.fr.eu.org>
References: <00a701c6cd0c$276f8740$ca01a8c0@munich>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
In-Reply-To: <00a701c6cd0c$276f8740$ca01a8c0@munich>
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information in GNAT Pro
Status: O
Content-Length: 4409
Lines: 138

* Michal Friess, 2006-08-31 :

> As per a discussion with Robert, Franco, Thomas and I, this TN is created
> for implementation of this scheme, and should of course first include a
> design for this implementation. Thomas is in charge of the design &
> implementation work. He will synchronize with the QA team to make sure the
> final scheme is compatible with the QA requirements.

Here is the design.

1. Format of license information file
-------------------------------------

The license information file is a human-readable text file comprising
two sections.

The first section is the header. The format of the header is as follows:

---------- BEGIN HEADER ----------
== GNAT License information ==

Licensee: <customer-id> <customer-name>

Product: <product-id>
Version: <version-string>
Host: <host-platform>
Target: <target-platform>

Validity: <validity-date>

Signature: <hex-digits>
---------- END HEADER ----------

The variable fields are as follows:

<customer-id>: optional string, '#' followed by account number
<customer-name>: name of licensee. May not start with '#'

For GPL versions, there is no customer-id, and the customer name is
'Free Software Developer'.

<product-id>: a string associated with a GNAT Build type:
  GNAT Pro
or
  GNAT GPL
(does not apply for FSF builds since these will not use the license
validation scheme) for the licensed product.

<version-string>
Licensed version, in the format of Gnatvsn.Static_Version_String.

<host-platform>, <target-platform>
The GCC names for the host and target for which the license is valid.
These two lines are optional. They shall be ommitted in the case of GPL
releases.

<validity-date>
Limit date for license validity. This line is optional.

<signature>
A string of hex digits, as specified below.

The second section is the text of the license, preceded by
"The technology specified above is licensed to the indicated licensee
under the following terms:"

2. The signature
----------------

The signature is a field that allows verifying whether the license file
has been altered. It is created as follows:

1. Prepare license file with no signature (i.e. with the last character
   of the Signature: line being the colon).
2. Compute a DSA signature of the file as specified in RFC 3174
3. The signature is the hexadecimal representation of the computed DSA
   signature.

It is verified as follows:

1. Search license file for Signature: line. If not found, verification
   fails.
2. Read sequence of hexadecimal digits from signature line.
3. Create a copy of license file with everything after the colon
   removed from the signature: line
4. Verify DSA signature as specified in RFC 3174

3. Key management
-----------------

The creation of DSA signatures requires DSA parameters, and a DSA
private key. The verification of DSA signatures requires the same DSA
parameters, and a DSA public key.

One set of parameters and one private/public key pair are created at
AdaCore.

The value of the numbers that constitute the parameters and the public
key are stored as named numbers in system.ads (so that they are
conveniently accessable as uint values through targparm and cannot be
altered without recompiling the whole runtime library).

4. License enforcement
----------------------

At startup, each licensed tool:

  1. looks for the license file in
     $prefix/gcc/TARGET/VERSION/gnat_license.txt
     if the file is missing, verification fails
     
  2. loads the license file

  3. checks the signature on the license file

  4. checks the product-id against Gnatvsn.Get_Build_Type

  5. checks the version-string against Gnatvsn.Static_Version_String

  6. if a host platform name is present in the license file, check it
     against the compiler's host platform name (if not skip to step 7)

  7. if a target platform name is present in the license file, check it
     against the compiler's target platform name (if not skip to step 8)

  8. if a validity date is present in the license file, check that the
     current date is not past the validity date

If any of these steps fails, an error message is emitted:
"fatal error: installation checks failed" followed by a numeric error
code indicating which of the steps above failed, and Unrecoverable_Error
is raised.

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From quinot@adacore.com  Fri Sep  1 17:08:17 2006
Date: Fri, 1 Sep 2006 23:08:15 +0200
From: Thomas Quinot <quinot@adacore.com>
To: =?iso-8859-1?Q?Micha=EBl?= Friess <friess@adacore.com>
Cc: report@adacore.com
Message-ID: <20060901210815.GA7135@melamine.cuivre.fr.eu.org>
References: <00a701c6cd0c$276f8740$ca01a8c0@munich> <20060901164322.GA92617@melamine.cuivre.fr.eu.org>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <20060901164322.GA92617@melamine.cuivre.fr.eu.org>
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information in GNAT Pro
Status: O
Content-Length: 988
Lines: 27

* Thomas Quinot, 2006-09-01 :

> Here is the design.

Small revision...

> The value of the numbers that constitute the parameters and the public
> key are stored as named numbers in system.ads (so that they are
> conveniently accessable as uint values through targparm and cannot be
> altered without recompiling the whole runtime library).

After more reflection, it appears that this may not be the best option
(necessity to maintain several copies of the keys, more maintainance
work for FSF repo). Maybe it's better just to hardcode the keys in the
unit that will do license validation. This will at least force anyone
who wants to change the key to rebuild the compiler, and that's good
enough protection for our purposes (since anyone who wants to get
through the hassle of rebuilding the compiler can just remove the check
altogether).

Thomas.

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From berrendo@adacore.com  Mon Sep  4 03:45:03 2006
Message-ID: <44FBD976.2090707@adacore.com>
Date: Mon, 04 Sep 2006 09:44:54 +0200
From: Romain Berrendonner <berrendo@adacore.com>
MIME-Version: 1.0
To: Thomas Quinot <quinot@adacore.com>
CC: =?ISO-8859-1?Q?Micha=EBl_Friess?= <friess@adacore.com>, 
 report@adacore.com
References: <00a701c6cd0c$276f8740$ca01a8c0@munich> <20060901164322.GA92617@melamine.cuivre.fr.eu.org>
In-Reply-To: <20060901164322.GA92617@melamine.cuivre.fr.eu.org>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information  in GNAT Pro
Status: O
Content-Length: 512
Lines: 17

Thomas Quinot a crit :
> The variable fields are as follows:
> 
> <customer-id>: optional string, '#' followed by account number
> <customer-name>: name of licensee. May not start with '#'
> 
> For GPL versions, there is no customer-id, and the customer name is
> 'Free Software Developer'.

What are the plans for the Pro version ?

-- 
Romain Berrendonner                           berrendonner@adacore.com
+33 (0)1 49 70 67 16                            http://www.adacore.com
8 rue de Milan F-75009 Paris



From quinot@adacore.com  Mon Sep  4 04:38:23 2006
Date: Mon, 4 Sep 2006 10:37:56 +0200
From: Thomas Quinot <quinot@adacore.com>
To: Romain Berrendonner <berrendo@adacore.com>
Cc: =?iso-8859-1?Q?Micha=EBl?= Friess <friess@adacore.com>,
	report@adacore.com
Message-ID: <20060904083756.GA23836@melamine.cuivre.fr.eu.org>
References: <00a701c6cd0c$276f8740$ca01a8c0@munich> <20060901164322.GA92617@melamine.cuivre.fr.eu.org> <44FBD976.2090707@adacore.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <44FBD976.2090707@adacore.com>
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information in GNAT Pro
Status: O
Content-Length: 525
Lines: 17

* Romain Berrendonner, 2006-09-04 :

> What are the plans for the Pro version ?

The proposed design allows implementation of many policies. All GNAT Pro
customers could share a common certificate licensed to 'GNAT Pro
Customer' without a customer ID, or each customer could download from GT
a distribution containing a branded certificate for their specific
customer information.

Thomas.

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From berrendo@adacore.com  Mon Sep  4 04:50:36 2006
Message-ID: <44FBE8D3.5040407@adacore.com>
Date: Mon, 04 Sep 2006 10:50:27 +0200
From: Romain Berrendonner <berrendo@adacore.com>
MIME-Version: 1.0
To: Thomas Quinot <quinot@adacore.com>
CC: =?ISO-8859-1?Q?Micha=EBl_Friess?= <friess@adacore.com>, 
 report@adacore.com
References: <00a701c6cd0c$276f8740$ca01a8c0@munich> <20060901164322.GA92617@melamine.cuivre.fr.eu.org> <44FBD976.2090707@adacore.com> <20060904083756.GA23836@melamine.cuivre.fr.eu.org>
In-Reply-To: <20060904083756.GA23836@melamine.cuivre.fr.eu.org>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information  in GNAT Pro
Status: O
Content-Length: 778
Lines: 19

Thomas Quinot a crit :
> The proposed design allows implementation of many policies. All GNAT Pro
> customers could share a common certificate licensed to 'GNAT Pro
> Customer' without a customer ID, or each customer could download from GT
> a distribution containing a branded certificate for their specific
> customer information.

This is precisely want I want to discuss ;-)

I strongly oppose personalized certificates for Professional customers. 
I think this would be 1) an overkill 2) weaken our "no-lock" message 3) 
be a nightmare to manage. I am fine with the generic certificate though.

-- 
Romain Berrendonner                           berrendonner@adacore.com
+33 (0)1 49 70 67 16                            http://www.adacore.com
8 rue de Milan F-75009 Paris



From charlet@ACT-Europe.FR  Mon Sep  4 04:55:29 2006
Date: Mon, 4 Sep 2006 10:55:28 +0200
From: Arnaud Charlet <charlet@adacore.com>
To: Romain Berrendonner <berrendo@adacore.com>
Cc: Thomas Quinot <quinot@adacore.com>,
	=?iso-8859-1?Q?Micha=EBl?= Friess <friess@adacore.com>,
	report@adacore.com
Message-ID: <20060904085528.GA45759@adacore.com>
References: <00a701c6cd0c$276f8740$ca01a8c0@munich> <20060901164322.GA92617@melamine.cuivre.fr.eu.org> <44FBD976.2090707@adacore.com> <20060904083756.GA23836@melamine.cuivre.fr.eu.org> <44FBE8D3.5040407@adacore.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <44FBE8D3.5040407@adacore.com>
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information in GNAT Pro
Status: O
Content-Length: 566
Lines: 16

> I strongly oppose personalized certificates for Professional customers. 
> I think this would be 1) an overkill 2) weaken our "no-lock" message 3) 
> be a nightmare to manage. I am fine with the generic certificate though.

Strongly agreed.

Part of the technical requirements for this issue are:

- do not generate more binaries than already done today (so one binary
  for pro, and one for gpl for a given platform)
- the generation of "license" files should be handled by the sales team,
  with no impact on the technical team or on report/GT suppport.

Arno



From charlet@ACT-Europe.FR  Mon Sep  4 04:57:36 2006
Date: Mon, 4 Sep 2006 10:57:35 +0200
From: Arnaud Charlet <charlet@adacore.com>
To: Romain Berrendonner <berrendo@adacore.com>
Cc: Thomas Quinot <quinot@adacore.com>,
	=?iso-8859-1?Q?Micha=EBl?= Friess <friess@adacore.com>,
	report@adacore.com
Message-ID: <20060904085735.GA51572@adacore.com>
References: <00a701c6cd0c$276f8740$ca01a8c0@munich> <20060901164322.GA92617@melamine.cuivre.fr.eu.org> <44FBD976.2090707@adacore.com> <20060904083756.GA23836@melamine.cuivre.fr.eu.org> <44FBE8D3.5040407@adacore.com> <20060904085528.GA45759@adacore.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <20060904085528.GA45759@adacore.com>
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information in GNAT Pro
Status: O
Content-Length: 479
Lines: 14

> - do not generate more binaries than already done today (so one binary
>   for pro, and one for gpl for a given platform)
> - the generation of "license" files should be handled by the sales team,
>   with no impact on the technical team or on report/GT suppport.

And I forgot:

- the check will not be included at the FSF, so the implementation should
  be clearly separated to avoid merge conflicts at the FSF, and be easily
  disabled (e.g. remove/modify one line)

Arno



From fofanov@adacore.com  Mon Sep  4 05:04:47 2006
Date: Mon, 4 Sep 2006 11:04:46 +0200
From: Vasiliy Fofanov <fofanov@adacore.com>
Message-ID: <1893010269.20060904110446@adacore.com>
To: Romain Berrendonner <berrendo@adacore.com>
CC: Thomas Quinot <quinot@adacore.com>, Michal Friess <friess@adacore.com>, report@adacore.com
In-Reply-To: <44FBE8D3.5040407@adacore.com>
References: <00a701c6cd0c$276f8740$ca01a8c0@munich>
 <20060901164322.GA92617@melamine.cuivre.fr.eu.org>
 <44FBD976.2090707@adacore.com>
 <20060904083756.GA23836@melamine.cuivre.fr.eu.org>
 <44FBE8D3.5040407@adacore.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information in GNAT Pro
Status: O
Content-Length: 344
Lines: 11

I would agree that generic certificate is the proper way. Per-customer
licenses are untenable in our model of support IMO. The only problem is
that it will take one Pro user to leak, say, 5.04a1 license file to allow
any eval to remove the restrictions. But I think this is the risk we'll
have to live with.

Best regards,
 Vasiliy Fofanov.




From friess@adacore.com  Mon Sep  4 05:06:02 2006
From: =?iso-8859-1?Q?Micha=EBl_Friess?= <friess@adacore.com>
To: "'Thomas Quinot'" <quinot@adacore.com>
Cc: <report@adacore.com>
Date: Mon, 4 Sep 2006 11:05:52 +0200
Message-ID: <000401c6d001$52883240$89000a0a@munich>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
In-Reply-To: <20060901164322.GA92617@melamine.cuivre.fr.eu.org>
Subject: RE: [F831-014] - #999 including GNAT Pro license and customer information in GNAT Pro
Status: O
Content-Length: 5156
Lines: 157

The proposal corresponds to what has been discussed, seems good to me.

Note that GNAT Pro Eval is not mentioned in the proposal. So product id
could also contain this value: "GNAT Pro Eval".

Michal

> -----Message d'origine-----
> De: Thomas Quinot [mailto:quinot@adacore.com]
> Envoy: Friday, September 01, 2006 6:43 PM
> : Michal Friess
> Cc: report@adacore.com
> Objet: Re: [F831-014] - #999 including GNAT Pro license and customer
> information in GNAT Pro
> 
> * Michal Friess, 2006-08-31 :
> 
> > As per a discussion with Robert, Franco, Thomas and I, this TN is
> created
> > for implementation of this scheme, and should of course first include a
> > design for this implementation. Thomas is in charge of the design &
> > implementation work. He will synchronize with the QA team to make sure
> the
> > final scheme is compatible with the QA requirements.
> 
> Here is the design.
> 
> 1. Format of license information file
> -------------------------------------
> 
> The license information file is a human-readable text file comprising
> two sections.
> 
> The first section is the header. The format of the header is as follows:
> 
> ---------- BEGIN HEADER ----------
> == GNAT License information ==
> 
> Licensee: <customer-id> <customer-name>
> 
> Product: <product-id>
> Version: <version-string>
> Host: <host-platform>
> Target: <target-platform>
> 
> Validity: <validity-date>
> 
> Signature: <hex-digits>
> ---------- END HEADER ----------
> 
> The variable fields are as follows:
> 
> <customer-id>: optional string, '#' followed by account number
> <customer-name>: name of licensee. May not start with '#'
> 
> For GPL versions, there is no customer-id, and the customer name is
> 'Free Software Developer'.
> 
> <product-id>: a string associated with a GNAT Build type:
>   GNAT Pro
> or
>   GNAT GPL
> (does not apply for FSF builds since these will not use the license
> validation scheme) for the licensed product.
> 
> <version-string>
> Licensed version, in the format of Gnatvsn.Static_Version_String.
> 
> <host-platform>, <target-platform>
> The GCC names for the host and target for which the license is valid.
> These two lines are optional. They shall be ommitted in the case of GPL
> releases.
> 
> <validity-date>
> Limit date for license validity. This line is optional.
> 
> <signature>
> A string of hex digits, as specified below.
> 
> The second section is the text of the license, preceded by
> "The technology specified above is licensed to the indicated licensee
> under the following terms:"
> 
> 2. The signature
> ----------------
> 
> The signature is a field that allows verifying whether the license file
> has been altered. It is created as follows:
> 
> 1. Prepare license file with no signature (i.e. with the last character
>    of the Signature: line being the colon).
> 2. Compute a DSA signature of the file as specified in RFC 3174
> 3. The signature is the hexadecimal representation of the computed DSA
>    signature.
> 
> It is verified as follows:
> 
> 1. Search license file for Signature: line. If not found, verification
>    fails.
> 2. Read sequence of hexadecimal digits from signature line.
> 3. Create a copy of license file with everything after the colon
>    removed from the signature: line
> 4. Verify DSA signature as specified in RFC 3174
> 
> 3. Key management
> -----------------
> 
> The creation of DSA signatures requires DSA parameters, and a DSA
> private key. The verification of DSA signatures requires the same DSA
> parameters, and a DSA public key.
> 
> One set of parameters and one private/public key pair are created at
> AdaCore.
> 
> The value of the numbers that constitute the parameters and the public
> key are stored as named numbers in system.ads (so that they are
> conveniently accessable as uint values through targparm and cannot be
> altered without recompiling the whole runtime library).
> 
> 4. License enforcement
> ----------------------
> 
> At startup, each licensed tool:
> 
>   1. looks for the license file in
>      $prefix/gcc/TARGET/VERSION/gnat_license.txt
>      if the file is missing, verification fails
> 
>   2. loads the license file
> 
>   3. checks the signature on the license file
> 
>   4. checks the product-id against Gnatvsn.Get_Build_Type
> 
>   5. checks the version-string against Gnatvsn.Static_Version_String
> 
>   6. if a host platform name is present in the license file, check it
>      against the compiler's host platform name (if not skip to step 7)
> 
>   7. if a target platform name is present in the license file, check it
>      against the compiler's target platform name (if not skip to step 8)
> 
>   8. if a validity date is present in the license file, check that the
>      current date is not past the validity date
> 
> If any of these steps fails, an error message is emitted:
> "fatal error: installation checks failed" followed by a numeric error
> code indicating which of the steps above failed, and Unrecoverable_Error
> is raised.
> 
> --
> Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
>                AdaCore -- Paris, France -- New York, USA





From friess@adacore.com  Mon Sep  4 05:07:27 2006
From: =?iso-8859-1?Q?Micha=EBl_Friess?= <friess@adacore.com>
To: "'Arnaud Charlet'" <charlet@adacore.com>,
	"'Romain Berrendonner'" <berrendo@adacore.com>
Cc: "'Thomas Quinot'" <quinot@adacore.com>,
	<report@adacore.com>
Date: Mon, 4 Sep 2006 11:07:17 +0200
Message-ID: <000501c6d001$85157b00$89000a0a@munich>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
In-Reply-To: <20060904085528.GA45759@adacore.com>
Subject: RE: [F831-014] - #999 including GNAT Pro license and customer information in GNAT Pro
Status: O
Content-Length: 844
Lines: 26

> -----Message d'origine-----
> De: Arnaud Charlet [mailto:charlet@adacore.com]
> Envoy: Monday, September 04, 2006 10:55 AM
> : Romain Berrendonner
> Cc: Thomas Quinot; Michal Friess; report@adacore.com
> Objet: Re: [F831-014] - #999 including GNAT Pro license and customer
> information in GNAT Pro
> 
> Part of the technical requirements for this issue are:
> 
> - do not generate more binaries than already done today (so one binary
>   for pro, and one for gpl for a given platform)

AFAIK the scheme does not require additional binaries.

> - the generation of "license" files should be handled by the sales team,
>   with no impact on the technical team or on report/GT suppport.

Agreed. Another way of seeing this is to handle the generation of the
license file automatically, for example by GNAT Tracker/the CRM.

Michal





From friess@adacore.com  Mon Sep  4 05:09:40 2006
From: =?iso-8859-1?Q?Micha=EBl_Friess?= <friess@adacore.com>
To: "'Arnaud Charlet'" <charlet@adacore.com>,
	"'Romain Berrendonner'" <berrendo@adacore.com>
Cc: "'Thomas Quinot'" <quinot@adacore.com>,
	<report@adacore.com>
Date: Mon, 4 Sep 2006 11:09:30 +0200
Message-ID: <000601c6d001$d4e85b20$89000a0a@munich>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
In-Reply-To: <20060904085735.GA51572@adacore.com>
Subject: RE: [F831-014] - #999 including GNAT Pro license and customer information in GNAT Pro
Status: O
Content-Length: 725
Lines: 23

> -----Message d'origine-----
> De: Arnaud Charlet [mailto:charlet@adacore.com]
> Envoy: Monday, September 04, 2006 10:58 AM
> : Romain Berrendonner
> Cc: Thomas Quinot; Michal Friess; report@adacore.com
> Objet: Re: [F831-014] - #999 including GNAT Pro license and customer
> information in GNAT Pro
> 
> And I forgot:
> 
> - the check will not be included at the FSF, so the implementation should
>   be clearly separated to avoid merge conflicts at the FSF, and be easily
>   disabled (e.g. remove/modify one line)

Right. This had indeed been discussed with Thomas and we agreed that the
scheme should not be visible at the FSF. So the fact that it was not
mentioned in the proposal is an oversight.

Michal





From friess@adacore.com  Mon Sep  4 05:15:09 2006
From: =?iso-8859-1?Q?Micha=EBl_Friess?= <friess@adacore.com>
To: "'Vasiliy Fofanov'" <fofanov@adacore.com>
Cc: <report@adacore.com>,
	"'Sales'" <sales@eu.adacore.com>
Date: Mon, 4 Sep 2006 11:15:00 +0200
Message-ID: <000701c6d002$9990e640$89000a0a@munich>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
In-Reply-To: <1893010269.20060904110446@adacore.com>
Subject: RE: [F831-014] - #999 including GNAT Pro license and customer information in GNAT Pro
Status: O
Content-Length: 876
Lines: 24

> -----Message d'origine-----
> De: Vasiliy Fofanov [mailto:fofanov@adacore.com]
> Envoy: Monday, September 04, 2006 11:05 AM
> : Romain Berrendonner
> Cc: Thomas Quinot; Michal Friess; report@adacore.com
> Objet: Re: [F831-014] - #999 including GNAT Pro license and customer
> information in GNAT Pro
> 
> Per-customer licenses are untenable in our model of support IMO.

Not sure to understand why you say so. This is exactly what we give to our
customers: a proper GNAT Pro license. We are just making it more explicit.

Note that this is also a useful way to understand where the GNAT Pro that
people use come from: at the sales level we have several examples of people
who are not customers and who are using GNAT Pro (Thales, DCN) and it's
always difficult to find out where they got their GNAT Pro version from.
This scheme would simplify our job.

Michal





From comar@adacore.com  Mon Sep  4 05:18:33 2006
Message-ID: <44FBEF66.4060908@adacore.com>
Date: Mon, 04 Sep 2006 11:18:30 +0200
From: Cyrille Comar <comar@adacore.com>
MIME-Version: 1.0
To: Thomas Quinot <quinot@adacore.com>
CC: =?ISO-8859-1?Q?Micha=EBl_Friess?= <friess@adacore.com>, 
 report@adacore.com
References: <00a701c6cd0c$276f8740$ca01a8c0@munich> <20060901164322.GA92617@melamine.cuivre.fr.eu.org> <20060901210815.GA7135@melamine.cuivre.fr.eu.org>
In-Reply-To: <20060901210815.GA7135@melamine.cuivre.fr.eu.org>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information  in GNAT Pro
Status: O
Content-Length: 612
Lines: 15

Thomas Quinot wrote:
> After more reflection, it appears that this may not be the best option
> (necessity to maintain several copies of the keys, more maintainance
> work for FSF repo). Maybe it's better just to hardcode the keys in the
> unit that will do license validation. This will at least force anyone
> who wants to change the key to rebuild the compiler, and that's good
> enough protection for our purposes (since anyone who wants to get
> through the hassle of rebuilding the compiler can just remove the check
> altogether).

but doesn't that require re-building a compiler for each evaluation?





From berrendo@adacore.com  Mon Sep  4 05:38:01 2006
Message-ID: <44FBF3F0.2050902@adacore.com>
Date: Mon, 04 Sep 2006 11:37:52 +0200
From: Romain Berrendonner <berrendo@adacore.com>
MIME-Version: 1.0
To: Cyrille Comar <comar@adacore.com>
CC: Thomas Quinot <quinot@adacore.com>, =?ISO-8859-1?Q?Micha=EBl_Friess?=
 <friess@adacore.com>,  report@adacore.com
References: <00a701c6cd0c$276f8740$ca01a8c0@munich> <20060901164322.GA92617@melamine.cuivre.fr.eu.org> <20060901210815.GA7135@melamine.cuivre.fr.eu.org> <44FBEF66.4060908@adacore.com>
In-Reply-To: <44FBEF66.4060908@adacore.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information  in GNAT Pro
Status: O
Content-Length: 584
Lines: 18

Cyrille Comar a crit :
> 
> but doesn't that require re-building a compiler for each evaluation?
> 

There are two solutions here:
- We have a script that repackages the compiler with the certificate;
- We leave the compiler untouched and add an additional level of 
wrapping which would add the certificate on the fly at install.

I prefer option 1 as it gives less opportunity to cheat with the system.

-- 
Romain Berrendonner                           berrendonner@adacore.com
+33 (0)1 49 70 67 16                            http://www.adacore.com
8 rue de Milan F-75009 Paris



From fofanov@adacore.com  Mon Sep  4 05:38:52 2006
Date: Mon, 4 Sep 2006 11:38:50 +0200
From: Vasiliy Fofanov <fofanov@adacore.com>
Message-ID: <424168041.20060904113850@adacore.com>
To: Michal Friess <friess@adacore.com>
CC: report@adacore.com, "'Sales'" <sales@eu.adacore.com>
In-Reply-To: <000701c6d002$9990e640$89000a0a@munich>
References: <1893010269.20060904110446@adacore.com>
 <000701c6d002$9990e640$89000a0a@munich>
MIME-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information in GNAT Pro
Status: O
Content-Length: 509
Lines: 17

> Not sure to understand why you say so. This is exactly what we give to our
> customers: a proper GNAT Pro license.

"License" in my case refers to a license key. "License" in your case refers
to a contract. One doesn't have anything to do with the other.

> We are just making it more explicit.

Do you suggest that we tailor-build a wavefront for each customer that
requires it so that it has a proper license key? That's what I call
untenable, surely you can see this?

Best regards,
 Vasiliy Fofanov.




From charlet@ACT-Europe.FR  Mon Sep  4 05:41:09 2006
Date: Mon, 4 Sep 2006 11:41:08 +0200
From: Arnaud Charlet <charlet@adacore.com>
To: Romain Berrendonner <berrendo@adacore.com>
Cc: Cyrille Comar <comar@adacore.com>, Thomas Quinot <quinot@adacore.com>,
	=?iso-8859-1?Q?Micha=EBl?= Friess <friess@adacore.com>,
	report@adacore.com
Message-ID: <20060904094108.GA7890@adacore.com>
References: <00a701c6cd0c$276f8740$ca01a8c0@munich> <20060901164322.GA92617@melamine.cuivre.fr.eu.org> <20060901210815.GA7135@melamine.cuivre.fr.eu.org> <44FBEF66.4060908@adacore.com> <44FBF3F0.2050902@adacore.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <44FBF3F0.2050902@adacore.com>
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information in GNAT Pro
Status: O
Content-Length: 754
Lines: 20

> There are two solutions here:
> - We have a script that repackages the compiler with the certificate;
> - We leave the compiler untouched and add an additional level of 
> wrapping which would add the certificate on the fly at install.
> 
> I prefer option 1 as it gives less opportunity to cheat with the system.
> 

How are evaluations handled currently ? Do people have access to the
download area of gnattracker ? (in which case we would need to change that),
or only to a "wavefront area" ? (in which case wavefront is no really
appropriate) or something else ?

Again, I assume no extra work for evaluations would be required on the
technical side, and that everything would be handled either automatically,
or by the sales team, right ?

Arno



From friess@adacore.com  Mon Sep  4 05:44:19 2006
From: =?iso-8859-1?Q?Micha=EBl_Friess?= <friess@adacore.com>
To: "'Romain Berrendonner'" <berrendo@adacore.com>
Cc: <report@adacore.com>,
	"'Sales'" <sales@eu.adacore.com>
Date: Mon, 4 Sep 2006 11:44:10 +0200
Message-ID: <000801c6d006$ac48a120$89000a0a@munich>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
In-Reply-To: <44FBE8D3.5040407@adacore.com>
Subject: RE: [F831-014] - #999 including GNAT Pro license and customer information in GNAT Pro
Status: O
Content-Length: 1327
Lines: 35

To summarize the discussion we have just had in our office: you would agree
with a certificate shared by all our Pro customers and licensed to 'GNAT Pro
Customer'.

I agree that this is a reasonable solution as it lightens the QA and release
process, and as I don't want the 'Pro' discussion to take over the 'Eval'
discussion I would say that we should go this route for the 'Pro' customers.

Michal

> -----Message d'origine-----
> De: Romain Berrendonner [mailto:berrendo@adacore.com]
> Envoy: Monday, September 04, 2006 10:50 AM
> : Thomas Quinot
> Cc: Michal Friess; report@adacore.com
> Objet: Re: [F831-014] - #999 including GNAT Pro license and customer
> information in GNAT Pro
> 
> Thomas Quinot a crit :
> > The proposed design allows implementation of many policies. All GNAT Pro
> > customers could share a common certificate licensed to 'GNAT Pro
> > Customer' without a customer ID, or each customer could download from GT
> > a distribution containing a branded certificate for their specific
> > customer information.
> 
> This is precisely want I want to discuss ;-)
> 
> I strongly oppose personalized certificates for Professional customers.
> I think this would be 1) an overkill 2) weaken our "no-lock" message 3)
> be a nightmare to manage. I am fine with the generic certificate though.






From friess@adacore.com  Mon Sep  4 05:47:36 2006
From: =?iso-8859-1?Q?Micha=EBl_Friess?= <friess@adacore.com>
To: "'Arnaud Charlet'" <charlet@adacore.com>
Cc: <report@adacore.com>,
	"'Sales'" <sales@eu.adacore.com>
Date: Mon, 4 Sep 2006 11:47:23 +0200
Message-ID: <000901c6d007$1f99a7f0$89000a0a@munich>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
In-Reply-To: <20060904094108.GA7890@adacore.com>
Subject: RE: [F831-014] - #999 including GNAT Pro license and customer information in GNAT Pro
Status: O
Content-Length: 888
Lines: 26

> -----Message d'origine-----
> De: Arnaud Charlet [mailto:charlet@adacore.com]
> Envoy: Monday, September 04, 2006 11:41 AM
> : Romain Berrendonner
> Cc: Cyrille Comar; Thomas Quinot; Michal Friess; report@adacore.com
> Objet: Re: [F831-014] - #999 including GNAT Pro license and customer
> information in GNAT Pro
> 
> How are evaluations handled currently ?

Currently evaluations are handled exactly the same way as standard GNAT Pro
subscriptions (in particular, they have exactly the same GNAT Tracker access
as standard GNAT Pro customers).
 
> Again, I assume no extra work for evaluations would be required on the
> technical side, and that everything would be handled either automatically,
> or by the sales team, right ?

You are right, everything should be handled automatically or by the sales
team. The exact procedure to follow needs to be determined.

Michal





From fofanov@adacore.com  Mon Sep  4 05:52:24 2006
Date: Mon, 4 Sep 2006 11:52:23 +0200
From: Vasiliy Fofanov <fofanov@adacore.com>
Message-ID: <1691129980.20060904115223@adacore.com>
To: Romain Berrendonner <berrendo@adacore.com>
CC: Cyrille Comar <comar@adacore.com>, Thomas Quinot <quinot@adacore.com>, Michal Friess <friess@adacore.com>, 
	<report@adacore.com>
In-Reply-To: <44FBF3F0.2050902@adacore.com>
References: <00a701c6cd0c$276f8740$ca01a8c0@munich>
 <20060901164322.GA92617@melamine.cuivre.fr.eu.org>
 <20060901210815.GA7135@melamine.cuivre.fr.eu.org>
 <44FBEF66.4060908@adacore.com> <44FBF3F0.2050902@adacore.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information in GNAT Pro
Status: O
Content-Length: 748
Lines: 20

> - We have a script that repackages the compiler with the certificate;
> - We leave the compiler untouched and add an additional level of 
> wrapping which would add the certificate on the fly at install.

This can't be done for Windows (both native and cross) and VMS, neither of
these packages allow changing their contents, and can only be generated on
a corresponding host.

I think the only way to make this work is to allow user to download a
license key separately from GNAT Tracker and manually copy it where
compiler expects it. Which is a bit of a hassle.

Note that we want to do it anyway if the client goes from eval to pro, we
don't want to force him to reinstall just to replace the license key.

Best regards,
 Vasiliy Fofanov.




From charlet@ACT-Europe.FR  Mon Sep  4 06:03:47 2006
Date: Mon, 4 Sep 2006 12:03:46 +0200
From: Arnaud Charlet <charlet@adacore.com>
To: =?iso-8859-1?Q?Micha=EBl?= Friess <friess@adacore.com>
Cc: report@adacore.com, 'Sales' <sales@eu.adacore.com>
Message-ID: <20060904100346.GA26769@adacore.com>
References: <20060904094108.GA7890@adacore.com> <000901c6d007$1f99a7f0$89000a0a@munich>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <000901c6d007$1f99a7f0$89000a0a@munich>
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information in GNAT Pro
Status: O
Content-Length: 306
Lines: 10

> Currently evaluations are handled exactly the same way as standard GNAT Pro
> subscriptions (in particular, they have exactly the same GNAT Tracker access
> as standard GNAT Pro customers).

So you are proposing to change that, right ?
Could you (or someone else) clarify what would be changed ?

Arno



From kenner@vlsi1.ultra.nyu.edu  Mon Sep  4 06:24:53 2006
From: kenner@vlsi1.ultra.nyu.edu (Richard Kenner)
Message-Id: <10609041027.AA23751@vlsi1.ultra.nyu.edu>
Date: Mon, 04 Sep 2006 06:27:44 EDT
To: fofanov@adacore.com
Cc: berrendo@adacore.com, comar@adacore.com, friess@adacore.com,
        quinot@adacore.com, report@adacore.com
In-Reply-To: <1691129980.20060904115223@adacore.com>
References: <00a701c6cd0c$276f8740$ca01a8c0@munich>
    <20060901164322.GA92617@melamine.cuivre.fr.eu.org>
    <20060901210815.GA7135@melamine.cuivre.fr.eu.org>
    <44FBEF66.4060908@adacore.com> <44FBF3F0.2050902@adacore.com>
    <1691129980.20060904115223@adacore.com>
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information in GNAT Pro
Status: O
Content-Length: 463
Lines: 10

> This can't be done for Windows (both native and cross) and VMS, neither o f
> these packages allow changing their contents, and can only be generated on
> a corresponding host.

Dunno about VMS, but it certainly can be done on Windows.  You take
the existing InstallShield file and put it and the license into a zip
file.  This zip can be built anywhere and the customer unpacks it and
runs the InstallShield file that looks for the license and copies it in.



From fofanov@adacore.com  Mon Sep  4 06:31:52 2006
Date: Mon, 4 Sep 2006 12:31:51 +0200
From: Vasiliy Fofanov <fofanov@adacore.com>
Message-ID: <9110287480.20060904123151@adacore.com>
To: kenner@vlsi1.ultra.nyu.edu (Richard Kenner)
CC: berrendo@adacore.com, comar@adacore.com, friess@adacore.com, quinot@adacore.com, 
	<report@adacore.com>
In-Reply-To: <10609041027.AA23751@vlsi1.ultra.nyu.edu>
References: <00a701c6cd0c$276f8740$ca01a8c0@munich>
 <20060901164322.GA92617@melamine.cuivre.fr.eu.org>
 <20060901210815.GA7135@melamine.cuivre.fr.eu.org>
 <44FBEF66.4060908@adacore.com> <44FBF3F0.2050902@adacore.com>
 <1691129980.20060904115223@adacore.com>
 <10609041027.AA23751@vlsi1.ultra.nyu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information in GNAT Pro
Status: O
Content-Length: 633
Lines: 16

> Dunno about VMS, but it certainly can be done on Windows.  You take
> the existing InstallShield file and put it and the license into a zip
> file.  This zip can be built anywhere and the customer unpacks it and
> runs the InstallShield file that looks for the license and copies it in.

First, it's not really technically possible. InstallScript runs in a
temporary directory and doesn't know the location where the initial package
resides. Second, this completely defeats the purpose of auto-installing
packages. It simply is never done on Windows so I am opposed to it for this
reason alone.

Best regards,
 Vasiliy Fofanov.




From kenner@vlsi1.ultra.nyu.edu  Mon Sep  4 06:40:17 2006
From: kenner@vlsi1.ultra.nyu.edu (Richard Kenner)
Message-Id: <10609041043.AA23954@vlsi1.ultra.nyu.edu>
Date: Mon, 04 Sep 2006 06:43:05 EDT
To: fofanov@adacore.com
Cc: report@adacore.com
In-Reply-To: <9110287480.20060904123151@adacore.com>
References: <00a701c6cd0c$276f8740$ca01a8c0@munich>
    <20060901164322.GA92617@melamine.cuivre.fr.eu.org>
    <20060901210815.GA7135@melamine.cuivre.fr.eu.org>
    <44FBEF66.4060908@adacore.com> <44FBF3F0.2050902@adacore.com>
    <1691129980.20060904115223@adacore.com>
    <10609041027.AA23751@vlsi1.ultra.nyu.edu>
    <9110287480.20060904123151@adacore.com>
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information in GNAT Pro
Status: O
Content-Length: 459
Lines: 10

> It simply is never done on Windows 

That's certainly not my experience.  I purchased a lot of software
from Handango for my new phone (a Windows Mobile device).  All of it was
an installer on XP that sent things to the phone.  Half the time, it was
a self-extracting exe file, but the other half it was a zip file that you
are supposed to unpack someplace.  I recall seeing a lot of other Windows
software done that way, but can't point to it right now.



From dewar@adacore.com  Mon Sep  4 06:48:27 2006
Message-ID: <44FC0483.7070401@adacore.com>
Date: Mon, 04 Sep 2006 06:48:35 -0400
From: Robert Dewar <dewar@adacore.com>
MIME-Version: 1.0
To: Richard Kenner <kenner@vlsi1.ultra.nyu.edu>
CC:  fofanov@adacore.com,  report@adacore.com
References: <00a701c6cd0c$276f8740$ca01a8c0@munich>    <20060901164322.GA92617@melamine.cuivre.fr.eu.org>    <20060901210815.GA7135@melamine.cuivre.fr.eu.org>    <44FBEF66.4060908@adacore.com> <44FBF3F0.2050902@adacore.com>    <1691129980.20060904115223@adacore.com>    <10609041027.AA23751@vlsi1.ultra.nyu.edu>    <9110287480.20060904123151@adacore.com> <10609041043.AA23954@vlsi1.ultra.nyu.edu>
In-Reply-To: <10609041043.AA23954@vlsi1.ultra.nyu.edu>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information  in GNAT Pro
Status: O
Content-Length: 576
Lines: 16

Richard Kenner wrote:
>> It simply is never done on Windows 
> 
> That's certainly not my experience.  I purchased a lot of software
> from Handango for my new phone (a Windows Mobile device).  All of it was
> an installer on XP that sent things to the phone.  Half the time, it was
> a self-extracting exe file, but the other half it was a zip file that you
> are supposed to unpack someplace.  I recall seeing a lot of other Windows
> software done that way, but can't point to it right now.

I agree with Richard, this is quite common in windows software
distribution.





From fofanov@adacore.com  Mon Sep  4 06:54:24 2006
Date: Mon, 4 Sep 2006 12:54:23 +0200
From: Vasiliy Fofanov <fofanov@adacore.com>
Message-ID: <213291368.20060904125423@adacore.com>
To: kenner@vlsi1.ultra.nyu.edu (Richard Kenner)
CC: report@adacore.com
In-Reply-To: <10609041043.AA23954@vlsi1.ultra.nyu.edu>
References: <00a701c6cd0c$276f8740$ca01a8c0@munich>
 <20060901164322.GA92617@melamine.cuivre.fr.eu.org>
 <20060901210815.GA7135@melamine.cuivre.fr.eu.org>
 <44FBEF66.4060908@adacore.com> <44FBF3F0.2050902@adacore.com>
 <1691129980.20060904115223@adacore.com>
 <10609041027.AA23751@vlsi1.ultra.nyu.edu>
 <9110287480.20060904123151@adacore.com>
 <10609041043.AA23954@vlsi1.ultra.nyu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information in GNAT Pro
Status: O
Content-Length: 948
Lines: 24

> but the other half it was a zip file that you are supposed to unpack
> someplace. I recall seeing a lot of other Windows software done that way,
> but can't point to it right now.

Hmm short of things like driver updates and similar things I would be quite
surprised if a product I got was a zip file. I'd like Richard and Robert to
give an example, it shouldn't be difficult if it is indeed "quite common".
I would be even more surprised if a zip file contained an installshield.
Zips that include the installer plus a license key file is something I
would mostly expect from some warez site.

But in any case, I don't know of the way to determine where the user
unpacked it...

And of course it is obvious that unzipping, running installer and cleaning
up afterward is certainly much more hassle for the user than just running the
installer, and I thought we agreed we don't want to inconvenience Pro
users.

Best regards,
 Vasiliy Fofanov.




From kenner@vlsi1.ultra.nyu.edu  Mon Sep  4 06:59:04 2006
From: kenner@vlsi1.ultra.nyu.edu (Richard Kenner)
Message-Id: <10609041101.AA24220@vlsi1.ultra.nyu.edu>
Date: Mon, 04 Sep 2006 07:01:54 EDT
To: fofanov@adacore.com
Cc: report@adacore.com
In-Reply-To: <213291368.20060904125423@adacore.com>
References: <00a701c6cd0c$276f8740$ca01a8c0@munich>
    <20060901164322.GA92617@melamine.cuivre.fr.eu.org>
    <20060901210815.GA7135@melamine.cuivre.fr.eu.org>
    <44FBEF66.4060908@adacore.com> <44FBF3F0.2050902@adacore.com>
    <1691129980.20060904115223@adacore.com>
    <10609041027.AA23751@vlsi1.ultra.nyu.edu>
    <9110287480.20060904123151@adacore.com>
    <10609041043.AA23954@vlsi1.ultra.nyu.edu>
    <213291368.20060904125423@adacore.com>
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information in GNAT Pro
Status: O
Content-Length: 276
Lines: 8

> I'd like Richard and Robert to
> give an example, it shouldn't be difficult if it is indeed "quite common".

I *did* give an example: exactly half of the Handango software for my
phone.  I install Windows software so rarely now that I have few examples
either way for it.



From dewar@adacore.com  Mon Sep  4 07:04:02 2006
Message-ID: <44FC0824.8020404@adacore.com>
Date: Mon, 04 Sep 2006 07:04:04 -0400
From: Robert Dewar <dewar@adacore.com>
MIME-Version: 1.0
To: Vasiliy Fofanov <fofanov@adacore.com>
CC: Richard Kenner <kenner@vlsi1.ultra.nyu.edu>,  report@adacore.com
References: <00a701c6cd0c$276f8740$ca01a8c0@munich> <20060901164322.GA92617@melamine.cuivre.fr.eu.org> <20060901210815.GA7135@melamine.cuivre.fr.eu.org> <44FBEF66.4060908@adacore.com> <44FBF3F0.2050902@adacore.com> <1691129980.20060904115223@adacore.com> <10609041027.AA23751@vlsi1.ultra.nyu.edu> <9110287480.20060904123151@adacore.com> <10609041043.AA23954@vlsi1.ultra.nyu.edu> <213291368.20060904125423@adacore.com>
In-Reply-To: <213291368.20060904125423@adacore.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information  in GNAT Pro
Status: O
Content-Length: 1138
Lines: 29

Vasiliy Fofanov wrote:
>> but the other half it was a zip file that you are supposed to unpack
>> someplace. I recall seeing a lot of other Windows software done that way,
>> but can't point to it right now.
> 
> Hmm short of things like driver updates and similar things I would be quite
> surprised if a product I got was a zip file.

A lot of palm software is shipped that way, plus a lot of other things.
I really don't feel like wasiting time looking for specific examples!


> And of course it is obvious that unzipping, running installer and cleaning
> up afterward is certainly much more hassle for the user than just running the
> installer, and I thought we agreed we don't want to inconvenience Pro
> users.

But right now, we are providing no license statement, which is a 
violation of the GPL, which must be fixed, non-discussable.

And if we do provide COPYING and not our own license, that's actively
confusing, and I think untenable.

I really think that individual license agreements are very valuable,
it is not as though we have thousands of customers, and I am sure we
can figure out the technical issues involved




From quinot@adacore.com  Mon Sep  4 07:23:32 2006
Date: Mon, 4 Sep 2006 13:23:30 +0200
From: Thomas Quinot <quinot@adacore.com>
To: Cyrille Comar <comar@adacore.com>
Cc: =?iso-8859-1?Q?Micha=EBl?= Friess <friess@adacore.com>,
	report@adacore.com
Message-ID: <20060904112330.GA30680@melamine.cuivre.fr.eu.org>
References: <00a701c6cd0c$276f8740$ca01a8c0@munich> <20060901164322.GA92617@melamine.cuivre.fr.eu.org> <20060901210815.GA7135@melamine.cuivre.fr.eu.org> <44FBEF66.4060908@adacore.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <44FBEF66.4060908@adacore.com>
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information in GNAT Pro
Status: O
Content-Length: 486
Lines: 14

* Cyrille Comar, 2006-09-04 :

> but doesn't that require re-building a compiler for each evaluation?

As discussed live, this requires repackaging, but not rebuilding, for
each case where we want to distribute a time-limited compiler (the
purpose of the whole design being essentially to dissociate the license
information from the compiler binary).

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From quinot@adacore.com  Mon Sep  4 07:25:39 2006
Date: Mon, 4 Sep 2006 13:25:37 +0200
From: Thomas Quinot <quinot@adacore.com>
To: Romain Berrendonner <berrendo@adacore.com>
Cc: =?iso-8859-1?Q?Micha=EBl?= Friess <friess@adacore.com>,
	report@adacore.com
Message-ID: <20060904112537.GB30680@melamine.cuivre.fr.eu.org>
References: <00a701c6cd0c$276f8740$ca01a8c0@munich> <20060901164322.GA92617@melamine.cuivre.fr.eu.org> <44FBD976.2090707@adacore.com> <20060904083756.GA23836@melamine.cuivre.fr.eu.org> <44FBE8D3.5040407@adacore.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <44FBE8D3.5040407@adacore.com>
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information in GNAT Pro
Status: O
Content-Length: 417
Lines: 13

* Romain Berrendonner, 2006-09-04 :

> This is precisely want I want to discuss ;-)

This discussion of policy does not belong under this TN, which is for
the design and implementation of the underlying supporting mechanism.
Please open a separate TN if you want to discuss policy.

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From quinot@adacore.com  Mon Sep  4 07:26:17 2006
Date: Mon, 4 Sep 2006 13:26:15 +0200
From: Thomas Quinot <quinot@adacore.com>
To: Arnaud Charlet <charlet@adacore.com>
Cc: Romain Berrendonner <berrendo@adacore.com>,
	=?iso-8859-1?Q?Micha=EBl?= Friess <friess@adacore.com>,
	report@adacore.com
Message-ID: <20060904112615.GC30680@melamine.cuivre.fr.eu.org>
References: <00a701c6cd0c$276f8740$ca01a8c0@munich> <20060901164322.GA92617@melamine.cuivre.fr.eu.org> <44FBD976.2090707@adacore.com> <20060904083756.GA23836@melamine.cuivre.fr.eu.org> <44FBE8D3.5040407@adacore.com> <20060904085528.GA45759@adacore.com> <20060904085735.GA51572@adacore.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <20060904085735.GA51572@adacore.com>
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information in GNAT Pro
Status: O
Content-Length: 398
Lines: 13

* Arnaud Charlet, 2006-09-04 :

> - the check will not be included at the FSF, so the implementation should
>   be clearly separated to avoid merge conflicts at the FSF, and be easily
>   disabled (e.g. remove/modify one line)

This is what I had in mind indeed.

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From dewar@adacore.com  Mon Sep  4 07:28:40 2006
Message-ID: <44FC0DF0.1030006@adacore.com>
Date: Mon, 04 Sep 2006 07:28:48 -0400
From: Robert Dewar <dewar@adacore.com>
MIME-Version: 1.0
To: Thomas Quinot <quinot@adacore.com>
CC: Romain Berrendonner <berrendo@adacore.com>, =?ISO-8859-1?Q?Micha=EBl_?=
 =?ISO-8859-1?Q?Friess?= <friess@adacore.com>,  report@adacore.com
References: <00a701c6cd0c$276f8740$ca01a8c0@munich> <20060901164322.GA92617@melamine.cuivre.fr.eu.org> <44FBD976.2090707@adacore.com> <20060904083756.GA23836@melamine.cuivre.fr.eu.org> <44FBE8D3.5040407@adacore.com> <20060904112537.GB30680@melamine.cuivre.fr.eu.org>
In-Reply-To: <20060904112537.GB30680@melamine.cuivre.fr.eu.org>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information  in GNAT Pro
Status: O
Content-Length: 407
Lines: 16

Thomas Quinot wrote:
> * Romain Berrendonner, 2006-09-04 :
> 
>> This is precisely want I want to discuss ;-)
> 
> This discussion of policy does not belong under this TN, which is for
> the design and implementation of the underlying supporting mechanism.
> Please open a separate TN if you want to discuss policy.

Actually discussion of the policy is not really appropriate for
email in any case.
> 





From fofanov@adacore.com  Mon Sep  4 07:32:33 2006
Date: Mon, 4 Sep 2006 13:32:32 +0200
From: Vasiliy Fofanov <fofanov@adacore.com>
Message-ID: <203231446.20060904133232@adacore.com>
To: Robert Dewar <dewar@adacore.com>
CC: Richard Kenner <kenner@vlsi1.ultra.nyu.edu>, report@adacore.com
In-Reply-To: <44FC0824.8020404@adacore.com>
References: <00a701c6cd0c$276f8740$ca01a8c0@munich>
 <20060901164322.GA92617@melamine.cuivre.fr.eu.org>
 <20060901210815.GA7135@melamine.cuivre.fr.eu.org>
 <44FBEF66.4060908@adacore.com> <44FBF3F0.2050902@adacore.com>
 <1691129980.20060904115223@adacore.com>
 <10609041027.AA23751@vlsi1.ultra.nyu.edu>
 <9110287480.20060904123151@adacore.com>
 <10609041043.AA23954@vlsi1.ultra.nyu.edu>
 <213291368.20060904125423@adacore.com> <44FC0824.8020404@adacore.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information in GNAT Pro
Status: O
Content-Length: 229
Lines: 10

> But right now, we are providing no license statement

This statement is wrong, at least as far as Windows kits are concerned.
Besides I am lost, I thought we are talking of license key here?

Best regards,
 Vasiliy Fofanov.




From kenner@vlsi1.ultra.nyu.edu  Mon Sep  4 07:40:54 2006
From: kenner@vlsi1.ultra.nyu.edu (Richard Kenner)
Message-Id: <10609041143.AA24473@vlsi1.ultra.nyu.edu>
Date: Mon, 04 Sep 2006 07:43:45 EDT
To: fofanov@adacore.com
Cc: dewar@adacore.com, report@adacore.com
In-Reply-To: <203231446.20060904133232@adacore.com>
References: <00a701c6cd0c$276f8740$ca01a8c0@munich>
    <20060901164322.GA92617@melamine.cuivre.fr.eu.org>
    <20060901210815.GA7135@melamine.cuivre.fr.eu.org>
    <44FBEF66.4060908@adacore.com> <44FBF3F0.2050902@adacore.com>
    <1691129980.20060904115223@adacore.com>
    <10609041027.AA23751@vlsi1.ultra.nyu.edu>
    <9110287480.20060904123151@adacore.com>
    <10609041043.AA23954@vlsi1.ultra.nyu.edu>
    <213291368.20060904125423@adacore.com>
    <44FC0824.8020404@adacore.com>
    <203231446.20060904133232@adacore.com>
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information in GNAT Pro
Status: O
Content-Length: 460
Lines: 13

> > But right now, we are providing no license statement
> 
> This statement is wrong, at least as far as Windows kits are concerned.

Do you provide the file COPYING and the file that's included as paper in
the box that gives the license conditions?

> Besides I am lost, I thought we are talking of license key here?

We are talking about a text file that contains the actual license.  It's
used somewhat as a "license key", but there isn't an actual key.



From quinot@adacore.com  Mon Sep  4 08:16:59 2006
Date: Mon, 4 Sep 2006 14:16:57 +0200
From: Thomas Quinot <quinot@adacore.com>
To: Richard Kenner <kenner@vlsi1.ultra.nyu.edu>
Cc: fofanov@adacore.com, dewar@adacore.com, report@adacore.com
Message-ID: <20060904121657.GA32121@melamine.cuivre.fr.eu.org>
References: <44FBEF66.4060908@adacore.com> <44FBF3F0.2050902@adacore.com> <1691129980.20060904115223@adacore.com> <10609041027.AA23751@vlsi1.ultra.nyu.edu> <9110287480.20060904123151@adacore.com> <10609041043.AA23954@vlsi1.ultra.nyu.edu> <213291368.20060904125423@adacore.com> <44FC0824.8020404@adacore.com> <203231446.20060904133232@adacore.com> <10609041143.AA24473@vlsi1.ultra.nyu.edu>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <10609041143.AA24473@vlsi1.ultra.nyu.edu>
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information in GNAT Pro
Status: O
Content-Length: 548
Lines: 15

* Richard Kenner, 2006-09-04 :

> We are talking about a text file that contains the actual license.  It's
> used somewhat as a "license key", but there isn't an actual key.

That's not accurate. In the proposed design (coming from the
brainstorming session of this topic), the license file does incorporate
a digital signature used to check its integrity, and the compiler won't
work without a proper signature.

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From berrendo@adacore.com  Mon Sep  4 08:20:08 2006
Message-ID: <44FC19EE.2080505@adacore.com>
Date: Mon, 04 Sep 2006 14:19:58 +0200
From: Romain Berrendonner <berrendo@adacore.com>
MIME-Version: 1.0
To: Robert Dewar <dewar@adacore.com>
CC: Thomas Quinot <quinot@adacore.com>, =?ISO-8859-1?Q?Micha=EBl_Friess?=
 <friess@adacore.com>,  report@adacore.com
References: <00a701c6cd0c$276f8740$ca01a8c0@munich> <20060901164322.GA92617@melamine.cuivre.fr.eu.org> <44FBD976.2090707@adacore.com> <20060904083756.GA23836@melamine.cuivre.fr.eu.org> <44FBE8D3.5040407@adacore.com> <20060904112537.GB30680@melamine.cuivre.fr.eu.org> <44FC0DF0.1030006@adacore.com>
In-Reply-To: <44FC0DF0.1030006@adacore.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information  in GNAT Pro
Status: O
Content-Length: 668
Lines: 18

Robert Dewar a crit :
> Actually discussion of the policy is not really appropriate for
> email in any case.

After discussing with Franco at lunch, I understand that this discussion 
took place during the summer meeting, at a breakout I could not attend.

My concern was that policy decisions we make have in this particular 
area have a huge technical impact. I would therefore not making design 
decisions without taking them into account. Michal's message lifted my 
concerns though.

-- 
Romain Berrendonner                           berrendonner@adacore.com
+33 (0)1 49 70 67 16                            http://www.adacore.com
8 rue de Milan F-75009 Paris



From fofanov@adacore.com  Mon Sep  4 08:21:33 2006
Date: Mon, 4 Sep 2006 14:21:31 +0200
From: Vasiliy Fofanov <fofanov@adacore.com>
Message-ID: <1132887335.20060904142131@adacore.com>
To: kenner@vlsi1.ultra.nyu.edu (Richard Kenner)
CC: dewar@adacore.com, report@adacore.com
In-Reply-To: <10609041143.AA24473@vlsi1.ultra.nyu.edu>
References: <00a701c6cd0c$276f8740$ca01a8c0@munich>
 <20060901164322.GA92617@melamine.cuivre.fr.eu.org>
 <20060901210815.GA7135@melamine.cuivre.fr.eu.org>
 <44FBEF66.4060908@adacore.com> <44FBF3F0.2050902@adacore.com>
 <1691129980.20060904115223@adacore.com>
 <10609041027.AA23751@vlsi1.ultra.nyu.edu>
 <9110287480.20060904123151@adacore.com>
 <10609041043.AA23954@vlsi1.ultra.nyu.edu>
 <213291368.20060904125423@adacore.com>    <44FC0824.8020404@adacore.com>
 <203231446.20060904133232@adacore.com> <10609041143.AA24473@vlsi1.ultra.nyu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information in GNAT Pro
Status: O
Content-Length: 606
Lines: 17

> Do you provide the file COPYING and the file that's included as paper in
> the box that gives the license conditions?

I provide a doc called "GNAT Pro License" that was sent to me by sales,
don't know in what relationship it is with the printed version, I presume
identical?

The COPYING file should be part of the common kit package, I don't control
it myself... And in fact I see that we don't supply it currently which I
agree is in violation of GPL and should be fixed immediately, irrespective
of this discussion. I am opening a separate TN for this right away.

Best regards,
 Vasiliy Fofanov.




From kenner@vlsi1.ultra.nyu.edu  Mon Sep  4 08:28:29 2006
From: kenner@vlsi1.ultra.nyu.edu (Richard Kenner)
Message-Id: <10609041231.AA24664@vlsi1.ultra.nyu.edu>
Date: Mon, 04 Sep 2006 08:31:19 EDT
To: quinot@adacore.com
Cc: dewar@adacore.com, fofanov@adacore.com, report@adacore.com
In-Reply-To: <20060904121657.GA32121@melamine.cuivre.fr.eu.org>
References: <44FBEF66.4060908@adacore.com>
    <44FBF3F0.2050902@adacore.com>
    <1691129980.20060904115223@adacore.com>
    <10609041027.AA23751@vlsi1.ultra.nyu.edu>
    <9110287480.20060904123151@adacore.com>
    <10609041043.AA23954@vlsi1.ultra.nyu.edu>
    <213291368.20060904125423@adacore.com>
    <44FC0824.8020404@adacore.com>
    <203231446.20060904133232@adacore.com>
    <10609041143.AA24473@vlsi1.ultra.nyu.edu>
    <20060904121657.GA32121@melamine.cuivre.fr.eu.org>
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information in GNAT Pro
Status: O
Content-Length: 986
Lines: 21

> * Richard Kenner, 2006-09-04 :
> 
> > We are talking about a text file that contains the actual license.  It's
> > used somewhat as a "license key", but there isn't an actual key.
> 
> That's not accurate. In the proposed design (coming from the
> brainstorming session of this topic), the license file does incorporate
> a digital signature used to check its integrity, and the compiler won't
> work without a proper signature.

Sure, but I mean from the perspective of the user.  The point is that we
aren't issuing some sort of separate "key", but merely using the license
itself for that.  I'm actually not sure I like the idea of a header because
it makes it look more like a key.  I favor a pure text approach, which we
parse.  As to the signature, that's not really a key and can be put at the
end, where it might not even be noticed.

Ideally, we don't even want people to be aware of the checking going on.
(We're not keeping it a secret, but don't want it to be obvious.)



From kenner@vlsi1.ultra.nyu.edu  Mon Sep  4 08:30:09 2006
From: kenner@vlsi1.ultra.nyu.edu (Richard Kenner)
Message-Id: <10609041232.AA24713@vlsi1.ultra.nyu.edu>
Date: Mon, 04 Sep 2006 08:32:55 EDT
To: fofanov@adacore.com
Cc: dewar@adacore.com, report@adacore.com
In-Reply-To: <1132887335.20060904142131@adacore.com>
References: <00a701c6cd0c$276f8740$ca01a8c0@munich>
    <20060901164322.GA92617@melamine.cuivre.fr.eu.org>
    <20060901210815.GA7135@melamine.cuivre.fr.eu.org>
    <44FBEF66.4060908@adacore.com> <44FBF3F0.2050902@adacore.com>
    <1691129980.20060904115223@adacore.com>
    <10609041027.AA23751@vlsi1.ultra.nyu.edu>
    <9110287480.20060904123151@adacore.com>
    <10609041043.AA23954@vlsi1.ultra.nyu.edu>
    <213291368.20060904125423@adacore.com>   
    <44FC0824.8020404@adacore.com>
    <203231446.20060904133232@adacore.com>
    <10609041143.AA24473@vlsi1.ultra.nyu.edu>
    <1132887335.20060904142131@adacore.com>
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information in GNAT Pro
Status: O
Content-Length: 461
Lines: 10

> And in fact I see that we don't supply it currently which I
> agree is in violation of GPL and should be fixed immediately, irrespective
> of this discussion. I am opening a separate TN for this right away.

No, it is the *same* as this discusion.  Because, as Robert said, if we
*do* provide COPYING, but not the license, it'll seriously confuse
people and cause problems.  *This* is the proper TN for dealing with the
missing COPYING file, not a new one.



From quinot@adacore.com  Mon Sep  4 08:33:08 2006
Date: Mon, 4 Sep 2006 14:33:06 +0200
From: Thomas Quinot <quinot@adacore.com>
To: Richard Kenner <kenner@vlsi1.ultra.nyu.edu>
Cc: dewar@adacore.com, fofanov@adacore.com, report@adacore.com
Message-ID: <20060904123306.GA33102@melamine.cuivre.fr.eu.org>
References: <1691129980.20060904115223@adacore.com> <10609041027.AA23751@vlsi1.ultra.nyu.edu> <9110287480.20060904123151@adacore.com> <10609041043.AA23954@vlsi1.ultra.nyu.edu> <213291368.20060904125423@adacore.com> <44FC0824.8020404@adacore.com> <203231446.20060904133232@adacore.com> <10609041143.AA24473@vlsi1.ultra.nyu.edu> <20060904121657.GA32121@melamine.cuivre.fr.eu.org> <10609041231.AA24664@vlsi1.ultra.nyu.edu>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <10609041231.AA24664@vlsi1.ultra.nyu.edu>
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information in GNAT Pro
Status: O
Content-Length: 510
Lines: 17

* Richard Kenner, 2006-09-04 :

> parse.  As to the signature, that's not really a key and can be put at the
> end, where it might not even be noticed.

I don't mind moving the signature to the end of the file, rather than in
the header.
 
> Ideally, we don't even want people to be aware of the checking going on.

People won't be aware of any check unless the check fails.

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From kenner@vlsi1.ultra.nyu.edu  Mon Sep  4 08:43:19 2006
From: kenner@vlsi1.ultra.nyu.edu (Richard Kenner)
Message-Id: <10609041246.AA24826@vlsi1.ultra.nyu.edu>
Date: Mon, 04 Sep 2006 08:46:10 EDT
To: quinot@adacore.com
Cc: dewar@adacore.com, fofanov@adacore.com, report@adacore.com
In-Reply-To: <20060904123306.GA33102@melamine.cuivre.fr.eu.org>
References: <1691129980.20060904115223@adacore.com>
    <10609041027.AA23751@vlsi1.ultra.nyu.edu>
    <9110287480.20060904123151@adacore.com>
    <10609041043.AA23954@vlsi1.ultra.nyu.edu>
    <213291368.20060904125423@adacore.com>
    <44FC0824.8020404@adacore.com>
    <203231446.20060904133232@adacore.com>
    <10609041143.AA24473@vlsi1.ultra.nyu.edu>
    <20060904121657.GA32121@melamine.cuivre.fr.eu.org>
    <10609041231.AA24664@vlsi1.ultra.nyu.edu>
    <20060904123306.GA33102@melamine.cuivre.fr.eu.org>
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information in GNAT Pro
Status: O
Content-Length: 354
Lines: 12

> I don't mind moving the signature to the end of the file, rather than in
> the header.

Why do you feel we need a "header" at all?  Why can't it just be what
looks like free text? E.g.,

"This is an license to evaluate GNAT Pro Eval issued to ABC Corp until
March, 1, 2007."

We can easily enough parse that and it doesn't look like a "license key".



From kenner@vlsi1.ultra.nyu.edu  Mon Sep  4 08:44:23 2006
From: kenner@vlsi1.ultra.nyu.edu (Richard Kenner)
Message-Id: <10609041247.AA24852@vlsi1.ultra.nyu.edu>
Date: Mon, 04 Sep 2006 08:47:14 EDT
To: quinot@adacore.com
Cc: dewar@adacore.com, fofanov@adacore.com, report@adacore.com
In-Reply-To: <20060904123306.GA33102@melamine.cuivre.fr.eu.org>
References: <1691129980.20060904115223@adacore.com>
    <10609041027.AA23751@vlsi1.ultra.nyu.edu>
    <9110287480.20060904123151@adacore.com>
    <10609041043.AA23954@vlsi1.ultra.nyu.edu>
    <213291368.20060904125423@adacore.com>
    <44FC0824.8020404@adacore.com>
    <203231446.20060904133232@adacore.com>
    <10609041143.AA24473@vlsi1.ultra.nyu.edu>
    <20060904121657.GA32121@melamine.cuivre.fr.eu.org>
    <10609041231.AA24664@vlsi1.ultra.nyu.edu>
    <20060904123306.GA33102@melamine.cuivre.fr.eu.org>
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information in GNAT Pro
Status: O
Content-Length: 266
Lines: 8

> I don't mind moving the signature to the end of the file, rather than in
> the header.

The idea is that there'll be an option to gnatls to list the entire file,
so it should be formatted like a license document, with all the particulars
in as part of the text.



From fofanov@adacore.com  Mon Sep  4 08:46:23 2006
Date: Mon, 4 Sep 2006 14:46:21 +0200
From: Vasiliy Fofanov <fofanov@adacore.com>
Message-ID: <1007826322.20060904144621@adacore.com>
To: kenner@vlsi1.ultra.nyu.edu (Richard Kenner)
CC: dewar@adacore.com, report@adacore.com
In-Reply-To: <10609041232.AA24713@vlsi1.ultra.nyu.edu>
References: <00a701c6cd0c$276f8740$ca01a8c0@munich>
 <20060901164322.GA92617@melamine.cuivre.fr.eu.org>
 <20060901210815.GA7135@melamine.cuivre.fr.eu.org>
 <44FBEF66.4060908@adacore.com> <44FBF3F0.2050902@adacore.com>
 <1691129980.20060904115223@adacore.com>
 <10609041027.AA23751@vlsi1.ultra.nyu.edu>
 <9110287480.20060904123151@adacore.com>
 <10609041043.AA23954@vlsi1.ultra.nyu.edu>
 <213291368.20060904125423@adacore.com>       <44FC0824.8020404@adacore.com>
 <203231446.20060904133232@adacore.com>
 <10609041143.AA24473@vlsi1.ultra.nyu.edu>
 <1132887335.20060904142131@adacore.com>
 <10609041232.AA24713@vlsi1.ultra.nyu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information in GNAT Pro
Status: O
Content-Length: 352
Lines: 12

> No, it is the *same* as this discusion. Because, as Robert said, if we
> *do* provide COPYING, but not the license, it'll seriously confuse people
> and cause problems.

Well, your call. I find issue of "violation of GPL" more severe than
"possibility of confusion" but if people feel otherwise it's fine by me...

Best regards,
 Vasiliy Fofanov.




From kenner@vlsi1.ultra.nyu.edu  Mon Sep  4 08:49:21 2006
From: kenner@vlsi1.ultra.nyu.edu (Richard Kenner)
Message-Id: <10609041252.AA24974@vlsi1.ultra.nyu.edu>
Date: Mon, 04 Sep 2006 08:52:10 EDT
To: fofanov@adacore.com
Cc: dewar@adacore.com, report@adacore.com
In-Reply-To: <1007826322.20060904144621@adacore.com>
References: <00a701c6cd0c$276f8740$ca01a8c0@munich>
    <20060901164322.GA92617@melamine.cuivre.fr.eu.org>
    <20060901210815.GA7135@melamine.cuivre.fr.eu.org>
    <44FBEF66.4060908@adacore.com> <44FBF3F0.2050902@adacore.com>
    <1691129980.20060904115223@adacore.com>
    <10609041027.AA23751@vlsi1.ultra.nyu.edu>
    <9110287480.20060904123151@adacore.com>
    <10609041043.AA23954@vlsi1.ultra.nyu.edu>
    <213291368.20060904125423@adacore.com>      
    <44FC0824.8020404@adacore.com>
    <203231446.20060904133232@adacore.com>
    <10609041143.AA24473@vlsi1.ultra.nyu.edu>
    <1132887335.20060904142131@adacore.com>
    <10609041232.AA24713@vlsi1.ultra.nyu.edu>
    <1007826322.20060904144621@adacore.com>
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information in GNAT Pro
Status: O
Content-Length: 443
Lines: 9

> Well, your call. I find issue of "violation of GPL" more severe than
> "possibility of confusion" but if people feel otherwise it's fine by me...

No, it's certainly not more severe.  The GPL violation is minor and the worst
that would happen is that RMS would ask us to fix it, which we plan to do
anyway.  Having customers think their code was suddenly covered by the GPL
is *extremely* serious and could easily result in lost business.



From quinot@adacore.com  Mon Sep  4 08:53:17 2006
Date: Mon, 4 Sep 2006 14:53:14 +0200
From: Thomas Quinot <quinot@adacore.com>
To: Richard Kenner <kenner@vlsi1.ultra.nyu.edu>
Cc: dewar@adacore.com, fofanov@adacore.com, report@adacore.com
Message-ID: <20060904125314.GB33102@melamine.cuivre.fr.eu.org>
References: <9110287480.20060904123151@adacore.com> <10609041043.AA23954@vlsi1.ultra.nyu.edu> <213291368.20060904125423@adacore.com> <44FC0824.8020404@adacore.com> <203231446.20060904133232@adacore.com> <10609041143.AA24473@vlsi1.ultra.nyu.edu> <20060904121657.GA32121@melamine.cuivre.fr.eu.org> <10609041231.AA24664@vlsi1.ultra.nyu.edu> <20060904123306.GA33102@melamine.cuivre.fr.eu.org> <10609041246.AA24826@vlsi1.ultra.nyu.edu>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <10609041246.AA24826@vlsi1.ultra.nyu.edu>
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information in GNAT Pro
Status: O
Content-Length: 672
Lines: 21

* Richard Kenner, 2006-09-04 :

> Why do you feel we need a "header" at all?  Why can't it just be what
> looks like free text? E.g.,
> 
> "This is an license to evaluate GNAT Pro Eval issued to ABC Corp until
> March, 1, 2007."
> We can easily enough parse that and it doesn't look like a "license key".

In my opinion, compared to a simple "keyword: value" format, this would
be a royal pain to parse. Also, your proposal as stated does not have
any provision for optional fields (limit date, host platform, target
platform).

Thomas.

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From friess@adacore.com  Mon Sep  4 09:13:59 2006
From: =?iso-8859-1?Q?Micha=EBl_Friess?= <friess@adacore.com>
To: "'Arnaud Charlet'" <charlet@adacore.com>
Cc: <report@adacore.com>,
	"'Sales'" <sales@eu.adacore.com>
Date: Mon, 4 Sep 2006 15:13:49 +0200
Message-ID: <001801c6d023$f6210630$89000a0a@munich>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
In-Reply-To: <20060904100346.GA26769@adacore.com>
Subject: RE: [F831-014] - #999 including GNAT Pro license and customer information in GNAT Pro
Status: O
Content-Length: 1445
Lines: 41

> -----Message d'origine-----
> De: Arnaud Charlet [mailto:charlet@adacore.com]
> Envoy: Monday, September 04, 2006 12:04 PM
> : Michal Friess
> Cc: report@adacore.com; 'Sales'
> Objet: Re: [F831-014] - #999 including GNAT Pro license and customer
> information in GNAT Pro
> 
> So you are proposing to change that, right ?
> Could you (or someone else) clarify what would be changed ?

The introduction of license verification for evaluations in Europe will
indeed need some modification in our eval scheme.
 
I can think of the following steps:

- take an existing GNAT Pro package that corresponds to the product the
prospect wants to evaluate.

- unpack this package if necessary and put an eval license file at the right
place in the package (license file generated by a license generation tool)

- package GNAT Pro that will now be GNAT Pro Eval and put it on the
prospect's account.

If all the previous steps are done manually we will probably have to provide
GNAT Pro Eval as a wavefront.

We could also consider an automatic process that could be handled by the
CRM/GNAT Tracker. But this would require more work & discussion on our tools
so I'd rather go for the manual process for now.

Now, of course, there are still some resources from the technical team that
are required to put GNAT Pro Eval on the prospect's account. I guess that it
would be reasonable in this context to let the sales team do this work.

Michal





From kenner@vlsi1.ultra.nyu.edu  Mon Sep  4 09:16:57 2006
From: kenner@vlsi1.ultra.nyu.edu (Richard Kenner)
Message-Id: <10609041319.AA25101@vlsi1.ultra.nyu.edu>
Date: Mon, 04 Sep 2006 09:19:46 EDT
To: quinot@adacore.com
Cc: dewar@adacore.com, fofanov@adacore.com, report@adacore.com
In-Reply-To: <20060904125314.GB33102@melamine.cuivre.fr.eu.org>
References: <9110287480.20060904123151@adacore.com>
    <10609041043.AA23954@vlsi1.ultra.nyu.edu>
    <213291368.20060904125423@adacore.com>
    <44FC0824.8020404@adacore.com>
    <203231446.20060904133232@adacore.com>
    <10609041143.AA24473@vlsi1.ultra.nyu.edu>
    <20060904121657.GA32121@melamine.cuivre.fr.eu.org>
    <10609041231.AA24664@vlsi1.ultra.nyu.edu>
    <20060904123306.GA33102@melamine.cuivre.fr.eu.org>
    <10609041246.AA24826@vlsi1.ultra.nyu.edu>
    <20060904125314.GB33102@melamine.cuivre.fr.eu.org>
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information in GNAT Pro
Status: O
Content-Length: 761
Lines: 19

> In my opinion, compared to a simple "keyword: value" format, this would
> be a royal pain to parse. 

Well, Robert volunteered to write the code and he seemed fine with parsing
something like the above.  The most important thing here is how the user
views the license, not a few more lines of code to parse it.

> Also, your proposal as stated does not have
> any provision for optional fields (limit date, host platform, target
> platform).

Sure it does.  I included the limit date and you can easily add "on
a Sparc/Solaris host" (or similar language) for other limitations.

Parsing general English is hard, but if we write the sentence in a very
structured form, it seems to be no harder at all than parsing fixed fields.
Why do you think it's harder?



From charlet@ACT-Europe.FR  Mon Sep  4 09:19:26 2006
Date: Mon, 4 Sep 2006 15:19:23 +0200
From: Arnaud Charlet <charlet@adacore.com>
To: =?iso-8859-1?Q?Micha=EBl?= Friess <friess@adacore.com>
Cc: report@adacore.com, 'Sales' <sales@eu.adacore.com>
Message-ID: <20060904131923.GA78669@adacore.com>
References: <20060904100346.GA26769@adacore.com> <001801c6d023$f6210630$89000a0a@munich>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <001801c6d023$f6210630$89000a0a@munich>
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information in GNAT Pro
Status: O
Content-Length: 1442
Lines: 34

> - take an existing GNAT Pro package that corresponds to the product the
> prospect wants to evaluate.
> 
> - unpack this package if necessary and put an eval license file at the right
> place in the package (license file generated by a license generation tool)

As Vasiliy explained, this is not possible under Windows with our
InstallShield technology, so it's not clear ewhat you are proposing instead.

> If all the previous steps are done manually we will probably have to provide
> GNAT Pro Eval as a wavefront.

And how would you configure their gnattracker account then ? You cannot
put e.g. x86-windows as part of their supported platforms since they would
have access to the x86-windows platform in the download area I guess ?

So are we talking about some kind of new account that would have no access
to the 'download' page ?

> We could also consider an automatic process that could be handled by the
> CRM/GNAT Tracker. But this would require more work & discussion on our tools
> so I'd rather go for the manual process for now.
> 
> Now, of course, there are still some resources from the technical team that
> are required to put GNAT Pro Eval on the prospect's account. I guess that it
> would be reasonable in this context to let the sales team do this work.

Yes, I would strongly prefer that all the manual steps described above
(which currently we do not know how to do technically) are done by the
sales team.

Arno



From dewar@adacore.com  Mon Sep  4 09:44:08 2006
Message-ID: <44FC2DB1.9080505@adacore.com>
Date: Mon, 04 Sep 2006 09:44:17 -0400
From: Robert Dewar <dewar@adacore.com>
MIME-Version: 1.0
To: Vasiliy Fofanov <fofanov@adacore.com>
CC: Richard Kenner <kenner@vlsi1.ultra.nyu.edu>,  report@adacore.com
References: <00a701c6cd0c$276f8740$ca01a8c0@munich> <20060901164322.GA92617@melamine.cuivre.fr.eu.org> <20060901210815.GA7135@melamine.cuivre.fr.eu.org> <44FBEF66.4060908@adacore.com> <44FBF3F0.2050902@adacore.com> <1691129980.20060904115223@adacore.com> <10609041027.AA23751@vlsi1.ultra.nyu.edu> <9110287480.20060904123151@adacore.com> <10609041043.AA23954@vlsi1.ultra.nyu.edu> <213291368.20060904125423@adacore.com> <44FC0824.8020404@adacore.com> <203231446.20060904133232@adacore.com>
In-Reply-To: <203231446.20060904133232@adacore.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information  in GNAT Pro
Status: O
Content-Length: 709
Lines: 23

Vasiliy Fofanov wrote:
>> But right now, we are providing no license statement
> 
> This statement is wrong, at least as far as Windows kits are concerned.
> Besides I am lost, I thought we are talking of license key here?

No, we are not talking about license keys, that's a *total* confusion,
we are talking about the license. If we are talking about license keys
that's the phrase we use!

We provide a license for GNAT Pro that is currently available ONLY
in the box, not in electronic form. The idea is to provide this
license in a more direct form.

At least for the unix distributions we are not even providing the
GPL COPYING file, which is a GPL violation.
> 
> Best regards,
>  Vasiliy Fofanov.





From dewar@adacore.com  Mon Sep  4 09:47:15 2006
Message-ID: <44FC2E6C.4050209@adacore.com>
Date: Mon, 04 Sep 2006 09:47:24 -0400
From: Robert Dewar <dewar@adacore.com>
MIME-Version: 1.0
To: Thomas Quinot <quinot@adacore.com>
CC: Richard Kenner <kenner@vlsi1.ultra.nyu.edu>,  fofanov@adacore.com, 
 report@adacore.com
References: <44FBEF66.4060908@adacore.com> <44FBF3F0.2050902@adacore.com> <1691129980.20060904115223@adacore.com> <10609041027.AA23751@vlsi1.ultra.nyu.edu> <9110287480.20060904123151@adacore.com> <10609041043.AA23954@vlsi1.ultra.nyu.edu> <213291368.20060904125423@adacore.com> <44FC0824.8020404@adacore.com> <203231446.20060904133232@adacore.com> <10609041143.AA24473@vlsi1.ultra.nyu.edu> <20060904121657.GA32121@melamine.cuivre.fr.eu.org>
In-Reply-To: <20060904121657.GA32121@melamine.cuivre.fr.eu.org>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information  in GNAT Pro
Status: O
Content-Length: 617
Lines: 19

Thomas Quinot wrote:
> * Richard Kenner, 2006-09-04 :
> 
>> We are talking about a text file that contains the actual license.  It's
>> used somewhat as a "license key", but there isn't an actual key.
> 
> That's not accurate. In the proposed design (coming from the
> brainstorming session of this topic), the license file does incorporate
> a digital signature used to check its integrity, and the compiler won't
> work without a proper signature.

Right, but to think of it as a license key is confusing, the signature
is there (in the case of GNAT Pro) simply to ensure integrity of the
license agreement.
> 





From dewar@adacore.com  Mon Sep  4 09:48:44 2006
Message-ID: <44FC2EC5.6090708@adacore.com>
Date: Mon, 04 Sep 2006 09:48:53 -0400
From: Robert Dewar <dewar@adacore.com>
MIME-Version: 1.0
To: Romain Berrendonner <berrendo@adacore.com>
CC: Thomas Quinot <quinot@adacore.com>, =?ISO-8859-1?Q?Micha=EBl_Friess?=
 <friess@adacore.com>,  report@adacore.com
References: <00a701c6cd0c$276f8740$ca01a8c0@munich> <20060901164322.GA92617@melamine.cuivre.fr.eu.org> <44FBD976.2090707@adacore.com> <20060904083756.GA23836@melamine.cuivre.fr.eu.org> <44FBE8D3.5040407@adacore.com> <20060904112537.GB30680@melamine.cuivre.fr.eu.org> <44FC0DF0.1030006@adacore.com> <44FC19EE.2080505@adacore.com>
In-Reply-To: <44FC19EE.2080505@adacore.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information  in GNAT Pro
Status: O
Content-Length: 551
Lines: 17

Romain Berrendonner wrote:

> After discussing with Franco at lunch, I understand that this discussion 
> took place during the summer meeting, at a breakout I could not attend.
> 
> My concern was that policy decisions we make have in this particular 
> area have a huge technical impact. I would therefore not making design 
> decisions without taking them into account. Michal's message lifted my 
> concerns though.

Right, at the breakout session we did have all points of concern
represented, including the concerns of the release team.
> 





From dewar@adacore.com  Mon Sep  4 09:49:37 2006
Message-ID: <44FC2EFA.70208@adacore.com>
Date: Mon, 04 Sep 2006 09:49:46 -0400
From: Robert Dewar <dewar@adacore.com>
MIME-Version: 1.0
To: Vasiliy Fofanov <fofanov@adacore.com>
CC: Richard Kenner <kenner@vlsi1.ultra.nyu.edu>,  report@adacore.com
References: <00a701c6cd0c$276f8740$ca01a8c0@munich> <20060901164322.GA92617@melamine.cuivre.fr.eu.org> <20060901210815.GA7135@melamine.cuivre.fr.eu.org> <44FBEF66.4060908@adacore.com> <44FBF3F0.2050902@adacore.com> <1691129980.20060904115223@adacore.com> <10609041027.AA23751@vlsi1.ultra.nyu.edu> <9110287480.20060904123151@adacore.com> <10609041043.AA23954@vlsi1.ultra.nyu.edu> <213291368.20060904125423@adacore.com>    <44FC0824.8020404@adacore.com> <203231446.20060904133232@adacore.com> <10609041143.AA24473@vlsi1.ultra.nyu.edu> <1132887335.20060904142131@adacore.com>
In-Reply-To: <1132887335.20060904142131@adacore.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information  in GNAT Pro
Status: O
Content-Length: 679
Lines: 20

Vasiliy Fofanov wrote:

> I provide a doc called "GNAT Pro License" that was sent to me by sales,
> don't know in what relationship it is with the printed version, I presume
> identical?
> 
> The COPYING file should be part of the common kit package, I don't control
> it myself... And in fact I see that we don't supply it currently which I
> agree is in violation of GPL and should be fixed immediately, irrespective
> of this discussion. I am opening a separate TN for this right away.

No need to open a separate TN for that, we certainly do NOT want to go
ahead and include COPYING without including our license statement as well.
> 
> Best regards,
>  Vasiliy Fofanov.





From dewar@adacore.com  Mon Sep  4 09:52:18 2006
Message-ID: <44FC2F9B.9030409@adacore.com>
Date: Mon, 04 Sep 2006 09:52:27 -0400
From: Robert Dewar <dewar@adacore.com>
MIME-Version: 1.0
To: Richard Kenner <kenner@vlsi1.ultra.nyu.edu>
CC:  quinot@adacore.com,  fofanov@adacore.com,  report@adacore.com
References: <44FBEF66.4060908@adacore.com>    <44FBF3F0.2050902@adacore.com>    <1691129980.20060904115223@adacore.com>    <10609041027.AA23751@vlsi1.ultra.nyu.edu>    <9110287480.20060904123151@adacore.com>    <10609041043.AA23954@vlsi1.ultra.nyu.edu>    <213291368.20060904125423@adacore.com>    <44FC0824.8020404@adacore.com>    <203231446.20060904133232@adacore.com>    <10609041143.AA24473@vlsi1.ultra.nyu.edu>    <20060904121657.GA32121@melamine.cuivre.fr.eu.org> <10609041231.AA24664@vlsi1.ultra.nyu.edu>
In-Reply-To: <10609041231.AA24664@vlsi1.ultra.nyu.edu>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information  in GNAT Pro
Status: O
Content-Length: 662
Lines: 19

Richard Kenner wrote:

> Sure, but I mean from the perspective of the user.  The point is that we
> aren't issuing some sort of separate "key", but merely using the license
> itself for that.  I'm actually not sure I like the idea of a header because
> it makes it look more like a key.  I favor a pure text approach, which we
> parse.  As to the signature, that's not really a key and can be put at the
> end, where it might not even be noticed.

I fully agree with all points in this paraghraph
> 
> Ideally, we don't even want people to be aware of the checking going on.
> (We're not keeping it a secret, but don't want it to be obvious.)

Exactly right





From dewar@adacore.com  Mon Sep  4 09:53:09 2006
Message-ID: <44FC2FCE.2040601@adacore.com>
Date: Mon, 04 Sep 2006 09:53:18 -0400
From: Robert Dewar <dewar@adacore.com>
MIME-Version: 1.0
To: Richard Kenner <kenner@vlsi1.ultra.nyu.edu>
CC:  fofanov@adacore.com,  report@adacore.com
References: <00a701c6cd0c$276f8740$ca01a8c0@munich>    <20060901164322.GA92617@melamine.cuivre.fr.eu.org>    <20060901210815.GA7135@melamine.cuivre.fr.eu.org>    <44FBEF66.4060908@adacore.com> <44FBF3F0.2050902@adacore.com>    <1691129980.20060904115223@adacore.com>    <10609041027.AA23751@vlsi1.ultra.nyu.edu>    <9110287480.20060904123151@adacore.com>    <10609041043.AA23954@vlsi1.ultra.nyu.edu>    <213291368.20060904125423@adacore.com>       <44FC0824.8020404@adacore.com>    <203231446.20060904133232@adacore.com>    <10609041143.AA24473@vlsi1.ultra.nyu.edu>    <1132887335.20060904142131@adacore.com> <10609041232.AA24713@vlsi1.ultra.nyu.edu>
In-Reply-To: <10609041232.AA24713@vlsi1.ultra.nyu.edu>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information  in GNAT Pro (also F904-003)
Status: O
Content-Length: 589
Lines: 15

Richard Kenner wrote:
>> And in fact I see that we don't supply it currently which I
>> agree is in violation of GPL and should be fixed immediately, irrespective
>> of this discussion. I am opening a separate TN for this right away.
> 
> No, it is the *same* as this discusion.  Because, as Robert said, if we
> *do* provide COPYING, but not the license, it'll seriously confuse
> people and cause problems.  *This* is the proper TN for dealing with the
> missing COPYING file, not a new one.

Reading Richard's comment here, I definitely agree, and I have closed
F904-003 accordingly




From dewar@adacore.com  Mon Sep  4 09:54:06 2006
Message-ID: <44FC3007.8050100@adacore.com>
Date: Mon, 04 Sep 2006 09:54:15 -0400
From: Robert Dewar <dewar@adacore.com>
MIME-Version: 1.0
To: Thomas Quinot <quinot@adacore.com>
CC: Richard Kenner <kenner@vlsi1.ultra.nyu.edu>,  fofanov@adacore.com, 
 report@adacore.com
References: <1691129980.20060904115223@adacore.com> <10609041027.AA23751@vlsi1.ultra.nyu.edu> <9110287480.20060904123151@adacore.com> <10609041043.AA23954@vlsi1.ultra.nyu.edu> <213291368.20060904125423@adacore.com> <44FC0824.8020404@adacore.com> <203231446.20060904133232@adacore.com> <10609041143.AA24473@vlsi1.ultra.nyu.edu> <20060904121657.GA32121@melamine.cuivre.fr.eu.org> <10609041231.AA24664@vlsi1.ultra.nyu.edu> <20060904123306.GA33102@melamine.cuivre.fr.eu.org>
In-Reply-To: <20060904123306.GA33102@melamine.cuivre.fr.eu.org>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information  in GNAT Pro
Status: O
Content-Length: 585
Lines: 20

Thomas Quinot wrote:
> * Richard Kenner, 2006-09-04 :
> 
>> parse.  As to the signature, that's not really a key and can be put at the
>> end, where it might not even be noticed.
> 
> I don't mind moving the signature to the end of the file, rather than in
> the header.
>  
>> Ideally, we don't even want people to be aware of the checking going on.
> 
> People won't be aware of any check unless the check fails.
> 

But I think Richard is right, this should be a text file which we
parse, and should not look at all like a license key, which the
current header does not satisfy.




From dewar@adacore.com  Mon Sep  4 09:56:18 2006
Message-ID: <44FC308B.1050503@adacore.com>
Date: Mon, 04 Sep 2006 09:56:27 -0400
From: Robert Dewar <dewar@adacore.com>
MIME-Version: 1.0
To: Vasiliy Fofanov <fofanov@adacore.com>
CC: Richard Kenner <kenner@vlsi1.ultra.nyu.edu>,  report@adacore.com
References: <00a701c6cd0c$276f8740$ca01a8c0@munich> <20060901164322.GA92617@melamine.cuivre.fr.eu.org> <20060901210815.GA7135@melamine.cuivre.fr.eu.org> <44FBEF66.4060908@adacore.com> <44FBF3F0.2050902@adacore.com> <1691129980.20060904115223@adacore.com> <10609041027.AA23751@vlsi1.ultra.nyu.edu> <9110287480.20060904123151@adacore.com> <10609041043.AA23954@vlsi1.ultra.nyu.edu> <213291368.20060904125423@adacore.com>       <44FC0824.8020404@adacore.com> <203231446.20060904133232@adacore.com> <10609041143.AA24473@vlsi1.ultra.nyu.edu> <1132887335.20060904142131@adacore.com> <10609041232.AA24713@vlsi1.ultra.nyu.edu> <1007826322.20060904144621@adacore.com>
In-Reply-To: <1007826322.20060904144621@adacore.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information  in GNAT Pro
Status: O
Content-Length: 788
Lines: 25

Vasiliy Fofanov wrote:
>> No, it is the *same* as this discusion. Because, as Robert said, if we
>> *do* provide COPYING, but not the license, it'll seriously confuse people
>> and cause problems.
> 
> Well, your call. I find issue of "violation of GPL" more severe than
> "possibility of confusion" but if people feel otherwise it's fine by me...

That's an overreaction indeed.

Confusing our customers and worrying them about a change to the 
licensing in the GPL direction could have serious immediate
consequences.

Continuing our failure to provide the COPYING file for another month
or so after doing this for 10 years hardly qualifies as a severe and
urgent matter. No one is going to complain and if they do, we just
stick in the file!
> 
> Best regards,
>  Vasiliy Fofanov.





From dewar@adacore.com  Mon Sep  4 09:57:55 2006
Message-ID: <44FC30EB.8080004@adacore.com>
Date: Mon, 04 Sep 2006 09:58:03 -0400
From: Robert Dewar <dewar@adacore.com>
MIME-Version: 1.0
To: Thomas Quinot <quinot@adacore.com>
CC: Richard Kenner <kenner@vlsi1.ultra.nyu.edu>,  fofanov@adacore.com, 
 report@adacore.com
References: <9110287480.20060904123151@adacore.com> <10609041043.AA23954@vlsi1.ultra.nyu.edu> <213291368.20060904125423@adacore.com> <44FC0824.8020404@adacore.com> <203231446.20060904133232@adacore.com> <10609041143.AA24473@vlsi1.ultra.nyu.edu> <20060904121657.GA32121@melamine.cuivre.fr.eu.org> <10609041231.AA24664@vlsi1.ultra.nyu.edu> <20060904123306.GA33102@melamine.cuivre.fr.eu.org> <10609041246.AA24826@vlsi1.ultra.nyu.edu> <20060904125314.GB33102@melamine.cuivre.fr.eu.org>
In-Reply-To: <20060904125314.GB33102@melamine.cuivre.fr.eu.org>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information  in GNAT Pro
Status: O
Content-Length: 907
Lines: 27

Thomas Quinot wrote:
> * Richard Kenner, 2006-09-04 :
> 
>> Why do you feel we need a "header" at all?  Why can't it just be what
>> looks like free text? E.g.,
>>
>> "This is an license to evaluate GNAT Pro Eval issued to ABC Corp until
>> March, 1, 2007."
>> We can easily enough parse that and it doesn't look like a "license key".
> 
> In my opinion, compared to a simple "keyword: value" format, this would
> be a royal pain to parse. Also, your proposal as stated does not have
> any provision for optional fields (limit date, host platform, target
> platform).

OK, I will be happy to do the code for the parsing, which seems trivial
to me. We should not base the design on one person's antipathy to 
writing what is really a simple task. Why don't I take over this TN
and propose something along the lines Richard suggests, and then people
can decide which approach they prefer.
> 
> Thomas.
> 





From berrendo@adacore.com  Mon Sep  4 10:01:03 2006
Message-ID: <44FC3194.5000203@adacore.com>
Date: Mon, 04 Sep 2006 16:00:52 +0200
From: Romain Berrendonner <berrendo@adacore.com>
MIME-Version: 1.0
To: Robert Dewar <dewar@adacore.com>
CC: Thomas Quinot <quinot@adacore.com>, =?ISO-8859-1?Q?Micha=EBl_Friess?=
 <friess@adacore.com>,  report@adacore.com
References: <00a701c6cd0c$276f8740$ca01a8c0@munich> <20060901164322.GA92617@melamine.cuivre.fr.eu.org> <44FBD976.2090707@adacore.com> <20060904083756.GA23836@melamine.cuivre.fr.eu.org> <44FBE8D3.5040407@adacore.com> <20060904112537.GB30680@melamine.cuivre.fr.eu.org> <44FC0DF0.1030006@adacore.com> <44FC19EE.2080505@adacore.com> <44FC2EC5.6090708@adacore.com>
In-Reply-To: <44FC2EC5.6090708@adacore.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information  in GNAT Pro
Status: O
Content-Length: 570
Lines: 16

Robert Dewar a crit :
> Right, at the breakout session we did have all points of concern
> represented, including the concerns of the release team.

You mean Nico R. I presume ? As this has definitely huge importance in 
term of packaging and software distribution (as shown by this whole 
thread), I would have preferred attending such an important meeting.

Anyway, let's steam ahead now.

-- 
Romain Berrendonner                           berrendonner@adacore.com
+33 (0)1 49 70 67 16                            http://www.adacore.com
8 rue de Milan F-75009 Paris



From dewar@adacore.com  Mon Sep  4 10:05:04 2006
Message-ID: <44FC3299.80201@adacore.com>
Date: Mon, 04 Sep 2006 10:05:13 -0400
From: Robert Dewar <dewar@adacore.com>
MIME-Version: 1.0
To: Romain Berrendonner <berrendo@adacore.com>
CC: Thomas Quinot <quinot@adacore.com>, =?ISO-8859-1?Q?Micha=EBl_Friess?=
 <friess@adacore.com>,  report@adacore.com
References: <00a701c6cd0c$276f8740$ca01a8c0@munich> <20060901164322.GA92617@melamine.cuivre.fr.eu.org> <44FBD976.2090707@adacore.com> <20060904083756.GA23836@melamine.cuivre.fr.eu.org> <44FBE8D3.5040407@adacore.com> <20060904112537.GB30680@melamine.cuivre.fr.eu.org> <44FC0DF0.1030006@adacore.com> <44FC19EE.2080505@adacore.com> <44FC2EC5.6090708@adacore.com> <44FC3194.5000203@adacore.com>
In-Reply-To: <44FC3194.5000203@adacore.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information  in GNAT Pro
Status: O
Content-Length: 625
Lines: 19

Romain Berrendonner wrote:
> Robert Dewar a crit :
>> Right, at the breakout session we did have all points of concern
>> represented, including the concerns of the release team.
> 
> You mean Nico R. I presume ? As this has definitely huge importance in 
> term of packaging and software distribution (as shown by this whole 
> thread), I would have preferred attending such an important meeting.

Well as you know, people choose the breakout sessions they want to
attend, and we can't all be everywhere, which is why we don't take
any final decisions, just make recommendations.
> 
> Anyway, let's steam ahead now.
> 





From quinot@adacore.com  Mon Sep  4 10:41:33 2006
Date: Mon, 4 Sep 2006 16:41:31 +0200
From: Thomas Quinot <quinot@adacore.com>
To: Robert Dewar <dewar@adacore.com>
Cc: Richard Kenner <kenner@vlsi1.ultra.nyu.edu>, fofanov@adacore.com,
	report@adacore.com
Message-ID: <20060904144131.GA37929@melamine.cuivre.fr.eu.org>
References: <9110287480.20060904123151@adacore.com> <10609041043.AA23954@vlsi1.ultra.nyu.edu> <213291368.20060904125423@adacore.com> <44FC0824.8020404@adacore.com> <203231446.20060904133232@adacore.com> <10609041143.AA24473@vlsi1.ultra.nyu.edu> <20060904121657.GA32121@melamine.cuivre.fr.eu.org> <10609041231.AA24664@vlsi1.ultra.nyu.edu> <20060904123306.GA33102@melamine.cuivre.fr.eu.org> <44FC3007.8050100@adacore.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <44FC3007.8050100@adacore.com>
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information in GNAT Pro
Status: O
Content-Length: 380
Lines: 13

* Robert Dewar, 2006-09-04 :

> But I think Richard is right, this should be a text file which we
> parse, and should not look at all like a license key, which the
> current header does not satisfy.

OK, please propose an alternate format then.

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From dewar@adacore.com  Mon Sep  4 11:41:25 2006
Message-ID: <44FC492F.1080208@adacore.com>
Date: Mon, 04 Sep 2006 11:41:35 -0400
From: Robert Dewar <dewar@adacore.com>
MIME-Version: 1.0
To: Thomas Quinot <quinot@adacore.com>
CC: Richard Kenner <kenner@vlsi1.ultra.nyu.edu>,  fofanov@adacore.com, 
 report@adacore.com
References: <9110287480.20060904123151@adacore.com> <10609041043.AA23954@vlsi1.ultra.nyu.edu> <213291368.20060904125423@adacore.com> <44FC0824.8020404@adacore.com> <203231446.20060904133232@adacore.com> <10609041143.AA24473@vlsi1.ultra.nyu.edu> <20060904121657.GA32121@melamine.cuivre.fr.eu.org> <10609041231.AA24664@vlsi1.ultra.nyu.edu> <20060904123306.GA33102@melamine.cuivre.fr.eu.org> <44FC3007.8050100@adacore.com> <20060904144131.GA37929@melamine.cuivre.fr.eu.org>
In-Reply-To: <20060904144131.GA37929@melamine.cuivre.fr.eu.org>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information  in GNAT Pro
Status: O
Content-Length: 350
Lines: 16

Thomas Quinot wrote:
> * Robert Dewar, 2006-09-04 :
> 
>> But I think Richard is right, this should be a text file which we
>> parse, and should not look at all like a license key, which the
>> current header does not satisfy.
> 
> OK, please propose an alternate format then.

Will do
will try to find a format that is happy for everyone :-)
> 





From quinot@adacore.com  Mon Sep 25 12:00:53 2006
Date: Mon, 25 Sep 2006 18:00:45 +0200
From: Thomas Quinot <quinot@adacore.com>
To: Robert Dewar <dewar@adacore.com>
Cc: Richard Kenner <kenner@vlsi1.ultra.nyu.edu>, fofanov@adacore.com,
	report@adacore.com
Message-ID: <20060925160045.GB90552@melamine.cuivre.fr.eu.org>
References: <213291368.20060904125423@adacore.com> <44FC0824.8020404@adacore.com> <203231446.20060904133232@adacore.com> <10609041143.AA24473@vlsi1.ultra.nyu.edu> <20060904121657.GA32121@melamine.cuivre.fr.eu.org> <10609041231.AA24664@vlsi1.ultra.nyu.edu> <20060904123306.GA33102@melamine.cuivre.fr.eu.org> <44FC3007.8050100@adacore.com> <20060904144131.GA37929@melamine.cuivre.fr.eu.org> <44FC492F.1080208@adacore.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <44FC492F.1080208@adacore.com>
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information in GNAT Pro
Status: O
Content-Length: 1015
Lines: 33

* Robert Dewar, 2006-09-04 :

> >>But I think Richard is right, this should be a text file which we
> >>parse, and should not look at all like a license key, which the
> >>current header does not satisfy.

OK, since there was no new on this TN, I prepared a parser that expects
the following format for the installation parameters file:

----------
The following license to [use|evaluate] GNAT <buildtype-version>
for <target-platform> [hosted on <host-platform>] is granted to
<licensee> until <validity-date>.

<text from the GMGPL, or whatever license>

Signature: rrrr ssss
----------

Does that look better?

If so, I'll proceed with completing implementation of the checks and
checking it required supporting code (inlcuding modular exponentiation
for uints, which should go either as a child of uintp or directly within
uintp, since it relies on Least_Sig_Digit).

Thomas.

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From dewar@adacore.com  Mon Sep 25 13:48:16 2006
Message-ID: <4518165E.1020008@adacore.com>
Date: Mon, 25 Sep 2006 13:48:14 -0400
From: Robert Dewar <dewar@adacore.com>
MIME-Version: 1.0
To: Thomas Quinot <quinot@adacore.com>
CC: Richard Kenner <kenner@vlsi1.ultra.nyu.edu>,  fofanov@adacore.com, 
 report@adacore.com
References: <213291368.20060904125423@adacore.com> <44FC0824.8020404@adacore.com> <203231446.20060904133232@adacore.com> <10609041143.AA24473@vlsi1.ultra.nyu.edu> <20060904121657.GA32121@melamine.cuivre.fr.eu.org> <10609041231.AA24664@vlsi1.ultra.nyu.edu> <20060904123306.GA33102@melamine.cuivre.fr.eu.org> <44FC3007.8050100@adacore.com> <20060904144131.GA37929@melamine.cuivre.fr.eu.org> <44FC492F.1080208@adacore.com> <20060925160045.GB90552@melamine.cuivre.fr.eu.org>
In-Reply-To: <20060925160045.GB90552@melamine.cuivre.fr.eu.org>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information  in GNAT Pro
Status: O
Content-Length: 1133
Lines: 39

Thomas Quinot wrote:
> * Robert Dewar, 2006-09-04 :
> 
>>>> But I think Richard is right, this should be a text file which we
>>>> parse, and should not look at all like a license key, which the
>>>> current header does not satisfy.
> 
> OK, since there was no new on this TN, 

That's odd, I did send some mail, must have disappeared, I see
that indeed it is not in comment file.

I prepared a parser that expects
> the following format for the installation parameters file:
> 
> ----------
> The following license to [use|evaluate] GNAT <buildtype-version>
> for <target-platform> [hosted on <host-platform>] is granted to
> <licensee> until <validity-date>.
> 
> <text from the GMGPL, or whatever license>
> 
> Signature: rrrr ssss

THat looks just perfect, better than what I had suggested anyway :-)
> ----------
> 
> Does that look better?
> 
> If so, I'll proceed with completing implementation of the checks and
> checking it required supporting code (inlcuding modular exponentiation
> for uints, which should go either as a child of uintp or directly within
> uintp, since it relies on Least_Sig_Digit).
> 
> Thomas.
> 




From quinot@gnat.com  Tue Sep 26 06:27:39 2006
To: file-ci@gnat.com
Message-Id: <20060926102738.DEB2648CBC1@nile.gnat.com>
Date: Tue, 26 Sep 2006 06:27:38 -0400 (EDT)
From: quinot@gnat.com (Thomas Quinot)
Subject: [ada] uintp.adb (1.89) checked in ada
Status: O
Content-Length: 110
Lines: 7


(UI_Modular_Exponentiation, UI_Modular_Inverse): New modular arithmetic
 functions for uint.
For F831-014




From quinot@gnat.com  Tue Sep 26 06:27:40 2006
To: file-ci@gnat.com
Message-Id: <20060926102739.5591648CCB7@nile.gnat.com>
Date: Tue, 26 Sep 2006 06:27:39 -0400 (EDT)
From: quinot@gnat.com (Thomas Quinot)
Subject: [ada] uintp.ads (1.70) checked in ada
Status: O
Content-Length: 110
Lines: 7


(UI_Modular_Exponentiation, UI_Modular_Inverse): New modular arithmetic
 functions for uint.
For F831-014




From quinot@malevil.act-europe.fr  Tue Sep 26 14:17:18 2006
Date: Tue, 26 Sep 2006 20:17:42 +0200 (CEST)
From: Thomas Quinot <quinot@malevil.act-europe.fr>
Message-Id: <200609261817.k8QIHgHI016860@malevil.act-europe.fr>
To: file@gnat.com
Subject: [F831-014] uintp.adb(1.90) RH change
Status: O
Content-Length: 110
Lines: 6

(UI_Modular_Inverse): Correct implementation.
----
date: 2006/09/26 18:17:39;  author: quinot;
For F831-014



From quinot@adacore.com  Tue Sep 26 14:28:41 2006
Date: Tue, 26 Sep 2006 20:28:37 +0200
From: Thomas Quinot <quinot@adacore.com>
To: Robert Dewar <dewar@adacore.com>
Cc: Richard Kenner <kenner@vlsi1.ultra.nyu.edu>, fofanov@adacore.com,
	report@adacore.com
Message-ID: <20060926182837.GA41748@melamine.cuivre.fr.eu.org>
References: <203231446.20060904133232@adacore.com> <10609041143.AA24473@vlsi1.ultra.nyu.edu> <20060904121657.GA32121@melamine.cuivre.fr.eu.org> <10609041231.AA24664@vlsi1.ultra.nyu.edu> <20060904123306.GA33102@melamine.cuivre.fr.eu.org> <44FC3007.8050100@adacore.com> <20060904144131.GA37929@melamine.cuivre.fr.eu.org> <44FC492F.1080208@adacore.com> <20060925160045.GB90552@melamine.cuivre.fr.eu.org> <4518165E.1020008@adacore.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <4518165E.1020008@adacore.com>
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information in GNAT Pro
Status: O
Content-Length: 1168
Lines: 28

* Robert Dewar, 2006-09-25 :

> >If so, I'll proceed with completing implementation of the checks and
> >checking it required supporting code (inlcuding modular exponentiation
> >for uints, which should go either as a child of uintp or directly within
> >uintp, since it relies on Least_Sig_Digit).

OK, this is mostly ready. The uintp code is now good (ugh, I had to
brush up my arithmetic, I had got Modular_Inverse wrong the first time
round...) The signature validation is OK as well, I checked it against a
signature generated by OpenSSL.

Now we need to remove the dependency upon Ada.Streams from GNAT.SHA1
so that it can be used in the compiler, and we also need support in
osint to tell us whether a given date is in the past or not.

I'm going to check in the new Instpar package (Installation Parameters
are the bits of information from the license file that we'll validate)
so that others can review it, but won't hook it into the build until
the various technical and procedural issues have all been addressed.

Thomas.

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From quinot@gnat.com  Tue Sep 26 14:33:33 2006
To: file-ci@gnat.com
Message-Id: <20060926183332.CE7DB48CC5B@nile.gnat.com>
Date: Tue, 26 Sep 2006 14:33:32 -0400 (EDT)
From: quinot@gnat.com (Thomas Quinot)
Subject: [ada] instpar.ads (1.1) checked in ada
Status: O
Content-Length: 87
Lines: 6


Instpar: New package providing validation of installation parameters.
For F831-014




From quinot@gnat.com  Tue Sep 26 14:33:37 2006
To: file-ci@gnat.com
Message-Id: <20060926183333.32D0C48CE49@nile.gnat.com>
Date: Tue, 26 Sep 2006 14:33:33 -0400 (EDT)
From: quinot@gnat.com (Thomas Quinot)
Subject: [ada] instpar.adb (1.1) checked in ada
Status: O
Content-Length: 87
Lines: 6


Instpar: New package providing validation of installation parameters.
For F831-014




From dismukes@gnat.com  Tue Sep 26 14:39:45 2006
To: quinot@adacore.com
Cc: dismukes@gnat.com, file@gnat.com
Message-Id: <20060926183944.A5B7F48CC5B@nile.gnat.com>
Date: Tue, 26 Sep 2006 14:39:44 -0400 (EDT)
From: dismukes@gnat.com (Gary Dismukes)
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information in GNAT Pro
Status: O
Content-Length: 669
Lines: 18

> OK, this is mostly ready. The uintp code is now good (ugh, I had to
> brush up my arithmetic, I had got Modular_Inverse wrong the first time
> round...)

Thomas,

I thought that first version didn't look quite right (I checked it for
3 mod 10 and came out with 3;-).  I was going to send you a note about that,
but I see you've fixed it.  One thing I'm wondering though is what the
behavior should be when there's not an inverse.  The spec comments don't
say anything about that.  If checks are on then you'll get an exception,
but if uintp is compiled with -gnatp I suppose the result isn't very
well defined.  What's the intended behavior in that case?

-- Gary




From quinot@adacore.com  Tue Sep 26 14:48:07 2006
Date: Tue, 26 Sep 2006 20:47:45 +0200
From: Thomas Quinot <quinot@adacore.com>
To: Gary Dismukes <dismukes@gnat.com>
Cc: file@gnat.com
Message-ID: <20060926184745.GA42708@melamine.cuivre.fr.eu.org>
References: <20060926183944.A5B7F48CC5B@nile.gnat.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <20060926183944.A5B7F48CC5B@nile.gnat.com>
Subject: Re: [F831-014] - #999 including GNAT Pro license and customer information in GNAT Pro
Status: O
Content-Length: 906
Lines: 21

* Gary Dismukes, 2006-09-26 :

> but I see you've fixed it.  One thing I'm wondering though is what the
> behavior should be when there's not an inverse.  The spec comments don't
> say anything about that.  If checks are on then you'll get an exception,
> but if uintp is compiled with -gnatp I suppose the result isn't very
> well defined.  What's the intended behavior in that case?

Um, I honestly haven't given a thought to that case. This code is not
actually intended to be used with a non-prime modulus, and undefined
behaviour would be perfectly fine (but we definitely are missing a note
of this assumption in the documentation, which I'll fix immediately).

As far as the current implementation goes, we'll get a nice division by
0 for anything not inversible.

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From quinot@gnat.com  Tue Sep 26 14:48:31 2006
To: file-ci@gnat.com
Message-Id: <20060926184830.DDC4048CD93@nile.gnat.com>
Date: Tue, 26 Sep 2006 14:48:30 -0400 (EDT)
From: quinot@gnat.com (Thomas Quinot)
Subject: [ada] uintp.ads (1.71) checked in ada
Status: O
Content-Length: 136
Lines: 7


(UI_Modular_Inverse): Add a note that the behaviour of this subprogram is
 undefined if the given n is not inversible.
For F831-014




From quinot@gnat.com  Wed Sep 27 07:21:32 2006
To: file-ci@gnat.com
Message-Id: <20060927112131.0AA3A48CEC8@nile.gnat.com>
Date: Wed, 27 Sep 2006 07:21:31 -0400 (EDT)
From: quinot@gnat.com (Thomas Quinot)
Subject: [ada] instpar.adb (1.2) checked in ada
Status: O
Content-Length: 258
Lines: 12


(Parse): Make target platform indication optional
(Check_Installation_Parameters): Do not check target platform ifr
 not specified.
 Check parameters expiration date against system clock.

For F831-014

This file is not to be propagated to the FSF repo.




From dewar@adacore.com  Thu Sep 28 06:31:59 2006
Message-ID: <451BA4A1.7020801@adacore.com>
Date: Thu, 28 Sep 2006 06:32:01 -0400
From: Robert Dewar <dewar@adacore.com>
MIME-Version: 1.0
To: Thomas Quinot <quinot@adacore.com>
CC: Robert Dewar <dewar@gnat.com>, Report <report@adacore.com>
References: <20060928102317.304DD48CF77@nile.gnat.com> <20060928102626.GE17834@melamine.cuivre.fr.eu.org>
In-Reply-To: <20060928102626.GE17834@melamine.cuivre.fr.eu.org>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 853
Lines: 24

Thomas Quinot wrote:
> * Robert Dewar, 2006-09-28 :
> 
>> great progress here!
> 
> Thanks!
> 
>> I did a fair amount of minor reformatting, pls double
>> check I did not screw something up.
> 
> Sure, will check. I'm also working at cleaning this up, and also
> improving uintp a bit so that the signature verification does not have a
> significant impact on compiler execution time.

My goodness, we don't need ANY kind of high security in the signature
check, a very simple checksum is perfectly good enough here. The idea
that this could be taking significant time is a bit appalling :-).

Actually I would prefer to keep DSA out of the picture, it seems a bit
provocative if you ask me, and instead use some very simple algorithm,
just a shift/xor loop would be fine I think. There is really no point
in high security here. What do others think?



From berrendo@adacore.com  Thu Sep 28 06:36:11 2006
Message-ID: <451BA597.7080206@adacore.com>
Date: Thu, 28 Sep 2006 12:36:07 +0200
From: Romain Berrendonner <berrendo@adacore.com>
MIME-Version: 1.0
To: Robert Dewar <dewar@adacore.com>
CC: Thomas Quinot <quinot@adacore.com>, Robert Dewar <dewar@gnat.com>, 
 Report <report@adacore.com>
References: <20060928102317.304DD48CF77@nile.gnat.com> <20060928102626.GE17834@melamine.cuivre.fr.eu.org> <451BA4A1.7020801@adacore.com>
In-Reply-To: <451BA4A1.7020801@adacore.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 689
Lines: 18

Robert Dewar a crit :
> Actually I would prefer to keep DSA out of the picture, it seems a bit
> provocative if you ask me, and instead use some very simple algorithm,
> just a shift/xor loop would be fine I think. There is really no point
> in high security here. What do others think?

As the work has been done, I don't see any problem having DSA signature 
rather than MD5 of content + some password or whatever other scheme we 
come up with.

If you are concerned by opposition, just remove the word "signature:" 
from the licence file. That way, it will just look like a random string 
of weird digits rather than a "DSA, cryptographically-secure, 
industry-proven, signature".





From quinot@adacore.com  Thu Sep 28 06:46:10 2006
Date: Thu, 28 Sep 2006 12:46:07 +0200
From: Thomas Quinot <quinot@adacore.com>
To: Robert Dewar <dewar@adacore.com>
Cc: Robert Dewar <dewar@gnat.com>, Report <report@adacore.com>
Message-ID: <20060928104607.GF17834@melamine.cuivre.fr.eu.org>
References: <20060928102317.304DD48CF77@nile.gnat.com> <20060928102626.GE17834@melamine.cuivre.fr.eu.org> <451BA4A1.7020801@adacore.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <451BA4A1.7020801@adacore.com>
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 867
Lines: 21

* Robert Dewar, 2006-09-28 :

> Actually I would prefer to keep DSA out of the picture, it seems a bit
> provocative if you ask me, and instead use some very simple algorithm,
> just a shift/xor loop would be fine I think. There is really no point
> in high security here. What do others think?

The requirement we wanted to implement is: bypassing the license file
enforcement must be at least as difficult as rebuilding the compiler.

If we just use a hash of the license file + some password (which then
needs to be hardcoded in source code so that we can verify it), then a
user who wants to bypass the protection has all the necessary
information to generate a new license file himself, without having to
rebuild the compiler.

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From dewar@adacore.com  Thu Sep 28 07:33:59 2006
Message-ID: <451BB32B.1080509@adacore.com>
Date: Thu, 28 Sep 2006 07:34:03 -0400
From: Robert Dewar <dewar@adacore.com>
MIME-Version: 1.0
To: Romain Berrendonner <berrendo@adacore.com>
CC: Thomas Quinot <quinot@adacore.com>, Robert Dewar <dewar@gnat.com>, 
 Report <report@adacore.com>
References: <20060928102317.304DD48CF77@nile.gnat.com> <20060928102626.GE17834@melamine.cuivre.fr.eu.org> <451BA4A1.7020801@adacore.com> <451BA597.7080206@adacore.com>
In-Reply-To: <451BA597.7080206@adacore.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 381
Lines: 14

Romain Berrendonner wrote:

> As the work has been done, I don't see any problem having DSA signature 
> rather than MD5 of content + some password or whatever other scheme we 
> come up with.

I would regard MD5 as overkill as well
> 
> If you are concerned by opposition, just remove the word "signature:" 
> from the licence file.

Yes, I think that's a good idea in any case



From quinot@gnat.com  Thu Sep 28 07:34:08 2006
To: file-ci@gnat.com
Message-Id: <20060928113407.63CAB48CDE1@nile.gnat.com>
Date: Thu, 28 Sep 2006 07:34:07 -0400 (EDT)
From: quinot@gnat.com (Thomas Quinot)
Subject: [ada] instpar.ads (1.4) checked in ada
Status: O
Content-Length: 27
Lines: 6


Add comments.
F831-014




From quinot@gnat.com  Thu Sep 28 07:35:13 2006
To: file-ci@gnat.com
Message-Id: <20060928113512.BD14048CBC6@nile.gnat.com>
Date: Thu, 28 Sep 2006 07:35:12 -0400 (EDT)
From: quinot@gnat.com (Thomas Quinot)
Subject: [ada] instpar.adb (1.3) checked in ada
Status: O
Content-Length: 82
Lines: 6


Minor reformatting (use uppercase identifiers, for consistency).
For F831-014




From dewar@adacore.com  Thu Sep 28 07:35:31 2006
Message-ID: <451BB386.2090606@adacore.com>
Date: Thu, 28 Sep 2006 07:35:34 -0400
From: Robert Dewar <dewar@adacore.com>
MIME-Version: 1.0
To: Thomas Quinot <quinot@adacore.com>
CC: Robert Dewar <dewar@gnat.com>, Report <report@adacore.com>
References: <20060928102317.304DD48CF77@nile.gnat.com> <20060928102626.GE17834@melamine.cuivre.fr.eu.org> <451BA4A1.7020801@adacore.com> <20060928104607.GF17834@melamine.cuivre.fr.eu.org>
In-Reply-To: <20060928104607.GF17834@melamine.cuivre.fr.eu.org>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 1027
Lines: 25

Thomas Quinot wrote:
> * Robert Dewar, 2006-09-28 :
> 
>> Actually I would prefer to keep DSA out of the picture, it seems a bit
>> provocative if you ask me, and instead use some very simple algorithm,
>> just a shift/xor loop would be fine I think. There is really no point
>> in high security here. What do others think?
> 
> The requirement we wanted to implement is: bypassing the license file
> enforcement must be at least as difficult as rebuilding the compiler.

Impossible to achieve, because very rapidly, someone will post license
files which can be obtained by people wanting to break things. And why
should people rebuild the compiler, they can see the algorithm we use
no matter how complex it is.
> 
> If we just use a hash of the license file + some password (which then
> needs to be hardcoded in source code so that we can verify it), then a
> user who wants to bypass the protection has all the necessary
> information to generate a new license file himself, without having to
> rebuild the compiler.
> 




From quinot@adacore.com  Thu Sep 28 07:41:28 2006
Date: Thu, 28 Sep 2006 13:41:26 +0200
From: Thomas Quinot <quinot@adacore.com>
To: Robert Dewar <dewar@adacore.com>
Cc: Report <report@adacore.com>
Message-ID: <20060928114126.GA21600@melamine.cuivre.fr.eu.org>
References: <20060928102317.304DD48CF77@nile.gnat.com> <20060928102626.GE17834@melamine.cuivre.fr.eu.org> <451BA4A1.7020801@adacore.com> <20060928104607.GF17834@melamine.cuivre.fr.eu.org> <451BB386.2090606@adacore.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <451BB386.2090606@adacore.com>
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 1538
Lines: 34

* Robert Dewar, 2006-09-28 :

> >The requirement we wanted to implement is: bypassing the license file
> >enforcement must be at least as difficult as rebuilding the compiler.
> Impossible to achieve, because very rapidly, someone will post license
> files which can be obtained by people wanting to break things.

Sure, we may assume that some people may share license files we send
them; however I doubt our customers would do that (just as we work under
the assumption that they are usually not redistributing GNAT Pro), and
as far as evals are concerned, the only license files they could share
would be the limited-time files we issued. The point in using a public
key signature is that they cannot *make their own* license files with
parameters of their choosing (e.g. with a different expiration date).

> And why should people rebuild the compiler, they can see the algorithm we use
> no matter how complex it is.

This is exactly the point of using a public-key signature scheme: knowledge
of the *verification* algorithm and keys (which everyone can see in
source code) is not sufficient to *generate* a new signature (because
generating the signature involves the private part of the key, while
checking it only requires the public part).

(OK, all this is based on the assumption that solving the discrete log
problem in reasonable time is harder than rebuilding the compiler :-) )

Thomas.

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From dewar@adacore.com  Thu Sep 28 07:49:54 2006
Message-ID: <451BB6E6.6030607@adacore.com>
Date: Thu, 28 Sep 2006 07:49:58 -0400
From: Robert Dewar <dewar@adacore.com>
MIME-Version: 1.0
To: Thomas Quinot <quinot@adacore.com>
CC: Report <report@adacore.com>, Franco Gasperoni <gasperoni@adacore.com>
References: <20060928102317.304DD48CF77@nile.gnat.com> <20060928102626.GE17834@melamine.cuivre.fr.eu.org> <451BA4A1.7020801@adacore.com> <20060928104607.GF17834@melamine.cuivre.fr.eu.org> <451BB386.2090606@adacore.com> <20060928114126.GA21600@melamine.cuivre.fr.eu.org>
In-Reply-To: <20060928114126.GA21600@melamine.cuivre.fr.eu.org>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 456
Lines: 10

I really do not like having the proper operation of the compiler
depend on a private key that only we know and that we do not
disclose in the sources. In fact I find this a violation of
GPL, at least in spirit and for sure it is a violation of GPL 3.
I think we are going too far here and treading on dangerous
ground. All we want to prevent is simple casual bypassing.
If someone wants to bypass it, they should be able to do so
without any big hassle.



From quinot@gnat.com  Thu Sep 28 08:13:14 2006
To: file-ci@gnat.com
Message-Id: <20060928121312.ECC2E48CBDD@nile.gnat.com>
Date: Thu, 28 Sep 2006 08:13:12 -0400 (EDT)
From: quinot@gnat.com (Thomas Quinot)
Subject: [ada] uintp.adb (1.93) checked in ada
Status: O
Content-Length: 741
Lines: 19


(UI_Div_Rem): New subprogram, extending previous implementation of UI_Div.
(UI_Div): Reimplement as a call to UI_Div_Rem.

(UI_Rem): Take advantage of the fact that UI_Div_Rem provides the remainder,
 avoiding the cost of a multiplication and a subtraction.
(UI_Modular_Inverse): Take advantage of the fact that UI_Div_Rem provides
 both quotient and remainder in a single computation.

Done while working on F831-014

This is a performance optimization (taking advantage of the fact that the
same division algorithm can compute both the quotient and the remainder
in parallel), with no testable behaviour change. It also fixes a memory
leak in UI_Rem where intermediate results in the fallback general case
were not properly discarded.




From quinot@gnat.com  Thu Sep 28 08:30:32 2006
To: file-ci@gnat.com
Message-Id: <20060928123032.197E348CDE4@nile.gnat.com>
Date: Thu, 28 Sep 2006 08:30:32 -0400 (EDT)
From: quinot@gnat.com (Thomas Quinot)
Subject: [ada] instpar.adb (1.4) checked in ada
Status: O
Content-Length: 159
Lines: 7


(Parse): Modify to not expect a "Signature:" keyword before the signature.
 The signature is just taken to be the last two words in the file.
For F831-014




From quinot@adacore.com  Thu Sep 28 08:30:39 2006
Date: Thu, 28 Sep 2006 14:30:35 +0200
From: Thomas Quinot <quinot@adacore.com>
To: Romain Berrendonner <berrendo@adacore.com>
Cc: Robert Dewar <dewar@adacore.com>, Robert Dewar <dewar@gnat.com>,
	Report <report@adacore.com>
Message-ID: <20060928123035.GA23095@melamine.cuivre.fr.eu.org>
References: <20060928102317.304DD48CF77@nile.gnat.com> <20060928102626.GE17834@melamine.cuivre.fr.eu.org> <451BA4A1.7020801@adacore.com> <451BA597.7080206@adacore.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <451BA597.7080206@adacore.com>
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 442
Lines: 14

* Romain Berrendonner, 2006-09-28 :

> If you are concerned by opposition, just remove the word "signature:" 
> from the licence file. That way, it will just look like a random string 
> of weird digits rather than a "DSA, cryptographically-secure, 
> industry-proven, signature".

This is now implemented.

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From kenner@vlsi1.ultra.nyu.edu  Thu Sep 28 08:38:53 2006
From: kenner@vlsi1.ultra.nyu.edu (Richard Kenner)
Message-Id: <10609281242.AA10030@vlsi1.ultra.nyu.edu>
Date: Thu, 28 Sep 2006 08:42:13 EDT
To: dewar@adacore.com
Cc: dewar@gnat.com, quinot@adacore.com, report@adacore.com
In-Reply-To: <451BA4A1.7020801@adacore.com>
References: <20060928102317.304DD48CF77@nile.gnat.com>
    <20060928102626.GE17834@melamine.cuivre.fr.eu.org>
    <451BA4A1.7020801@adacore.com>
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 460
Lines: 10

> Actually I would prefer to keep DSA out of the picture, it seems a bit
> provocative if you ask me, and instead use some very simple algorithm,
> just a shift/xor loop would be fine I think. There is really no point
> in high security here. What do others think?

I agree with you.  The *most* I could see doing is MD5.  But even a
simple adding of all the values mod 32 is good enough.  After all, the
check digit on airplane ticket number is just mod 7!



From kenner@vlsi1.ultra.nyu.edu  Thu Sep 28 08:40:35 2006
From: kenner@vlsi1.ultra.nyu.edu (Richard Kenner)
Message-Id: <10609281243.AA10078@vlsi1.ultra.nyu.edu>
Date: Thu, 28 Sep 2006 08:43:53 EDT
To: berrendo@adacore.com
Cc: dewar@adacore.com, dewar@gnat.com, quinot@adacore.com, report@adacore.com
In-Reply-To: <451BA597.7080206@adacore.com>
References: <20060928102317.304DD48CF77@nile.gnat.com>
    <20060928102626.GE17834@melamine.cuivre.fr.eu.org>
    <451BA4A1.7020801@adacore.com> <451BA597.7080206@adacore.com>
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 374
Lines: 9

> As the work has been done, I don't see any problem having DSA signature 
> rather than MD5 of content + some password or whatever other scheme we 
> come up with.

I do see a problem.  I really don't want to have to worry about export
laws and cryptography.  The intent here was never for anything that was
remotely "secure", just preventing completely obvious mucking.



From kenner@vlsi1.ultra.nyu.edu  Thu Sep 28 08:43:33 2006
From: kenner@vlsi1.ultra.nyu.edu (Richard Kenner)
Message-Id: <10609281246.AA10286@vlsi1.ultra.nyu.edu>
Date: Thu, 28 Sep 2006 08:46:53 EDT
To: dewar@adacore.com
Cc: gasperoni@adacore.com, quinot@adacore.com, report@adacore.com
In-Reply-To: <451BB6E6.6030607@adacore.com>
References: <20060928102317.304DD48CF77@nile.gnat.com>
    <20060928102626.GE17834@melamine.cuivre.fr.eu.org>
    <451BA4A1.7020801@adacore.com>
    <20060928104607.GF17834@melamine.cuivre.fr.eu.org>
    <451BB386.2090606@adacore.com>
    <20060928114126.GA21600@melamine.cuivre.fr.eu.org>
    <451BB6E6.6030607@adacore.com>
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 649
Lines: 14

> I really do not like having the proper operation of the compiler
> depend on a private key that only we know and that we do not
> disclose in the sources. In fact I find this a violation of
> GPL, at least in spirit and for sure it is a violation of GPL 3.
> I think we are going too far here and treading on dangerous
> ground. All we want to prevent is simple casual bypassing.
> If someone wants to bypass it, they should be able to do so
> without any big hassle.

I strongly agree.  Just break the license file into four-character
segment, add them all up mod 2**32, and output the result.  That's more
than secure enough for our purposes.



From quinot@adacore.com  Thu Sep 28 08:49:03 2006
Date: Thu, 28 Sep 2006 14:49:01 +0200
From: Thomas Quinot <quinot@adacore.com>
To: Richard Kenner <kenner@vlsi1.ultra.nyu.edu>
Cc: report@adacore.com
Message-ID: <20060928124901.GB23095@melamine.cuivre.fr.eu.org>
References: <20060928102317.304DD48CF77@nile.gnat.com> <20060928102626.GE17834@melamine.cuivre.fr.eu.org> <451BA4A1.7020801@adacore.com> <451BA597.7080206@adacore.com> <10609281243.AA10078@vlsi1.ultra.nyu.edu>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <10609281243.AA10078@vlsi1.ultra.nyu.edu>
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 966
Lines: 26

* Richard Kenner, 2006-09-28 :

> laws and cryptography.  The intent here was never for anything that was
> remotely "secure", just preventing completely obvious mucking.

That's not the requirement that was expressed by the sales team as the
initial motivation for this TN. I would suggest to involve Franco in
this discussion.

I also seem to recall a warning expressed by Robert that we *had to*
make the protection sufficiently non-trivial that it is actually
considered a protection mechanism, and that a simple checksum was
definitely not appropriate for that.

(Also, please note that what is being implemented at this point is
precisely the spec that was posted on this TN 4 weeks ago, and that the
only comments that were made at that point were purely cosmetic
considerations on the format of the license file).

Thomas.

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From comar@adacore.com  Thu Sep 28 08:53:18 2006
Message-ID: <451BC5BC.9040106@adacore.com>
Date: Thu, 28 Sep 2006 14:53:16 +0200
From: Cyrille Comar <comar@adacore.com>
MIME-Version: 1.0
To: Robert Dewar <dewar@adacore.com>
CC: Thomas Quinot <quinot@adacore.com>, Report <report@adacore.com>, 
 Franco Gasperoni <gasperoni@adacore.com>,
 Friess Michael <friess@adacore.com>
References: <20060928102317.304DD48CF77@nile.gnat.com> <20060928102626.GE17834@melamine.cuivre.fr.eu.org> <451BA4A1.7020801@adacore.com> <20060928104607.GF17834@melamine.cuivre.fr.eu.org> <451BB386.2090606@adacore.com> <20060928114126.GA21600@melamine.cuivre.fr.eu.org> <451BB6E6.6030607@adacore.com>
In-Reply-To: <451BB6E6.6030607@adacore.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 890
Lines: 19

Robert Dewar a crit :
> I really do not like having the proper operation of the compiler
> depend on a private key that only we know and that we do not
> disclose in the sources. In fact I find this a violation of
> GPL, at least in spirit and for sure it is a violation of GPL 3.
> I think we are going too far here and treading on dangerous
> ground.

I, too, am not at ease with this approach... I managed to get convinced 
that we needed to deal with the eval problem in Europe but the current 
proposed approach is potentially too visible to standard GNAT Pro 
customers and I start to have second thoughts about the whole thing 
wondering if we are not generating more problems than solutions...

Maybe we should revisit the decision to put this technology in all GNAT 
Pro technology and use it (or a modified version of it) only a specific 
compilers only used for evaluation...



From berrendo@adacore.com  Thu Sep 28 08:53:32 2006
Message-ID: <451BC5C8.8020405@adacore.com>
Date: Thu, 28 Sep 2006 14:53:28 +0200
From: Romain Berrendonner <berrendo@adacore.com>
MIME-Version: 1.0
To: Thomas Quinot <quinot@adacore.com>
CC: Richard Kenner <kenner@vlsi1.ultra.nyu.edu>,  report@adacore.com
References: <20060928102317.304DD48CF77@nile.gnat.com> <20060928102626.GE17834@melamine.cuivre.fr.eu.org> <451BA4A1.7020801@adacore.com> <451BA597.7080206@adacore.com> <10609281243.AA10078@vlsi1.ultra.nyu.edu> <20060928124901.GB23095@melamine.cuivre.fr.eu.org>
In-Reply-To: <20060928124901.GB23095@melamine.cuivre.fr.eu.org>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 613
Lines: 17

Thomas Quinot a crit :
> * Richard Kenner, 2006-09-28 :
> 
> (Also, please note that what is being implemented at this point is
> precisely the spec that was posted on this TN 4 weeks ago, and that the
> only comments that were made at that point were purely cosmetic
> considerations on the format of the license file).

In addition, I recall discussing very similar concerns in a meeting 
where these issues have been discussed, in particular 1) the GPL 
compatibility of such code and 2) the confusing message that we could 
end up sending to customers with what looks like a locking mechanism.

-- 
Romain



From dewar@adacore.com  Thu Sep 28 10:59:48 2006
Message-ID: <451BE368.5040005@adacore.com>
Date: Thu, 28 Sep 2006 10:59:52 -0400
From: Robert Dewar <dewar@adacore.com>
MIME-Version: 1.0
To: Richard Kenner <kenner@vlsi1.ultra.nyu.edu>
CC:  berrendo@adacore.com,  dewar@gnat.com,  quinot@adacore.com, 
 report@adacore.com
References: <20060928102317.304DD48CF77@nile.gnat.com>    <20060928102626.GE17834@melamine.cuivre.fr.eu.org>    <451BA4A1.7020801@adacore.com> <451BA597.7080206@adacore.com> <10609281243.AA10078@vlsi1.ultra.nyu.edu>
In-Reply-To: <10609281243.AA10078@vlsi1.ultra.nyu.edu>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 473
Lines: 12

Richard Kenner wrote:
>> As the work has been done, I don't see any problem having DSA signature 
>> rather than MD5 of content + some password or whatever other scheme we 
>> come up with.
> 
> I do see a problem.  I really don't want to have to worry about export
> laws and cryptography.  The intent here was never for anything that was
> remotely "secure", just preventing completely obvious mucking.

Ouch yes! the export restrictions are a real problem here I fear



From dewar@adacore.com  Thu Sep 28 11:05:07 2006
Message-ID: <451BE4A6.7050807@adacore.com>
Date: Thu, 28 Sep 2006 11:05:10 -0400
From: Robert Dewar <dewar@adacore.com>
MIME-Version: 1.0
To: Thomas Quinot <quinot@adacore.com>
CC: Richard Kenner <kenner@vlsi1.ultra.nyu.edu>,  report@adacore.com
References: <20060928102317.304DD48CF77@nile.gnat.com> <20060928102626.GE17834@melamine.cuivre.fr.eu.org> <451BA4A1.7020801@adacore.com> <451BA597.7080206@adacore.com> <10609281243.AA10078@vlsi1.ultra.nyu.edu> <20060928124901.GB23095@melamine.cuivre.fr.eu.org>
In-Reply-To: <20060928124901.GB23095@melamine.cuivre.fr.eu.org>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 1131
Lines: 26

Thomas Quinot wrote:

> I also seem to recall a warning expressed by Robert that we *had to*
> make the protection sufficiently non-trivial that it is actually
> considered a protection mechanism, and that a simple checksum was
> definitely not appropriate for that.

That was the DMCA issue, but in practice ANYTHING is secure enough
to be considered a protection mechanism (someone even decided that
putting an autorun file on a CD was a protection mechanism, and that
advising people to press the shift key while loading the CD was a
DMCA violation -- this did not come to court, but because of bad
publicity reasons, not legal reasons, we don't know what would
happen in court in that case.
> 
> (Also, please note that what is being implemented at this point is
> precisely the spec that was posted on this TN 4 weeks ago, and that the
> only comments that were made at that point were purely cosmetic
> considerations on the format of the license file).

Well the spec simply mentioned a DSA signature, and I just did not
recognize the acronym at the time, as always you cannot assume that
lack of a comment means assent.




From dewar@adacore.com  Thu Sep 28 11:06:35 2006
Message-ID: <451BE4FE.5040502@adacore.com>
Date: Thu, 28 Sep 2006 11:06:38 -0400
From: Robert Dewar <dewar@adacore.com>
MIME-Version: 1.0
To: Cyrille Comar <comar@adacore.com>
CC: Thomas Quinot <quinot@adacore.com>, Report <report@adacore.com>, 
 Franco Gasperoni <gasperoni@adacore.com>,
 Friess Michael <friess@adacore.com>
References: <20060928102317.304DD48CF77@nile.gnat.com> <20060928102626.GE17834@melamine.cuivre.fr.eu.org> <451BA4A1.7020801@adacore.com> <20060928104607.GF17834@melamine.cuivre.fr.eu.org> <451BB386.2090606@adacore.com> <20060928114126.GA21600@melamine.cuivre.fr.eu.org> <451BB6E6.6030607@adacore.com> <451BC5BC.9040106@adacore.com>
In-Reply-To: <451BC5BC.9040106@adacore.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 863
Lines: 21

Cyrille Comar wrote:

> I, too, am not at ease with this approach... I managed to get convinced 
> that we needed to deal with the eval problem in Europe but the current 
> proposed approach is potentially too visible to standard GNAT Pro 
> customers and I start to have second thoughts about the whole thing 
> wondering if we are not generating more problems than solutions...
> 
> Maybe we should revisit the decision to put this technology in all GNAT 
> Pro technology and use it (or a modified version of it) only a specific 
> compilers only used for evaluation...

Well I am comfortable with that decision, and see no basis for
revisiting it (I know Cyrille was never very comfortable with
this decision, but I still don't see that anything has changed
that would warrant revisiting it).

But I do think we should avoid cryptographic type signatures.




From quinot@adacore.com  Thu Sep 28 11:07:55 2006
Date: Thu, 28 Sep 2006 17:07:52 +0200
From: Thomas Quinot <quinot@adacore.com>
To: Robert Dewar <dewar@adacore.com>
Cc: Richard Kenner <kenner@vlsi1.ultra.nyu.edu>, berrendo@adacore.com,
	dewar@gnat.com, report@adacore.com
Message-ID: <20060928150752.GB28880@melamine.cuivre.fr.eu.org>
References: <20060928102317.304DD48CF77@nile.gnat.com> <20060928102626.GE17834@melamine.cuivre.fr.eu.org> <451BA4A1.7020801@adacore.com> <451BA597.7080206@adacore.com> <10609281243.AA10078@vlsi1.ultra.nyu.edu> <451BE368.5040005@adacore.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <451BE368.5040005@adacore.com>
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 286
Lines: 11

* Robert Dewar, 2006-09-28 :

> Ouch yes! the export restrictions are a real problem here I fear

Even for a signature-only (no encryption) algorithm?

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From berrendo@adacore.com  Thu Sep 28 11:08:55 2006
Message-ID: <451BE582.50503@adacore.com>
Date: Thu, 28 Sep 2006 17:08:50 +0200
From: Romain Berrendonner <berrendo@adacore.com>
MIME-Version: 1.0
To: Robert Dewar <dewar@adacore.com>
CC: Richard Kenner <kenner@vlsi1.ultra.nyu.edu>,  dewar@gnat.com, 
 quinot@adacore.com,  report@adacore.com
References: <20060928102317.304DD48CF77@nile.gnat.com>    <20060928102626.GE17834@melamine.cuivre.fr.eu.org>    <451BA4A1.7020801@adacore.com> <451BA597.7080206@adacore.com> <10609281243.AA10078@vlsi1.ultra.nyu.edu> <451BE368.5040005@adacore.com>
In-Reply-To: <451BE368.5040005@adacore.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 607
Lines: 16

Robert Dewar a crit :
> Richard Kenner wrote:
>>> As the work has been done, I don't see any problem having DSA 
>>> signature rather than MD5 of content + some password or whatever 
>>> other scheme we come up with.
>>
>> I do see a problem.  I really don't want to have to worry about export
>> laws and cryptography.  The intent here was never for anything that was
>> remotely "secure", just preventing completely obvious mucking.
> 
> Ouch yes! the export restrictions are a real problem here I fear

Are you sure of that ? I thought that export regulations had been widely 
relaxed in 96 or 98 ...



From quinot@adacore.com  Thu Sep 28 11:12:46 2006
Date: Thu, 28 Sep 2006 17:12:44 +0200
From: Thomas Quinot <quinot@adacore.com>
To: Robert Dewar <dewar@adacore.com>
Cc: Cyrille Comar <comar@adacore.com>, Report <report@adacore.com>,
	Franco Gasperoni <gasperoni@adacore.com>,
	Friess Michael <friess@adacore.com>
Message-ID: <20060928151244.GC28880@melamine.cuivre.fr.eu.org>
References: <20060928102317.304DD48CF77@nile.gnat.com> <20060928102626.GE17834@melamine.cuivre.fr.eu.org> <451BA4A1.7020801@adacore.com> <20060928104607.GF17834@melamine.cuivre.fr.eu.org> <451BB386.2090606@adacore.com> <20060928114126.GA21600@melamine.cuivre.fr.eu.org> <451BB6E6.6030607@adacore.com> <451BC5BC.9040106@adacore.com> <451BE4FE.5040502@adacore.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <451BE4FE.5040502@adacore.com>
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 511
Lines: 16

* Robert Dewar, 2006-09-28 :

> But I do think we should avoid cryptographic type signatures.

Question is, is it acceptable to us that anyone can trivially bypass the
protection mechanism by looking at compiler source code and regenerating
a new license file, not having to rebuild the compiler?

It is not clear to me that we have a clear consensus on this point
currently.

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From berrendo@adacore.com  Thu Sep 28 11:25:44 2006
Message-ID: <451BE973.8080502@adacore.com>
Date: Thu, 28 Sep 2006 17:25:39 +0200
From: Romain Berrendonner <berrendo@adacore.com>
MIME-Version: 1.0
To: Romain Berrendonner <berrendo@adacore.com>
CC: Robert Dewar <dewar@adacore.com>, 
 Richard Kenner <kenner@vlsi1.ultra.nyu.edu>,
  dewar@gnat.com,  quinot@adacore.com,  report@adacore.com
References: <20060928102317.304DD48CF77@nile.gnat.com>    <20060928102626.GE17834@melamine.cuivre.fr.eu.org>    <451BA4A1.7020801@adacore.com> <451BA597.7080206@adacore.com> <10609281243.AA10078@vlsi1.ultra.nyu.edu> <451BE368.5040005@adacore.com> <451BE582.50503@adacore.com>
In-Reply-To: <451BE582.50503@adacore.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 713
Lines: 18

Romain Berrendonner a crit :
> Are you sure of that ? I thought that export regulations had been widely 
> relaxed in 96 or 98 ...

As a matter of fact, the Wassenaar Arrangement - to which the US are 
part - allows the use of cryptography where "the cryptographic 
capability is not-user accessible and which is specially designed and 
limited to allow any of the following: 1) Execution of copy-protected 
software" (section 5.A.2.a.9, in 
http://www.wassenaar.org/controllists/WA-LIST%20(05)%201%20Corr..pdf)

However I don't know if the US fully comply with this requirement. I 
would not bet on that ;-)

(DISCLAIMER: do NOT consider this as a supporting statement for 
introducing DRM in our technology)



From dewar@adacore.com  Thu Sep 28 11:55:56 2006
Message-ID: <451BF091.2010100@adacore.com>
Date: Thu, 28 Sep 2006 11:56:01 -0400
From: Robert Dewar <dewar@adacore.com>
MIME-Version: 1.0
To: Thomas Quinot <quinot@adacore.com>
CC: Richard Kenner <kenner@vlsi1.ultra.nyu.edu>,  berrendo@adacore.com, 
 dewar@gnat.com,  report@adacore.com
References: <20060928102317.304DD48CF77@nile.gnat.com> <20060928102626.GE17834@melamine.cuivre.fr.eu.org> <451BA4A1.7020801@adacore.com> <451BA597.7080206@adacore.com> <10609281243.AA10078@vlsi1.ultra.nyu.edu> <451BE368.5040005@adacore.com> <20060928150752.GB28880@melamine.cuivre.fr.eu.org>
In-Reply-To: <20060928150752.GB28880@melamine.cuivre.fr.eu.org>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 265
Lines: 13

Thomas Quinot wrote:
> * Robert Dewar, 2006-09-28 :
> 
>> Ouch yes! the export restrictions are a real problem here I fear
> 
> Even for a signature-only (no encryption) algorithm?

It's the algorithms that cane be troublesome, but I am not sure
of this ....
> 




From dewar@adacore.com  Thu Sep 28 11:59:15 2006
Message-ID: <451BF158.1090809@adacore.com>
Date: Thu, 28 Sep 2006 11:59:20 -0400
From: Robert Dewar <dewar@adacore.com>
MIME-Version: 1.0
To: Thomas Quinot <quinot@adacore.com>
CC: Cyrille Comar <comar@adacore.com>, Report <report@adacore.com>, 
 Franco Gasperoni <gasperoni@adacore.com>,
 Friess Michael <friess@adacore.com>
References: <20060928102317.304DD48CF77@nile.gnat.com> <20060928102626.GE17834@melamine.cuivre.fr.eu.org> <451BA4A1.7020801@adacore.com> <20060928104607.GF17834@melamine.cuivre.fr.eu.org> <451BB386.2090606@adacore.com> <20060928114126.GA21600@melamine.cuivre.fr.eu.org> <451BB6E6.6030607@adacore.com> <451BC5BC.9040106@adacore.com> <451BE4FE.5040502@adacore.com> <20060928151244.GC28880@melamine.cuivre.fr.eu.org>
In-Reply-To: <20060928151244.GC28880@melamine.cuivre.fr.eu.org>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 452
Lines: 14

Thomas Quinot wrote:
> * Robert Dewar, 2006-09-28 :
> 
>> But I do think we should avoid cryptographic type signatures.
> 
> Question is, is it acceptable to us that anyone can trivially bypass the
> protection mechanism by looking at compiler source code and regenerating
> a new license file, not having to rebuild the compiler?


I would say it's fine, and don't be so quick to use trivial. Finding
this in the compiler sources is not so easy ...



From dewar@adacore.com  Thu Sep 28 11:59:50 2006
Message-ID: <451BF17B.7020708@adacore.com>
Date: Thu, 28 Sep 2006 11:59:55 -0400
From: Robert Dewar <dewar@adacore.com>
MIME-Version: 1.0
To: Thomas Quinot <quinot@adacore.com>
CC: Cyrille Comar <comar@adacore.com>, Report <report@adacore.com>, 
 Franco Gasperoni <gasperoni@adacore.com>,
 Friess Michael <friess@adacore.com>
References: <20060928102317.304DD48CF77@nile.gnat.com> <20060928102626.GE17834@melamine.cuivre.fr.eu.org> <451BA4A1.7020801@adacore.com> <20060928104607.GF17834@melamine.cuivre.fr.eu.org> <451BB386.2090606@adacore.com> <20060928114126.GA21600@melamine.cuivre.fr.eu.org> <451BB6E6.6030607@adacore.com> <451BC5BC.9040106@adacore.com> <451BE4FE.5040502@adacore.com> <20060928151244.GC28880@melamine.cuivre.fr.eu.org>
In-Reply-To: <20060928151244.GC28880@melamine.cuivre.fr.eu.org>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 567
Lines: 18

Thomas Quinot wrote:
> * Robert Dewar, 2006-09-28 :
> 
>> But I do think we should avoid cryptographic type signatures.
> 
> Question is, is it acceptable to us that anyone can trivially bypass the
> protection mechanism by looking at compiler source code and regenerating
> a new license file, not having to rebuild the compiler?

In fact I would make this more difficult by burying the check somewhere
strange (e.g. sem_ch3.adb) rather than advertising clearly where it is.
> 
> It is not clear to me that we have a clear consensus on this point
> currently.
> 




From dewar@adacore.com  Thu Sep 28 12:01:06 2006
Message-ID: <451BF1C5.8030205@adacore.com>
Date: Thu, 28 Sep 2006 12:01:09 -0400
From: Robert Dewar <dewar@adacore.com>
MIME-Version: 1.0
To: Romain Berrendonner <berrendo@adacore.com>
CC: Richard Kenner <kenner@vlsi1.ultra.nyu.edu>,  dewar@gnat.com, 
 quinot@adacore.com,  report@adacore.com
References: <20060928102317.304DD48CF77@nile.gnat.com>    <20060928102626.GE17834@melamine.cuivre.fr.eu.org>    <451BA4A1.7020801@adacore.com> <451BA597.7080206@adacore.com> <10609281243.AA10078@vlsi1.ultra.nyu.edu> <451BE368.5040005@adacore.com> <451BE582.50503@adacore.com> <451BE973.8080502@adacore.com>
In-Reply-To: <451BE973.8080502@adacore.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 848
Lines: 24

Romain Berrendonner wrote:
> Romain Berrendonner a crit :
>> Are you sure of that ? I thought that export regulations had been widely 
>> relaxed in 96 or 98 ...
> 
> As a matter of fact, the Wassenaar Arrangement - to which the US are 
> part - allows the use of cryptography where "the cryptographic 
> capability is not-user accessible and which is specially designed and 
> limited to allow any of the following: 1) Execution of copy-protected 
> software" (section 5.A.2.a.9, in 
> http://www.wassenaar.org/controllists/WA-LIST%20(05)%201%20Corr..pdf)
> 
> However I don't know if the US fully comply with this requirement. I 
> would not bet on that ;-)

Right, I don't know.
> 
> (DISCLAIMER: do NOT consider this as a supporting statement for 
> introducing DRM in our technology)

This whole approach is supposed to fall short of DRM!




From brobecker@adacore.com  Thu Sep 28 12:10:20 2006
Date: Thu, 28 Sep 2006 09:10:17 -0700
From: Joel Brobecker <brobecker@adacore.com>
To: Robert Dewar <dewar@adacore.com>
Cc: Cyrille Comar <comar@adacore.com>,
	Thomas Quinot <quinot@adacore.com>, Report <report@adacore.com>,
	Franco Gasperoni <gasperoni@adacore.com>,
	Friess Michael <friess@adacore.com>
Message-ID: <20060928161017.GD1366@adacore.com>
References: <20060928102317.304DD48CF77@nile.gnat.com> <20060928102626.GE17834@melamine.cuivre.fr.eu.org> <451BA4A1.7020801@adacore.com> <20060928104607.GF17834@melamine.cuivre.fr.eu.org> <451BB386.2090606@adacore.com> <20060928114126.GA21600@melamine.cuivre.fr.eu.org> <451BB6E6.6030607@adacore.com> <451BC5BC.9040106@adacore.com> <451BE4FE.5040502@adacore.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <451BE4FE.5040502@adacore.com>
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 1474
Lines: 34

Robert,

I am sending this mesasge privately to you in order to avoid
rediscussing an issue which has probably been discussed at length
(and for which I probably do not have all the necessary info to make
an informed judgement). I just wanted to express my concern with
respect to license keys. I feel strongly against these license
keys, mostly moraly I suppose, but also because of how this might
impact our image.

I really don't know how much business we "lose" because of these
problematic evals. Nor do I know whether these eval would turn
into paying customers should they no longer have unprotected evals.
I am a bit concerned however, of how this will be received by our
current or potential customers.

That reminds me of the tremendous amount of efforts spent by the
music industry trying to protect their music and avoid music sharing.
I think of downloading "free" music on the internet as a form of theft,
so I tried downloading music from sites that sell it legally. It turned
out that, I, the one of the few honest people, had to endure the pains
of protected music: I could not play the music in anything but a Windows
PC, and only on that PC. That machine crashed at one point, and after
reinstall, the player refused to play the music I paid for.

The situation with GNAT is of course only midly comparable, since our
customers cannot continue without a compiler... But my concern about
our image is still present.

Just some of my thoughts...
-- 
Joel



From celier@adacore.com  Thu Sep 28 12:11:22 2006
Message-ID: <451BF427.9090100@adacore.com>
Date: Thu, 28 Sep 2006 09:11:19 -0700
From: Vincent Celier <celier@adacore.com>
MIME-Version: 1.0
To: Thomas Quinot <quinot@adacore.com>
CC: Robert Dewar <dewar@adacore.com>, Cyrille Comar <comar@adacore.com>, 
 Report <report@adacore.com>,
 Franco Gasperoni <gasperoni@adacore.com>, 
 Friess Michael <friess@adacore.com>
References: <20060928102317.304DD48CF77@nile.gnat.com> <20060928102626.GE17834@melamine.cuivre.fr.eu.org> <451BA4A1.7020801@adacore.com> <20060928104607.GF17834@melamine.cuivre.fr.eu.org> <451BB386.2090606@adacore.com> <20060928114126.GA21600@melamine.cuivre.fr.eu.org> <451BB6E6.6030607@adacore.com> <451BC5BC.9040106@adacore.com> <451BE4FE.5040502@adacore.com> <20060928151244.GC28880@melamine.cuivre.fr.eu.org>
In-Reply-To: <20060928151244.GC28880@melamine.cuivre.fr.eu.org>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 1353
Lines: 36


>>But I do think we should avoid cryptographic type signatures.
>>    
>>
>
>Question is, is it acceptable to us that anyone can trivially bypass the
>protection mechanism by looking at compiler source code and regenerating
>a new license file, not having to rebuild the compiler?
>  
>

This is my first comment on this TN.

I was not very confortable with the way this license business was going, 
but I understand the goal: to prevent dishonnest evaluations from easily 
using GNAT Pro after their evaluation has ended.

As for "trivially bypass the protection mechanism by looking at compiler 
source code", there is not such thing: it really take dedication to 
decide to look at the compiler code to find the part that is checking 
for the license, then to understand how the signature is built, then to 
write a program to build another license file. It is easier to get a 
license file on the black market. Also, if you have this kind of 
dedication, it is not so difficult to remove the license checking code 
and to rebuild the compiler.

The goal of the license checking code is to prevent dishonnest people to 
"easily" cheat. Reading the compiler sources is not trivial or easy.

So, I also think that we don't need to get ourselves in possible trouble 
with cryptographic problems and that simple checksums could be used.

--  Vincent




From brobecker@adacore.com  Thu Sep 28 12:20:25 2006
Date: Thu, 28 Sep 2006 09:20:23 -0700
From: Joel Brobecker <brobecker@adacore.com>
To: Robert Dewar <dewar@adacore.com>
Cc: Cyrille Comar <comar@adacore.com>,
	Thomas Quinot <quinot@adacore.com>, Report <report@adacore.com>,
	Franco Gasperoni <gasperoni@adacore.com>,
	Friess Michael <friess@adacore.com>
Message-ID: <20060928162023.GF1366@adacore.com>
References: <20060928102317.304DD48CF77@nile.gnat.com> <20060928102626.GE17834@melamine.cuivre.fr.eu.org> <451BA4A1.7020801@adacore.com> <20060928104607.GF17834@melamine.cuivre.fr.eu.org> <451BB386.2090606@adacore.com> <20060928114126.GA21600@melamine.cuivre.fr.eu.org> <451BB6E6.6030607@adacore.com> <451BC5BC.9040106@adacore.com> <451BE4FE.5040502@adacore.com> <20060928161017.GD1366@adacore.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <20060928161017.GD1366@adacore.com>
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 272
Lines: 12

> I am sending this mesasge privately to you in order to avoid

Just excellent! A private message sent publicly...

Let's say I am glad I didn't pick on anyone in the company :-).
Sorry for the noise, guys. I really do not want to start this discussion
again!

-- 
Joel



From dewar@adacore.com  Thu Sep 28 12:25:23 2006
Message-ID: <451BF777.9080105@adacore.com>
Date: Thu, 28 Sep 2006 12:25:27 -0400
From: Robert Dewar <dewar@adacore.com>
MIME-Version: 1.0
To: Joel Brobecker <brobecker@adacore.com>
CC: Cyrille Comar <comar@adacore.com>, Thomas Quinot <quinot@adacore.com>, 
 Report <report@adacore.com>,
 Franco Gasperoni <gasperoni@adacore.com>, 
 Friess Michael <friess@adacore.com>
References: <20060928102317.304DD48CF77@nile.gnat.com> <20060928102626.GE17834@melamine.cuivre.fr.eu.org> <451BA4A1.7020801@adacore.com> <20060928104607.GF17834@melamine.cuivre.fr.eu.org> <451BB386.2090606@adacore.com> <20060928114126.GA21600@melamine.cuivre.fr.eu.org> <451BB6E6.6030607@adacore.com> <451BC5BC.9040106@adacore.com> <451BE4FE.5040502@adacore.com> <20060928161017.GD1366@adacore.com>
In-Reply-To: <20060928161017.GD1366@adacore.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 2292
Lines: 55

Joel Brobecker wrote:
> Robert,
> 
> I am sending this mesasge privately to you in order to avoid
> rediscussing an issue which has probably been discussed at length
> (and for which I probably do not have all the necessary info to make
> an informed judgement). I just wanted to express my concern with
> respect to license keys. I feel strongly against these license
> keys, mostly moraly I suppose, but also because of how this might
> impact our image.

Well this was certainly not a private message :-)
Hence not a private reply
> 
> I really don't know how much business we "lose" because of these
> problematic evals. Nor do I know whether these eval would turn
> into paying customers should they no longer have unprotected evals.
> I am a bit concerned however, of how this will be received by our
> current or potential customers.

We have discussed this extensively, and this is a decision taken
with a lot of input, and nothing has changed, please let's not
use the occasion of discussing integrity checking on the license
as a license to rediscuss the basic issue!

Remember we are solving several things here, including correctly
a fairly serious GPL violation (failure to provide a copy of the
GPL in our distributions).
> 
> That reminds me of the tremendous amount of efforts spent by the
> music industry trying to protect their music and avoid music sharing.
> I think of downloading "free" music on the internet as a form of theft,
> so I tried downloading music from sites that sell it legally. It turned
> out that, I, the one of the few honest people, had to endure the pains
> of protected music: I could not play the music in anything but a Windows
> PC, and only on that PC. That machine crashed at one point, and after
> reinstall, the player refused to play the music I paid for.

I don't see any anology at all
> 
> The situation with GNAT is of course only midly comparable, since our
> customers cannot continue without a compiler... But my concern about
> our image is still present.

I don't see that this will in anyway affect our customers and that
is the basis of the decision.
> 
> Just some of my thoughts...

We cannot spend too much time discussing stuff that is already
decided unless there are new factors. In this case there are none
that I see.




From dewar@adacore.com  Thu Sep 28 12:30:02 2006
Message-ID: <451BF88E.2040003@adacore.com>
Date: Thu, 28 Sep 2006 12:30:06 -0400
From: Robert Dewar <dewar@adacore.com>
MIME-Version: 1.0
To: Vincent Celier <celier@adacore.com>
CC: Thomas Quinot <quinot@adacore.com>, Cyrille Comar <comar@adacore.com>, 
 Report <report@adacore.com>,
 Franco Gasperoni <gasperoni@adacore.com>, 
 Friess Michael <friess@adacore.com>
References: <20060928102317.304DD48CF77@nile.gnat.com> <20060928102626.GE17834@melamine.cuivre.fr.eu.org> <451BA4A1.7020801@adacore.com> <20060928104607.GF17834@melamine.cuivre.fr.eu.org> <451BB386.2090606@adacore.com> <20060928114126.GA21600@melamine.cuivre.fr.eu.org> <451BB6E6.6030607@adacore.com> <451BC5BC.9040106@adacore.com> <451BE4FE.5040502@adacore.com> <20060928151244.GC28880@melamine.cuivre.fr.eu.org> <451BF427.9090100@adacore.com>
In-Reply-To: <451BF427.9090100@adacore.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 2612
Lines: 60

Vincent Celier wrote:
>>> But I do think we should avoid cryptographic type signatures.
>>>    
>>>
>> Question is, is it acceptable to us that anyone can trivially bypass the
>> protection mechanism by looking at compiler source code and regenerating
>> a new license file, not having to rebuild the compiler?
>>  
>>
> 
> This is my first comment on this TN.
> 
> I was not very confortable with the way this license business was going, 
> but I understand the goal: to prevent dishonnest evaluations from easily 
> using GNAT Pro after their evaluation has ended.

No, that's only one goal, and indeed thinking of it only this way will
cause an incorrect evaluation of the whole approach.

The breakthrough in our discussions was realizing that

a) we are not currently supplying license agreements to our users
b) we aren't even providing the GPL, which is a GPL violation
c) we don't want to correct b) without also correcting a)

One of the important points is that this new scheme means that
every user of GNAT has a proper license agreement that shows
just wnat the licensing conditions are. The checksum is just
meant to help ensure that the license agreement is a valid one,
nothing more nothing less. It can only help int his goal, it
cannot make a gaurantee.
> 
> As for "trivially bypass the protection mechanism by looking at compiler 
> source code", there is not such thing: it really take dedication to 
> decide to look at the compiler code to find the part that is checking 
> for the license, then to understand how the signature is built, then to 
> write a program to build another license file. It is easier to get a 
> license file on the black market. Also, if you have this kind of 
> dedication, it is not so difficult to remove the license checking code 
> and to rebuild the compiler.
> 
> The goal of the license checking code is to prevent dishonnest people to 
> "easily" cheat. Reading the compiler sources is not trivial or easy.

Nope, that's wrong, it is to remind people that they are using 
unsupported software and that they do not have a license from
us anymore. They can continue to use it if they take the effort
(and we cannot prevent this). It's actually not dishonest to continue
to use the software, but the point is that it is only subject to
a GPL license after the trial period has expired, and we want to
emphasize that we are not providing a perpetual GMGPL to eval
customers (if we chose to do so).
> 
> So, I also think that we don't need to get ourselves in possible trouble 
> with cryptographic problems and that simple checksums could be used.

indeed




From quinot@adacore.com  Thu Sep 28 13:01:50 2006
Date: Thu, 28 Sep 2006 19:01:48 +0200
From: Thomas Quinot <quinot@adacore.com>
To: Robert Dewar <dewar@adacore.com>
Cc: Report <report@adacore.com>
Message-ID: <20060928170148.GA33317@melamine.cuivre.fr.eu.org>
References: <20060928102317.304DD48CF77@nile.gnat.com> <20060928102626.GE17834@melamine.cuivre.fr.eu.org> <451BA4A1.7020801@adacore.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <451BA4A1.7020801@adacore.com>
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 772
Lines: 20

* Robert Dewar, 2006-09-28 :

> Actually I would prefer to keep DSA out of the picture, it seems a bit
> provocative if you ask me, and instead use some very simple algorithm,
> just a shift/xor loop would be fine I think. There is really no point
> in high security here. What do others think?

I have an alternative version of instpar that just does a CRC32 of the
license file (reusing s-crc32, which is already part of gnat1). Maybe we
could go with that version for now, knowing that we can always resurrect
the more elaborate DSA-based version from CVS if we see that we have too
many bogus license files floating around.

Thomas.

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From dewar@adacore.com  Thu Sep 28 14:50:36 2006
Message-ID: <451C197E.1080503@adacore.com>
Date: Thu, 28 Sep 2006 14:50:38 -0400
From: Robert Dewar <dewar@adacore.com>
MIME-Version: 1.0
To: Thomas Quinot <quinot@adacore.com>
CC: Report <report@adacore.com>
References: <20060928102317.304DD48CF77@nile.gnat.com> <20060928102626.GE17834@melamine.cuivre.fr.eu.org> <451BA4A1.7020801@adacore.com> <20060928170148.GA33317@melamine.cuivre.fr.eu.org>
In-Reply-To: <20060928170148.GA33317@melamine.cuivre.fr.eu.org>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 705
Lines: 17

Thomas Quinot wrote:
> * Robert Dewar, 2006-09-28 :
> 
>> Actually I would prefer to keep DSA out of the picture, it seems a bit
>> provocative if you ask me, and instead use some very simple algorithm,
>> just a shift/xor loop would be fine I think. There is really no point
>> in high security here. What do others think?
> 
> I have an alternative version of instpar that just does a CRC32 of the
> license file (reusing s-crc32, which is already part of gnat1). Maybe we
> could go with that version for now, knowing that we can always resurrect
> the more elaborate DSA-based version from CVS if we see that we have too
> many bogus license files floating around.

That sounds reasonable to me ...



From kenner@vlsi1.ultra.nyu.edu  Thu Sep 28 21:47:41 2006
From: kenner@vlsi1.ultra.nyu.edu (Richard Kenner)
Message-Id: <10609290151.AA19245@vlsi1.ultra.nyu.edu>
Date: Thu, 28 Sep 2006 21:50:59 EDT
To: berrendo@adacore.com
Cc: dewar@adacore.com, dewar@gnat.com, quinot@adacore.com, report@adacore.com
In-Reply-To: <451BE582.50503@adacore.com>
References: <20060928102317.304DD48CF77@nile.gnat.com>   
    <20060928102626.GE17834@melamine.cuivre.fr.eu.org>   
    <451BA4A1.7020801@adacore.com> <451BA597.7080206@adacore.com>
    <10609281243.AA10078@vlsi1.ultra.nyu.edu>
    <451BE368.5040005@adacore.com> <451BE582.50503@adacore.com>
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 253
Lines: 8

> Are you sure of that ? I thought that export regulations had been widely 
> relaxed in 96 or 98 ...

They were, but people keep asking us for this export code (I forget what it's
called) and if we have encryption, it's different and more complex.
 



From kenner@vlsi1.ultra.nyu.edu  Thu Sep 28 21:48:24 2006
From: kenner@vlsi1.ultra.nyu.edu (Richard Kenner)
Message-Id: <10609290151.AA19286@vlsi1.ultra.nyu.edu>
Date: Thu, 28 Sep 2006 21:51:44 EDT
To: quinot@adacore.com
Cc: comar@adacore.com, dewar@adacore.com, friess@adacore.com,
        gasperoni@adacore.com, report@adacore.com
In-Reply-To: <20060928151244.GC28880@melamine.cuivre.fr.eu.org>
References: <20060928102317.304DD48CF77@nile.gnat.com>
    <20060928102626.GE17834@melamine.cuivre.fr.eu.org>
    <451BA4A1.7020801@adacore.com>
    <20060928104607.GF17834@melamine.cuivre.fr.eu.org>
    <451BB386.2090606@adacore.com>
    <20060928114126.GA21600@melamine.cuivre.fr.eu.org>
    <451BB6E6.6030607@adacore.com> <451BC5BC.9040106@adacore.com>
    <451BE4FE.5040502@adacore.com>
    <20060928151244.GC28880@melamine.cuivre.fr.eu.org>
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 301
Lines: 8

> Question is, is it acceptable to us that anyone can trivially bypass the
> protection mechanism by looking at compiler source code and regenerating
> a new license file, not having to rebuild the compiler?

I'd say "yes" and I think that was the concensus at the discussion
at the annual meeting.



From kenner@vlsi1.ultra.nyu.edu  Thu Sep 28 21:50:21 2006
From: kenner@vlsi1.ultra.nyu.edu (Richard Kenner)
Message-Id: <10609290153.AA19358@vlsi1.ultra.nyu.edu>
Date: Thu, 28 Sep 2006 21:53:39 EDT
To: berrendo@adacore.com
Cc: dewar@adacore.com, dewar@gnat.com, quinot@adacore.com, report@adacore.com
In-Reply-To: <451BE973.8080502@adacore.com>
References: <20060928102317.304DD48CF77@nile.gnat.com>   
    <20060928102626.GE17834@melamine.cuivre.fr.eu.org>   
    <451BA4A1.7020801@adacore.com> <451BA597.7080206@adacore.com>
    <10609281243.AA10078@vlsi1.ultra.nyu.edu>
    <451BE368.5040005@adacore.com> <451BE582.50503@adacore.com>
    <451BE973.8080502@adacore.com>
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 547
Lines: 12

> As a matter of fact, the Wassenaar Arrangement - to which the US are 
> part - allows the use of cryptography where "the cryptographic 
> capability is not-user accessible and which is specially designed and 
> limited to allow any of the following: 1) Execution of copy-protected 
> software" (section 5.A.2.a.9, in 
> http://www.wassenaar.org/controllists/WA-LIST%20(05)%201%20Corr..pdf)

Might one argue that because it's open-source, it's user-accessible?
My point is that I don't even want to get even *slightly* into this sort
of issue.



From kenner@vlsi1.ultra.nyu.edu  Thu Sep 28 21:53:03 2006
From: kenner@vlsi1.ultra.nyu.edu (Richard Kenner)
Message-Id: <10609290156.AA19571@vlsi1.ultra.nyu.edu>
Date: Thu, 28 Sep 2006 21:56:23 EDT
To: dewar@adacore.com
Cc: comar@adacore.com, friess@adacore.com, gasperoni@adacore.com,
        quinot@adacore.com, report@adacore.com
In-Reply-To: <451BF17B.7020708@adacore.com>
References: <20060928102317.304DD48CF77@nile.gnat.com>
    <20060928102626.GE17834@melamine.cuivre.fr.eu.org>
    <451BA4A1.7020801@adacore.com>
    <20060928104607.GF17834@melamine.cuivre.fr.eu.org>
    <451BB386.2090606@adacore.com>
    <20060928114126.GA21600@melamine.cuivre.fr.eu.org>
    <451BB6E6.6030607@adacore.com> <451BC5BC.9040106@adacore.com>
    <451BE4FE.5040502@adacore.com>
    <20060928151244.GC28880@melamine.cuivre.fr.eu.org>
    <451BF17B.7020708@adacore.com>
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 260
Lines: 7

> In fact I would make this more difficult by burying the check somewhere
> strange (e.g. sem_ch3.adb) rather than advertising clearly where it is.

I thought this has to be in the one file that we don't have in common
with the FSF (I forget which this is)?



From kenner@vlsi1.ultra.nyu.edu  Thu Sep 28 21:58:48 2006
From: kenner@vlsi1.ultra.nyu.edu (Richard Kenner)
Message-Id: <10609290202.AA19780@vlsi1.ultra.nyu.edu>
Date: Thu, 28 Sep 2006 22:02:02 EDT
To: brobecker@adacore.com
Cc: comar@adacore.com, dewar@adacore.com, friess@adacore.com,
        gasperoni@adacore.com, quinot@adacore.com, report@adacore.com
In-Reply-To: <20060928161017.GD1366@adacore.com>
References: <20060928102317.304DD48CF77@nile.gnat.com>
    <20060928102626.GE17834@melamine.cuivre.fr.eu.org>
    <451BA4A1.7020801@adacore.com>
    <20060928104607.GF17834@melamine.cuivre.fr.eu.org>
    <451BB386.2090606@adacore.com>
    <20060928114126.GA21600@melamine.cuivre.fr.eu.org>
    <451BB6E6.6030607@adacore.com> <451BC5BC.9040106@adacore.com>
    <451BE4FE.5040502@adacore.com> <20060928161017.GD1366@adacore.com>
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 825
Lines: 18

> I am sending this mesasge privately to you in order to avoid
> rediscussing an issue which has probably been discussed at length
> (and for which I probably do not have all the necessary info to make
> an informed judgement). I just wanted to express my concern with
> respect to license keys. I feel strongly against these license
> keys, mostly moraly I suppose, but also because of how this might
> impact our image.

That is *precisely* why Robert and I are avocating a very *weak*
signature, so that this isn't a "license key", but a license.  And there
has never been a plan to enforce this for non-eval customers

> That reminds me of the tremendous amount of efforts spent by the
> music industry trying to protect their music and avoid music sharing.

... which is exactly why we're advocating a weak mechanism.



From dewar@adacore.com  Fri Sep 29 00:14:33 2006
Message-ID: <451C9DAF.6090909@adacore.com>
Date: Fri, 29 Sep 2006 00:14:39 -0400
From: Robert Dewar <dewar@adacore.com>
MIME-Version: 1.0
To: Richard Kenner <kenner@vlsi1.ultra.nyu.edu>
CC:  comar@adacore.com,  friess@adacore.com,  gasperoni@adacore.com, 
 quinot@adacore.com,  report@adacore.com
References: <20060928102317.304DD48CF77@nile.gnat.com>    <20060928102626.GE17834@melamine.cuivre.fr.eu.org>    <451BA4A1.7020801@adacore.com>    <20060928104607.GF17834@melamine.cuivre.fr.eu.org>    <451BB386.2090606@adacore.com>    <20060928114126.GA21600@melamine.cuivre.fr.eu.org>    <451BB6E6.6030607@adacore.com> <451BC5BC.9040106@adacore.com>    <451BE4FE.5040502@adacore.com>    <20060928151244.GC28880@melamine.cuivre.fr.eu.org>    <451BF17B.7020708@adacore.com> <10609290156.AA19571@vlsi1.ultra.nyu.edu>
In-Reply-To: <10609290156.AA19571@vlsi1.ultra.nyu.edu>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 349
Lines: 11

Richard Kenner wrote:
>> In fact I would make this more difficult by burying the check somewhere
>> strange (e.g. sem_ch3.adb) rather than advertising clearly where it is.
> 
> I thought this has to be in the one file that we don't have in common
> with the FSF (I forget which this is)?

Well we could have another non-common file I suppose ...




From berrendo@adacore.com  Fri Sep 29 03:40:31 2006
Message-ID: <451CCDED.7060703@adacore.com>
Date: Fri, 29 Sep 2006 09:40:29 +0200
From: Romain Berrendonner <berrendo@adacore.com>
MIME-Version: 1.0
To: Robert Dewar <dewar@adacore.com>
CC: Vincent Celier <celier@adacore.com>, 
 Thomas Quinot <quinot@adacore.com>,
 Cyrille Comar <comar@adacore.com>, Report <report@adacore.com>, 
 Franco Gasperoni <gasperoni@adacore.com>,
 Friess Michael <friess@adacore.com>
References: <20060928102317.304DD48CF77@nile.gnat.com> <20060928102626.GE17834@melamine.cuivre.fr.eu.org> <451BA4A1.7020801@adacore.com> <20060928104607.GF17834@melamine.cuivre.fr.eu.org> <451BB386.2090606@adacore.com> <20060928114126.GA21600@melamine.cuivre.fr.eu.org> <451BB6E6.6030607@adacore.com> <451BC5BC.9040106@adacore.com> <451BE4FE.5040502@adacore.com> <20060928151244.GC28880@melamine.cuivre.fr.eu.org> <451BF427.9090100@adacore.com> <451BF88E.2040003@adacore.com>
In-Reply-To: <451BF88E.2040003@adacore.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 482
Lines: 16

Robert Dewar a crit :
> Vincent Celier wrote:
> The breakthrough in our discussions was realizing that
> 
> a) we are not currently supplying license agreements to our users
> b) we aren't even providing the GPL, which is a GPL violation
> c) we don't want to correct b) without also correcting a)

I just checked and we do provide in binary distributions a file called 
COPYING which includes the GPL. I checked linux only, but this should be 
platform independent.

-- 
Romain



From friess@adacore.com  Fri Sep 29 03:49:40 2006
From: =?iso-8859-1?Q?Micha=EBl_Friess?= <friess@adacore.com>
To: "'Romain Berrendonner'" <berrendo@adacore.com>,
	"'Robert Dewar'" <dewar@adacore.com>
Cc: "'Report'" <report@adacore.com>
Date: Fri, 29 Sep 2006 09:49:06 +0200
Message-ID: <00c801c6e39b$bd677d70$0400a8c0@munich>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
In-Reply-To: <451CCDED.7060703@adacore.com>
Subject: RE: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 925
Lines: 31

You will have a hard time finding this file on Windows. It's simply not
there.

Michal

> -----Message d'origine-----
> De: Romain Berrendonner [mailto:berrendo@adacore.com]
> Envoy: Friday, September 29, 2006 9:40 AM
> : Robert Dewar
> Cc: Vincent Celier; Thomas Quinot; Cyrille Comar; Report; Franco
> Gasperoni; Friess Michael
> Objet: Re: F831-014 (license check signature discussion, input requested)
> 
> Robert Dewar a crit :
> > Vincent Celier wrote:
> > The breakthrough in our discussions was realizing that
> >
> > a) we are not currently supplying license agreements to our users
> > b) we aren't even providing the GPL, which is a GPL violation
> > c) we don't want to correct b) without also correcting a)
> 
> I just checked and we do provide in binary distributions a file called
> COPYING which includes the GPL. I checked linux only, but this should be
> platform independent.
> 
> --
> Romain





From berrendo@adacore.com  Fri Sep 29 04:00:42 2006
Message-ID: <451CD2A6.2040406@adacore.com>
Date: Fri, 29 Sep 2006 10:00:38 +0200
From: Romain Berrendonner <berrendo@adacore.com>
MIME-Version: 1.0
To: =?ISO-8859-1?Q?Micha=EBl_Friess?= <friess@adacore.com>
CC: 'Robert Dewar' <dewar@adacore.com>, 'Report' <report@adacore.com>
References: <00c801c6e39b$bd677d70$0400a8c0@munich>
In-Reply-To: <00c801c6e39b$bd677d70$0400a8c0@munich>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 231
Lines: 11

Michal Friess a crit :
> You will have a hard time finding this file on Windows. It's simply not
> there.

Nothing that can't be fixed, obviously. We could even make the GPL 
accessible from GPS without difficulty.

-- 
Romain



From comar@adacore.com  Fri Sep 29 04:14:40 2006
Message-ID: <451CD5EE.60008@adacore.com>
Date: Fri, 29 Sep 2006 10:14:38 +0200
From: Cyrille Comar <comar@adacore.com>
MIME-Version: 1.0
To: =?ISO-8859-1?Q?Micha=EBl_Friess?= <friess@adacore.com>
CC: 'Romain Berrendonner' <berrendo@adacore.com>, 
 'Robert Dewar' <dewar@adacore.com>,
 'Report' <report@adacore.com>
References: <00c801c6e39b$bd677d70$0400a8c0@munich>
In-Reply-To: <00c801c6e39b$bd677d70$0400a8c0@munich>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 227
Lines: 8

Michal Friess a crit :
> You will have a hard time finding this file on Windows. It's simply not
> there.

on Windows, users have to explicitly accept the GPL as part of the 
install process so the GPL is definitely there.



From dewar@adacore.com  Fri Sep 29 06:05:33 2006
Message-ID: <451CEFF3.9050409@adacore.com>
Date: Fri, 29 Sep 2006 06:05:39 -0400
From: Robert Dewar <dewar@adacore.com>
MIME-Version: 1.0
To: Romain Berrendonner <berrendo@adacore.com>
CC: Vincent Celier <celier@adacore.com>, 
 Thomas Quinot <quinot@adacore.com>,
 Cyrille Comar <comar@adacore.com>, Report <report@adacore.com>, 
 Franco Gasperoni <gasperoni@adacore.com>,
 Friess Michael <friess@adacore.com>
References: <20060928102317.304DD48CF77@nile.gnat.com> <20060928102626.GE17834@melamine.cuivre.fr.eu.org> <451BA4A1.7020801@adacore.com> <20060928104607.GF17834@melamine.cuivre.fr.eu.org> <451BB386.2090606@adacore.com> <20060928114126.GA21600@melamine.cuivre.fr.eu.org> <451BB6E6.6030607@adacore.com> <451BC5BC.9040106@adacore.com> <451BE4FE.5040502@adacore.com> <20060928151244.GC28880@melamine.cuivre.fr.eu.org> <451BF427.9090100@adacore.com> <451BF88E.2040003@adacore.com> <451CCDED.7060703@adacore.com>
In-Reply-To: <451CCDED.7060703@adacore.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 829
Lines: 23

Romain Berrendonner wrote:
> Robert Dewar a crit :
>> Vincent Celier wrote:
>> The breakthrough in our discussions was realizing that
>>
>> a) we are not currently supplying license agreements to our users
>> b) we aren't even providing the GPL, which is a GPL violation
>> c) we don't want to correct b) without also correcting a)
> 
> I just checked and we do provide in binary distributions a file called 
> COPYING which includes the GPL. I checked linux only, but this should be 
> platform independent.

Interesting, when we checked before, we were told (I forget by
whom) that COPYING was not in the binary distribution.

This of course solves b), but is a seriously bad situation since
it means that the only license agreement we provide if people do
not have the box is completely wrong and seriously misleading.
> 




From dewar@adacore.com  Fri Sep 29 06:06:10 2006
Message-ID: <451CF018.7090304@adacore.com>
Date: Fri, 29 Sep 2006 06:06:16 -0400
From: Robert Dewar <dewar@adacore.com>
MIME-Version: 1.0
To: Romain Berrendonner <berrendo@adacore.com>
CC: =?ISO-8859-1?Q?Micha=EBl_Friess?= <friess@adacore.com>, 
 'Report' <report@adacore.com>
References: <00c801c6e39b$bd677d70$0400a8c0@munich> <451CD2A6.2040406@adacore.com>
In-Reply-To: <451CD2A6.2040406@adacore.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 398
Lines: 14

Romain Berrendonner wrote:
> Michal Friess a crit :
>> You will have a hard time finding this file on Windows. It's simply not
>> there.
> 
> Nothing that can't be fixed, obviously. We could even make the GPL 
> accessible from GPS without difficulty.

Yes, but the whole point is that we don't WANT to fix this without
also making sure that people have access to the real AdaCore license.
> 




From dewar@adacore.com  Fri Sep 29 06:07:17 2006
Message-ID: <451CF05B.7000706@adacore.com>
Date: Fri, 29 Sep 2006 06:07:23 -0400
From: Robert Dewar <dewar@adacore.com>
MIME-Version: 1.0
To: Robert Dewar <dewar@adacore.com>
CC: Romain Berrendonner <berrendo@adacore.com>, =?ISO-8859-1?Q?Micha=EBl_?=
 =?ISO-8859-1?Q?Friess?= <friess@adacore.com>, 
 'Report' <report@adacore.com>
References: <00c801c6e39b$bd677d70$0400a8c0@munich> <451CD2A6.2040406@adacore.com> <451CF018.7090304@adacore.com>
In-Reply-To: <451CF018.7090304@adacore.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 587
Lines: 16

Robert Dewar wrote:
> Romain Berrendonner wrote:
>> Michal Friess a crit :
>>> You will have a hard time finding this file on Windows. It's simply not
>>> there.
>> Nothing that can't be fixed, obviously. We could even make the GPL 
>> accessible from GPS without difficulty.
> 
> Yes, but the whole point is that we don't WANT to fix this without
> also making sure that people have access to the real AdaCore license.

Actually if people have to accept the GPL on windows for a GNAT Pro
installation, that's a potential real problem. I am surprised it has
not caused more trouble!



From berrendo@adacore.com  Fri Sep 29 06:18:18 2006
Message-ID: <451CF2E3.8000008@adacore.com>
Date: Fri, 29 Sep 2006 12:18:11 +0200
From: Romain Berrendonner <berrendo@adacore.com>
MIME-Version: 1.0
To: Robert Dewar <dewar@adacore.com>
CC: =?ISO-8859-1?Q?Micha=EBl_Friess?= <friess@adacore.com>, 
 'Report' <report@adacore.com>
References: <00c801c6e39b$bd677d70$0400a8c0@munich> <451CD2A6.2040406@adacore.com> <451CF018.7090304@adacore.com>
In-Reply-To: <451CF018.7090304@adacore.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 925
Lines: 29

Robert Dewar a crit :
> Romain Berrendonner wrote:
>> Michal Friess a crit :
>>> You will have a hard time finding this file on Windows. It's simply not
>>> there.
>>
>> Nothing that can't be fixed, obviously. We could even make the GPL 
>> accessible from GPS without difficulty.
> 
> Yes, but the whole point is that we don't WANT to fix this without
> also making sure that people have access to the real AdaCore license.

So shouldn't we have a scheme like:

o On UNIX, the license file contains either the GNAT Pro license 
statement or the GNAT Pro eval license statement;

o On Windows, the same as above + click to accept either the GNAT Pro 
license statement or the GNAT Pro eval license statement in the IS;

o On both systems, keep the COPYING file with the GPL in it at the top 
level of the installation prefix because the GNAT Pro licenses point to 
it and because it is required by the GPL.

-- 
Romain




From quinot@gnat.com  Fri Sep 29 06:32:44 2006
To: file-ci@gnat.com
Message-Id: <20060929103242.0EFAE48CBAF@nile.gnat.com>
Date: Fri, 29 Sep 2006 06:32:42 -0400 (EDT)
From: quinot@gnat.com (Thomas Quinot)
Subject: [ada] instpar.adb (1.6) checked in ada
Status: O
Content-Length: 83
Lines: 6


Use CRC-32 integrity check instead of a full-blown DSA signature.
For F831-014




From quinot@gnat.com  Fri Sep 29 06:32:46 2006
To: file-ci@gnat.com
Message-Id: <20060929103243.3A53148CBF1@nile.gnat.com>
Date: Fri, 29 Sep 2006 06:32:43 -0400 (EDT)
From: quinot@gnat.com (Thomas Quinot)
Subject: [ada] instpar.ads (1.5) checked in ada
Status: O
Content-Length: 83
Lines: 6


Use CRC-32 integrity check instead of a full-blown DSA signature.
For F831-014




From dewar@adacore.com  Fri Sep 29 06:36:40 2006
Message-ID: <451CF73E.8020100@adacore.com>
Date: Fri, 29 Sep 2006 06:36:46 -0400
From: Robert Dewar <dewar@adacore.com>
MIME-Version: 1.0
To: Romain Berrendonner <berrendo@adacore.com>
CC: =?ISO-8859-1?Q?Micha=EBl_Friess?= <friess@adacore.com>, 
 'Report' <report@adacore.com>
References: <00c801c6e39b$bd677d70$0400a8c0@munich> <451CD2A6.2040406@adacore.com> <451CF018.7090304@adacore.com> <451CF2E3.8000008@adacore.com>
In-Reply-To: <451CF2E3.8000008@adacore.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 567
Lines: 19

Romain Berrendonner wrote:

> So shouldn't we have a scheme like:
> 
> o On UNIX, the license file contains either the GNAT Pro license 
> statement or the GNAT Pro eval license statement;
> 
> o On Windows, the same as above + click to accept either the GNAT Pro 
> license statement or the GNAT Pro eval license statement in the IS;
> 
> o On both systems, keep the COPYING file with the GPL in it at the top 
> level of the installation prefix because the GNAT Pro licenses point to 
> it and because it is required by the GPL.

That is exactly the plan ...
> 




From quinot@gnat.com  Fri Sep 29 07:09:50 2006
To: file-ci@gnat.com
Message-Id: <20060929110950.77BCD48CBBD@nile.gnat.com>
Date: Fri, 29 Sep 2006 07:09:50 -0400 (EDT)
From: quinot@gnat.com (Thomas Quinot)
Subject: [scripts] crc32.adb (1.1) checked in scripts/ada_scripts
Status: O
Content-Length: 54
Lines: 6


New tool: compute and output CRC-32.
For F831-014




From quinot@gnat.com  Fri Sep 29 09:31:45 2006
To: file-ci@gnat.com
Message-Id: <20060929133144.6306C48CBF2@nile.gnat.com>
Date: Fri, 29 Sep 2006 09:31:44 -0400 (EDT)
From: quinot@gnat.com (Thomas Quinot)
Subject: [ada] instpar.adb (1.9) checked in ada
Status: O
Content-Length: 199
Lines: 8


Load gnatlic.txt as a source file, so that the default search path is used.
This means that a gnatlic.txt file from the current directory can be used
when building the GNAT runtime.
For F831-014




From quinot@adacore.com  Fri Sep 29 09:37:00 2006
Date: Fri, 29 Sep 2006 15:36:58 +0200
From: Thomas Quinot <quinot@adacore.com>
To: Romain Berrendonner <berrendo@adacore.com>
Cc: Robert Dewar <dewar@adacore.com>,
	=?iso-8859-1?Q?Micha=EBl?= Friess <friess@adacore.com>,
	'Report' <report@adacore.com>
Message-ID: <20060929133658.GA73015@melamine.cuivre.fr.eu.org>
References: <00c801c6e39b$bd677d70$0400a8c0@munich> <451CD2A6.2040406@adacore.com> <451CF018.7090304@adacore.com> <451CF2E3.8000008@adacore.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <451CF2E3.8000008@adacore.com>
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 580
Lines: 19

* Romain Berrendonner, 2006-09-29 :

> o On UNIX, the license file contains either the GNAT Pro license 
> statement or the GNAT Pro eval license statement;

Speaking of which...

In order to make progress on the implementation of this scheme, I need
to be able to generate an appropriate license file as part of our build
process. Where can I find the text of the GNAT Pro license that we want
to include in the GNAT Pro license file?

Thomas.

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From friess@adacore.com  Fri Sep 29 11:09:15 2006
From: =?iso-8859-1?Q?Micha=EBl_Friess?= <friess@adacore.com>
To: "'Romain Berrendonner'" <berrendo@adacore.com>
Cc: "'Report'" <report@adacore.com>
Date: Fri, 29 Sep 2006 17:08:24 +0200
Message-ID: <001601c6e3d9$1c5e7a80$89000a0a@munich>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
In-Reply-To: <451CF2E3.8000008@adacore.com>
Subject: RE: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 802
Lines: 26

> -----Message d'origine-----
> De: Romain Berrendonner [mailto:berrendo@adacore.com]
> Envoy: Friday, September 29, 2006 12:18 PM
> : Robert Dewar
> Cc: Michal Friess; 'Report'
> Objet: Re: F831-014 (license check signature discussion, input requested)
> 
> So shouldn't we have a scheme like:
> 
> o On UNIX, the license file contains either the GNAT Pro license
> statement or the GNAT Pro eval license statement;
> 
> o On Windows, the same as above + click to accept either the GNAT Pro
> license statement or the GNAT Pro eval license statement in the IS;
> 
> o On both systems, keep the COPYING file with the GPL in it at the top
> level of the installation prefix because the GNAT Pro licenses point to
> it and because it is required by the GPL.

Seems reasonable to me.

Michal





From fofanov@adacore.com  Fri Sep 29 11:24:18 2006
Date: Fri, 29 Sep 2006 17:24:16 +0200
From: Vasiliy Fofanov <fofanov@adacore.com>
Message-ID: <57146900.20060929172416@adacore.com>
To: Robert Dewar <dewar@adacore.com>
CC: Romain Berrendonner <berrendo@adacore.com>, Michal Friess <friess@adacore.com>, 
	'Report' <report@adacore.com>
In-Reply-To: <451CF05B.7000706@adacore.com>
References: <00c801c6e39b$bd677d70$0400a8c0@munich>
 <451CD2A6.2040406@adacore.com> <451CF018.7090304@adacore.com>
 <451CF05B.7000706@adacore.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 346
Lines: 13

> Actually if people have to accept the GPL on windows for a GNAT Pro
> installation, that's a potential real problem. I am surprised it has
> not caused more trouble!

It has not caused more trouble because it is a red herring. When installing
GNAT Pro people are presented with "GNAT Pro License", not GPL.


Best regards,
 Vasiliy Fofanov.




From fofanov@adacore.com  Fri Sep 29 11:28:57 2006
Date: Fri, 29 Sep 2006 17:28:55 +0200
From: Vasiliy Fofanov <fofanov@adacore.com>
Message-ID: <1017009894.20060929172855@adacore.com>
To: Robert Dewar <dewar@adacore.com>
CC: Romain Berrendonner <berrendo@adacore.com>, Vincent Celier <celier@adacore.com>, 
	Thomas Quinot <quinot@adacore.com>, Cyrille Comar <comar@adacore.com>, Report <report@adacore.com>, 
	Franco Gasperoni <gasperoni@adacore.com>, Friess Michael <friess@adacore.com>
In-Reply-To: <451CEFF3.9050409@adacore.com>
References: <20060928102317.304DD48CF77@nile.gnat.com>
 <20060928102626.GE17834@melamine.cuivre.fr.eu.org>
 <451BA4A1.7020801@adacore.com>
 <20060928104607.GF17834@melamine.cuivre.fr.eu.org>
 <451BB386.2090606@adacore.com>
 <20060928114126.GA21600@melamine.cuivre.fr.eu.org>
 <451BB6E6.6030607@adacore.com> <451BC5BC.9040106@adacore.com>
 <451BE4FE.5040502@adacore.com>
 <20060928151244.GC28880@melamine.cuivre.fr.eu.org>
 <451BF427.9090100@adacore.com> <451BF88E.2040003@adacore.com>
 <451CCDED.7060703@adacore.com> <451CEFF3.9050409@adacore.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 408
Lines: 13

> Interesting, when we checked before, we were told (I forget by
> whom) that COPYING was not in the binary distribution.

That was me, but Romain and I checked different things apparently - I
checked the actual installed compiler, and it doesn't have this file.
Romain checked the tarball, and apparently that one does. So I guess this
file is shipped but not installed.

Best regards,
 Vasiliy Fofanov.




From dismukes@gnat.com  Fri Sep 29 13:36:53 2006
To: berrendo@adacore.com, dewar@adacore.com
Cc: friess@adacore.com, report@adacore.com
Message-Id: <20060929173652.8296848CBAE@nile.gnat.com>
Date: Fri, 29 Sep 2006 13:36:52 -0400 (EDT)
From: dismukes@gnat.com (Gary Dismukes)
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 1312
Lines: 36

Romain Berrendonner wrote:

> Robert Dewar a crit :
> > Romain Berrendonner wrote:
> >> Michal Friess a crit :
> >>> You will have a hard time finding this file on Windows. It's simply not
> >>> there.
> >>
> >> Nothing that can't be fixed, obviously. We could even make the GPL 
> >> accessible from GPS without difficulty.
> > 
> > Yes, but the whole point is that we don't WANT to fix this without
> > also making sure that people have access to the real AdaCore license.
> 
> So shouldn't we have a scheme like:
> 
> o On UNIX, the license file contains either the GNAT Pro license 
> statement or the GNAT Pro eval license statement;
> 
> o On Windows, the same as above + click to accept either the GNAT Pro 
> license statement or the GNAT Pro eval license statement in the IS;
> 
> o On both systems, keep the COPYING file with the GPL in it at the top 
> level of the installation prefix because the GNAT Pro licenses point to 
> it and because it is required by the GPL.

Related question for the GNAAMP release: what license files should I be
providing to Rockwell?  I currently include COPYING, but it sounds like
there's an AdaCore license file (the "real AdaCore license" referred to
above) that should also be included.  If so, what's the name of that
file and where can I find it?

-- Gary




From dewar@adacore.com  Fri Sep 29 15:25:38 2006
Message-ID: <451D7332.5050609@adacore.com>
Date: Fri, 29 Sep 2006 15:25:38 -0400
From: Robert Dewar <dewar@adacore.com>
MIME-Version: 1.0
To: Vasiliy Fofanov <fofanov@adacore.com>
CC: Romain Berrendonner <berrendo@adacore.com>, =?ISO-8859-1?Q?Micha=EBl_?=
 =?ISO-8859-1?Q?Friess?= <friess@adacore.com>, 
 'Report' <report@adacore.com>
References: <00c801c6e39b$bd677d70$0400a8c0@munich> <451CD2A6.2040406@adacore.com> <451CF018.7090304@adacore.com> <451CF05B.7000706@adacore.com> <57146900.20060929172416@adacore.com>
In-Reply-To: <57146900.20060929172416@adacore.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 542
Lines: 19

Vasiliy Fofanov wrote:
>> Actually if people have to accept the GPL on windows for a GNAT Pro
>> installation, that's a potential real problem. I am surprised it has
>> not caused more trouble!
> 
> It has not caused more trouble because it is a red herring. When installing
> GNAT Pro people are presented with "GNAT Pro License", not GPL.

OK, so the claim that people had to accept the GPL was not a red
herring, it was just wrong.

Red herring refers to something that is real but irrelevant!
> 
> 
> Best regards,
>  Vasiliy Fofanov.




From berrendo@adacore.com  Fri Sep 29 16:16:01 2006
Message-ID: <451D7EDA.1000607@adacore.com>
Date: Fri, 29 Sep 2006 22:15:22 +0200
From: Romain Berrendonner <berrendo@adacore.com>
MIME-Version: 1.0
To: Gary Dismukes <dismukes@gnat.com>
CC:  dewar@adacore.com,  friess@adacore.com,  report@adacore.com
References: <20060929173652.8296848CBAE@nile.gnat.com>
In-Reply-To: <20060929173652.8296848CBAE@nile.gnat.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 640
Lines: 15

Gary Dismukes a crit :

> Related question for the GNAAMP release: what license files should I be
> providing to Rockwell?  I currently include COPYING, but it sounds like
> there's an AdaCore license file (the "real AdaCore license" referred to
> above) that should also be included.  If so, what's the name of that
> file and where can I find it?

As far as I know, there are an HTML only versions of these files 
available at: https://www2.adacore.com/internal.auth/Sales/
but they are marked as "experimental". I don't know their current 
status. We need a text version for inclusion in the licence files and in 
the Install Shield.



From comar@adacore.com  Sat Sep 30 04:53:35 2006
Message-ID: <451E308A.5060306@adacore.com>
Date: Sat, 30 Sep 2006 10:53:30 +0200
From: Cyrille Comar <comar@adacore.com>
MIME-Version: 1.0
To: Robert Dewar <dewar@adacore.com>
CC: Vasiliy Fofanov <fofanov@adacore.com>, 
 Romain Berrendonner <berrendo@adacore.com>,
 =?ISO-8859-1?Q?Micha=EBl_Friess?= <friess@adacore.com>, 
 'Report' <report@adacore.com>
References: <00c801c6e39b$bd677d70$0400a8c0@munich> <451CD2A6.2040406@adacore.com> <451CF018.7090304@adacore.com> <451CF05B.7000706@adacore.com> <57146900.20060929172416@adacore.com> <451D7332.5050609@adacore.com>
In-Reply-To: <451D7332.5050609@adacore.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 787
Lines: 21

Robert Dewar wrote:
> Vasiliy Fofanov wrote:
>>> Actually if people have to accept the GPL on windows for a GNAT Pro
>>> installation, that's a potential real problem. I am surprised it has
>>> not caused more trouble!
>>
>> It has not caused more trouble because it is a red herring. When 
>> installing
>> GNAT Pro people are presented with "GNAT Pro License", not GPL.
> 
> OK, so the claim that people had to accept the GPL was not a red
> herring, it was just wrong.

Sorry for the confusion I generated here. The only intent of my claim 
was that we already distribute the text of the GPL and/or a GNAT Pro 
license in our distribs and each time we have found a case where it was 
missing in the past, we have fixed it. Maybe there are some other cases, 
but they are marginal.




From dewar@adacore.com  Sat Sep 30 08:16:25 2006
Message-ID: <451E6018.8060003@adacore.com>
Date: Sat, 30 Sep 2006 08:16:24 -0400
From: Robert Dewar <dewar@adacore.com>
MIME-Version: 1.0
To: Cyrille Comar <comar@adacore.com>
CC: Vasiliy Fofanov <fofanov@adacore.com>, 
 Romain Berrendonner <berrendo@adacore.com>,
 =?ISO-8859-1?Q?Micha=EBl_Friess?= <friess@adacore.com>, 
 'Report' <report@adacore.com>
References: <00c801c6e39b$bd677d70$0400a8c0@munich> <451CD2A6.2040406@adacore.com> <451CF018.7090304@adacore.com> <451CF05B.7000706@adacore.com> <57146900.20060929172416@adacore.com> <451D7332.5050609@adacore.com> <451E308A.5060306@adacore.com>
In-Reply-To: <451E308A.5060306@adacore.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 1023
Lines: 28

Cyrille Comar wrote:
> Robert Dewar wrote:
>> Vasiliy Fofanov wrote:
>>>> Actually if people have to accept the GPL on windows for a GNAT Pro
>>>> installation, that's a potential real problem. I am surprised it has
>>>> not caused more trouble!
>>> It has not caused more trouble because it is a red herring. When 
>>> installing
>>> GNAT Pro people are presented with "GNAT Pro License", not GPL.
>> OK, so the claim that people had to accept the GPL was not a red
>> herring, it was just wrong.
> 
> Sorry for the confusion I generated here. The only intent of my claim 
> was that we already distribute the text of the GPL and/or a GNAT Pro 
> license

Well distributing our license without the GPL is a GPL violation, and
distributing the GPL without our license is a potential serious problem,
so the "or" here is a bit misleading!

It is very important that we always distribute both, and this new 
license approach will ensure that

a) we distribute both
b) we properly install them
c) they are easily accessible



From berrendo@adacore.com  Sat Sep 30 13:51:53 2006
Message-ID: <451EAEB4.2050003@adacore.com>
Date: Sat, 30 Sep 2006 19:51:48 +0200
From: Romain Berrendonner <berrendo@adacore.com>
MIME-Version: 1.0
To: Robert Dewar <dewar@adacore.com>
CC: Cyrille Comar <comar@adacore.com>, 
 Vasiliy Fofanov <fofanov@adacore.com>,
 =?ISO-8859-1?Q?Micha=EBl_Friess?= <friess@adacore.com>, 
 'Report' <report@adacore.com>
References: <00c801c6e39b$bd677d70$0400a8c0@munich> <451CD2A6.2040406@adacore.com> <451CF018.7090304@adacore.com> <451CF05B.7000706@adacore.com> <57146900.20060929172416@adacore.com> <451D7332.5050609@adacore.com> <451E308A.5060306@adacore.com> <451E6018.8060003@adacore.com>
In-Reply-To: <451E6018.8060003@adacore.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: F831-014 (license check signature discussion, input requested)
Status: O
Content-Length: 243
Lines: 12

Robert Dewar a crit :
> Well distributing our license without the GPL is a GPL violation

In addition, it is a violation of our own license ;-)

> a) we distribute both
> b) we properly install them
> c) they are easily accessible

Agreed.



From quinot@malevil.act-europe.fr  Mon Oct  2 05:49:12 2006
Date: Mon, 2 Oct 2006 11:49:36 +0200 (CEST)
From: Thomas Quinot <quinot@malevil.act-europe.fr>
Message-Id: <200610020949.k929naZm007863@malevil.act-europe.fr>
To: file@gnat.com
Subject: [F831-014] crc32.adb(1.2) RH change
Status: O
Content-Length: 136
Lines: 6

Reimplement using Stream_IO to avoid issues with whitespace characters.
----
date: 2006/10/02 09:49:33;  author: quinot;
For F831-014



From quinot@gnat.com  Thu Oct  5 06:38:49 2006
To: file-ci@gnat.com
Message-Id: <20061005103848.AC6E148CEDC@nile.gnat.com>
Date: Thu,  5 Oct 2006 06:38:48 -0400 (EDT)
From: quinot@gnat.com (Thomas Quinot)
Subject: [scripts] license-utils (1.1) checked in scripts/nightly
Status: O
Content-Length: 82
Lines: 6


Add utility routines to support the generation of license files.
For F831-014




From quinot@gnat.com  Thu Oct  5 06:39:16 2006
To: file-ci@gnat.com
Message-Id: <20061005103916.43ADC48CEDC@nile.gnat.com>
Date: Thu,  5 Oct 2006 06:39:16 -0400 (EDT)
From: quinot@gnat.com (Thomas Quinot)
Subject: [scripts] cosrc (1.162) checked in scripts/nightly
Status: O
Content-Length: 65
Lines: 6


Generate a gnatlic.txt file along with sources.
For F831-014




From quinot@gnat.com  Thu Oct  5 06:47:18 2006
To: file-ci@gnat.com
Message-Id: <20061005104718.1DF0F48CEDC@nile.gnat.com>
Date: Thu,  5 Oct 2006 06:47:18 -0400 (EDT)
From: quinot@gnat.com (Thomas Quinot)
Subject: [scripts] cosrc (1.163) checked in scripts/nightly
Status: O
Content-Length: 124
Lines: 7


Fix incorrect progression message for license file generation.
Fix PATH for GPL text (case of GPL package)
For F831-014




From quinot@adacore.com  Thu Oct  5 10:58:31 2006
Date: Thu, 5 Oct 2006 16:58:28 +0200
From: Thomas Quinot <quinot@adacore.com>
To: report@adacore.com
Cc: banner@adacore.com
Message-ID: <20061005145828.GB44537@melamine.cuivre.fr.eu.org>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Subject: F726-008 is merged into F831-014
Status: O
Content-Length: 1019
Lines: 25

There's really no point in having two different TNs here, since both TNs
concern implementation of the technical infrastructure for limited time
evals.

The one point from F726-008 that has not been captured yet into F831-014
is that eval packages should display 'GNAT Pro Eval X.XXx (YYYYMMDD-VV)'
instead of just 'GNAT Pro ...'.

A proposed technical implementation of that is to have a boolean
variable in gnatvsn determining whether or not to insert the word 'Eval'
in Gnat_Version_String, with a default value of False (no 'Eval'), which
could be changed to True when parsing the license file.

*However* another ramification of this is if we want the version string
to be controlled by the license file, then *all tools* (not just gnat1)
need to check the license file. Otherwise, we will have to specialize
the build (ie have different sources) for the eval packages.

Thomas.

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From quinot@adacore.com  Tue Oct 10 05:53:27 2006
Date: Tue, 10 Oct 2006 11:53:58 +0200
From: Thomas Quinot <quinot@adacore.com>
To: scripts@adacore.com
Message-ID: <20061010095358.GA2687@malevil.act-europe.fr>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Subject: F831-014 cosrc change for review (check in gnatlic.txt)
Status: O
Content-Length: 955
Lines: 28

Review of the enclosed diff against cosrc is kindly requested. The
purpose of the change is to have a checked-in copy of gnatlic.txt
consistent with the current version of gnatvsn.ads in CVS, for the
convenience of those who build the compiler directly from the
repository.

Index: cosrc
===================================================================
RCS file: /nile.c/cvs/Dev/scripts/nightly/cosrc,v
retrieving revision 1.163
diff -u -r1.163 cosrc
--- cosrc	5 Oct 2006 10:47:17 -0000	1.163
+++ cosrc	10 Oct 2006 09:51:31 -0000
@@ -217,6 +217,9 @@
     sign_gnatlic_file gnatlic.txt.in > $work/src/ada/gnatlic.txt &&
     cd .. &&
     rm -fr $tmpdir
+    if [ "$enable_checkins" = "true" ]; then
+       cvs commit -m "Automated updated of license file" gnatlic.txt
+    fi
 )
 if [ "$?" -ne 0 ]; then exit 1; fi
 
-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From quinot@gnat.com  Tue Oct 10 11:51:25 2006
To: file-ci@gnat.com
Message-Id: <20061010155125.4CE4248CCBE@nile.gnat.com>
Date: Tue, 10 Oct 2006 11:51:25 -0400 (EDT)
From: quinot@gnat.com (Thomas Quinot)
Subject: [ada] gnatlic.txt (1.1) checked in ada
Status: O
Content-Length: 120
Lines: 7


Check in a default GNATPro license file (to be regenerated automatically
each time gnatvsn is updated)
For F831-014




From quinot@gnat.com  Tue Oct 10 11:59:14 2006
To: file-ci@gnat.com
Message-Id: <20061010155913.D9F4D48CBDC@nile.gnat.com>
Date: Tue, 10 Oct 2006 11:59:13 -0400 (EDT)
From: quinot@gnat.com (Thomas Quinot)
Subject: [scripts] cosrc (1.164) checked in scripts/nightly
Status: O
Content-Length: 76
Lines: 6


If $enable_checkins is true, check in updated gnatlic.txt.
For F831-014




From quinot@gnat.com  Wed Oct 11 03:50:27 2006
To: file-ci@gnat.com
Message-Id: <20061011075026.D34C748CBCB@nile.gnat.com>
Date: Wed, 11 Oct 2006 03:50:26 -0400 (EDT)
From: quinot@gnat.com (Thomas Quinot)
Subject: [scripts] cosrc (1.165) checked in scripts/nightly
Status: O
Content-Length: 160
Lines: 8


Leave CVS symbolic link in place in $work/src/ada after updating gnatvsn.ads,
so we can check in gnatlic.txt.
Fixes sources packaging failure.
For F831-014




From quinot@gnat.com  Wed Oct 11 05:19:40 2006
To: file-ci@gnat.com
Message-Id: <20061011091939.9F89948CE17@nile.gnat.com>
Date: Wed, 11 Oct 2006 05:19:39 -0400 (EDT)
From: quinot@gnat.com (Thomas Quinot)
Subject: [ada] MANIFEST.GNAT (1.300) checked in ada
Status: O
Content-Length: 50
Lines: 6


Add gnatlic.txt to MANIFEST.GNAT
For F831-014




From quinot@adacore.com  Wed Oct 11 11:02:49 2006
Date: Wed, 11 Oct 2006 17:02:51 +0200
From: Thomas Quinot <quinot@adacore.com>
To: Nicolas Roche <roche@gnat.com>
Cc: file@adacore.com
Message-ID: <20061011150251.GA25830@melamine.cuivre.fr.eu.org>
References: <20061011145616.7849648CEEC@nile.gnat.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <20061011145616.7849648CEEC@nile.gnat.com>
Subject: Re: F831-014 [gnat] gnatlic.txt (1.1.8.2:gpl-branch) checked in gnat
Status: O
Content-Length: 495
Lines: 17

* Nicolas Roche, 2006-10-11 :

> Subject: [gnat] gnatlic.txt (1.1.8.2:gpl-branch) checked in gnat
> +The following license to use GNAT Pro 5.05w (20061009-34)
> +is granted to You until 2009-01-01.

Hmmm, strange, aren't we supposed to edit gnatvsn to use a different
version number, and display 'GPL' instead of 'Pro' for the case of the
GPL branch?

Thomas.

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From quinot@gnat.com  Wed Oct 18 11:24:19 2006
To: file-ci@gnat.com
Message-Id: <20061018152418.E1D6E48CDA0@nile.gnat.com>
Date: Wed, 18 Oct 2006 11:24:18 -0400 (EDT)
From: quinot@gnat.com (Thomas Quinot)
Subject: [ada] Makefile.in (1.2027) checked in ada
Status: O
Content-Length: 73
Lines: 6


Install gnatlic.txt along with the GNAT runtime library
For F831-014




From quinot@gnat.com  Wed Oct 18 11:25:27 2006
To: file-ci@gnat.com
Message-Id: <20061018152526.A937348CF09@nile.gnat.com>
Date: Wed, 18 Oct 2006 11:25:26 -0400 (EDT)
From: quinot@gnat.com (Thomas Quinot)
Subject: [gnat] Makefile.in (1.1679.2.206.2.15:gcc-4_1) checked in gnat
Status: O
Content-Length: 102
Lines: 7


Merge rev. 1.2027 from HEAD:
Install gnatlic.txt along with the GNAT runtime library
For F831-014




From quinot@gnat.com  Wed Oct 18 11:26:15 2006
To: file-ci@gnat.com
Message-Id: <20061018152614.A8C0C48CF07@nile.gnat.com>
Date: Wed, 18 Oct 2006 11:26:14 -0400 (EDT)
From: quinot@gnat.com (Thomas Quinot)
Subject: [ada] Makefile.in (1.1679.2.220:gcc-head) checked in ada
Status: O
Content-Length: 102
Lines: 7


Merge rev. 1.2027 from HEAD:
Install gnatlic.txt along with the GNAT runtime library
For F831-014




From quinot@adacore.com  Wed Oct 18 11:27:43 2006
Date: Wed, 18 Oct 2006 17:27:42 +0200
From: Thomas Quinot <quinot@adacore.com>
To: report@adacore.com
Message-ID: <20061018152741.GB78988@melamine.cuivre.fr.eu.org>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Subject: F831-014 progress
Status: O
Content-Length: 461
Lines: 13

We now generate a new license file (gnatlic.txt) nightly for GNAT Pro,
right after updating gnatvsn.ads. I have just checked in Makefile
changes so that this file is installed along with the runtime library
sources. If these changes work correctly, I'll hook the license file
validation code into the build tomorrow.

Thomas.

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From quinot@gnat.com  Fri Oct 20 12:16:54 2006
To: file-ci@gnat.com
Message-Id: <20061020161654.477F048CBBF@nile.gnat.com>
Date: Fri, 20 Oct 2006 12:16:54 -0400 (EDT)
From: quinot@gnat.com (Thomas Quinot)
Subject: [scripts] make-bin-5 (1.140) checked in scripts/nightly
Status: O
Content-Length: 122
Lines: 7


When packaging each version of the GNAT runtime, copy the gnatlic.txt
file to the destination directory.
For F831-014




From quinot@adacore.com  Fri Oct 20 12:18:48 2006
Date: Fri, 20 Oct 2006 18:18:46 +0200
From: Thomas Quinot <quinot@adacore.com>
To: fofanov@adacore.com
Cc: file@adacore.com
Message-ID: <20061020161846.GA74972@melamine.cuivre.fr.eu.org>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Subject: F831-014 gnatlic.txt: InstallShield update needed?
Status: O
Content-Length: 1693
Lines: 51

I've modified the binaries packaging to include gnatlic.txt, which must
exist in each rts-* directory. make-bin-5 has a comment that such
changes should be notified to you for the purpose of updating the
InstallShield scripts, so here's your notification :-)

Thomas.

----- Forwarded message from Thomas Quinot <quinot@gnat.com> -----

Old-To: gci@gnat.com,roche@act-europe.fr
Subject: [scripts] make-bin-5 (1.140) checked in scripts/nightly
X-ACT-Mailfilter: check-in in scripts
X-ACT-GTI: http://www.adacore.com/gti/tn.html?View_TN=F831-014
Date: Fri, 20 Oct 2006 12:16:54 -0400 (EDT)
From: Thomas Quinot <quinot@gnat.com>
X-ACT-List: gci
X-ACT-TN: 
Reply-To: quinot@gnat.com
To: gci-do-not-reply@adacore.com


When packaging each version of the GNAT runtime, copy the gnatlic.txt
file to the destination directory.
For F831-014


Index: scripts/nightly/make-bin-5
===================================================================
RCS file: make-bin-5,v
retrieving revision 1.139
retrieving revision 1.140
diff -u -p -r1.139 -r1.140
--- make-bin-5	2006/09/25 17:28:20	1.139
+++ make-bin-5	2006/10/20 16:16:54	1.140
@@ -172,7 +172,7 @@ log_misc "run-times"
       mkdir $brtsdir/adainclude &&
       mkdir $brtsdir/adalib &&
       chmod u+w $brtsdir $brtsdir/adainclude $brtsdir/adalib &&
-      for i in $rtsdir/*.ads $rtsdir/*.adb; do
+      for i in $rtsdir/gnatlic.txt $rtsdir/*.ads $rtsdir/*.adb; do
         cp -p $i $brtsdir/adainclude 
       done &&
       (cp -p $rtsdir/*.c $brtsdir/adainclude || true ) &&

----- End forwarded message -----

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From fofanov@adacore.com  Sat Oct 21 05:11:22 2006
Date: Sat, 21 Oct 2006 11:11:24 +0200
From: Vasiliy Fofanov <fofanov@adacore.com>
Message-ID: <1223788986.20061021111124@adacore.com>
To: Thomas Quinot <quinot@adacore.com>
CC: file@adacore.com
In-Reply-To: <20061020161846.GA74972@melamine.cuivre.fr.eu.org>
References: <20061020161846.GA74972@melamine.cuivre.fr.eu.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Subject: Re: F831-014 gnatlic.txt: InstallShield update needed?
Status: O
Content-Length: 370
Lines: 13


> I've modified the binaries packaging to include gnatlic.txt, which must
> exist in each rts-* directory. make-bin-5 has a comment that such
> changes should be notified to you for the purpose of updating the
> InstallShield scripts, so here's your notification :-)

Thanks. I think this change should not incur any change to IS...

Best regards,
 Vasiliy Fofanov.




From quinot@gnat.com  Tue Oct 24 11:04:56 2006
To: file-ci@gnat.com
Message-Id: <20061024150455.0307648CBEB@nile.gnat.com>
Date: Tue, 24 Oct 2006 11:04:55 -0400 (EDT)
From: quinot@gnat.com (Thomas Quinot)
Subject: [ada] MANIFEST.GNAT (1.303) checked in ada
Status: O
Content-Length: 87
Lines: 7


Add instpar.ads and instpar.adb
For F831-014
NOT TO BE REFLECTED IN FSF REPOSITORY




From quinot@adacore.com  Tue Oct 24 13:01:47 2006
Date: Tue, 24 Oct 2006 19:01:45 +0200
From: Thomas Quinot <quinot@adacore.com>
To: friess@adacore.com, gasperoni@adacore.com, berrendonner@adacore.com,
	dewar@adacore.com
Cc: file@adacore.com
Message-ID: <20061024170145.GD39589@melamine.cuivre.fr.eu.org>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Subject: F831-014 status update
Status: O
Content-Length: 816
Lines: 19

The license file circuitry is now almost ready to be activated. I want
to wait for tonight's builds to verify that the file is correctly
packaged in HIE binary distributions, and once this is validated I will
send an announcement to report@ to let people know that the check is
about to be activated. So, unless there is a major opposition to this,
you can expect that this will be done early next week.

>From then on, there will be only one missing piece to start using this
feature for evals: defining a procedure to prepare a specific license
file for an eval customer and to prepare a binary package containing
that license file instead of the standard GNAT Pro one.

Thomas.

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From friess@adacore.com  Wed Oct 25 10:22:36 2006
From: =?iso-8859-1?Q?Micha=EBl_Friess?= <friess@adacore.com>
To: "'Thomas Quinot'" <quinot@adacore.com>
Cc: <file@adacore.com>
References: <20061024170145.GD39589@melamine.cuivre.fr.eu.org>
In-Reply-To: <20061024170145.GD39589@melamine.cuivre.fr.eu.org>
Date: Wed, 25 Oct 2006 16:22:26 +0200
Message-ID: <00a601c6f840$ff19e3c0$fd4dab40$@com>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: RE: F831-014 status update
Status: O
Content-Length: 1161
Lines: 32

Very good news! Thanks :)

Michal

> -----Original Message-----
> From: Thomas Quinot [mailto:quinot@adacore.com]
> Sent: Tuesday, October 24, 2006 7:02 PM
> To: friess@adacore.com; gasperoni@adacore.com;
> berrendonner@adacore.com; dewar@adacore.com
> Cc: file@adacore.com
> Subject: F831-014 status update
> 
> The license file circuitry is now almost ready to be activated. I want
> to wait for tonight's builds to verify that the file is correctly
> packaged in HIE binary distributions, and once this is validated I will
> send an announcement to report@ to let people know that the check is
> about to be activated. So, unless there is a major opposition to this,
> you can expect that this will be done early next week.
> 
> From then on, there will be only one missing piece to start using this
> feature for evals: defining a procedure to prepare a specific license
> file for an eval customer and to prepare a binary package containing
> that license file instead of the standard GNAT Pro one.
> 
> Thomas.
> 
> --
> Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
>                AdaCore -- Paris, France -- New York, USA




From quinot@adacore.com  Mon Nov 13 08:35:18 2006
Date: Mon, 13 Nov 2006 14:35:17 +0100
From: Thomas Quinot <quinot@adacore.com>
To: report@adacore.com
Message-ID: <20061113133517.GC28678@melamine.cuivre.fr.eu.org>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Subject: F831-014 HEADS-UP: about to activate license file verification
Status: O
Content-Length: 919
Lines: 28

Attention all GNAT developers!

I am about to enable verification of the license file in the compiler.
A valid gnatlic.txt file is always available for the current wavefront
in CVS; this file is updated nightly at the same time as gnatvsn.ads.
It is installed automatically in the GNAT runtime directory
(adainclude).

What to do if something goes wrong
==================================

If for some reason the license file verification fails on your setup,
you will get an error message of the form:
"fatal error, installation parameters check failed: NNN"

If this happens:
  * please report the problem under this TN;
  * if this is blocking for you, you can disable the check locally
    by commenting out the call to Check_Installation_Parameters
    in gnat1drv.adb.

Thomas.

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From quinot@gnat.com  Tue Nov 14 09:18:47 2006
To: file-ci@gnat.com
Message-Id: <20061114141846.AF0BA48D02A@nile.gnat.com>
Date: Tue, 14 Nov 2006 09:18:46 -0500 (EST)
From: quinot@gnat.com (Thomas Quinot)
Subject: [scripts] daemon (1.133) checked in scripts/nightly
Status: O
Content-Length: 74
Lines: 6


When installing the new compiler, also copy gnatlic.txt.
For F831-014




From quinot@adacore.com  Tue Nov 14 11:08:59 2006
Date: Tue, 14 Nov 2006 17:08:31 +0100
From: Thomas Quinot <quinot@adacore.com>
To: report@adacore.com
Message-ID: <20061114160831.GB88168@melamine.cuivre.fr.eu.org>
References: <20061113133517.GC28678@melamine.cuivre.fr.eu.org>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <20061113133517.GC28678@melamine.cuivre.fr.eu.org>
Subject: Re: F831-014 HEADS-UP: about to activate license file verification
Status: O
Content-Length: 799
Lines: 23

* Thomas Quinot, 2006-11-13 :

> I am about to enable verification of the license file in the compiler.
> A valid gnatlic.txt file is always available for the current wavefront
> in CVS; this file is updated nightly at the same time as gnatvsn.ads.
> It is installed automatically in the GNAT runtime directory
> (adainclude).

This is being delayed a bit as a mailserver run showed issues with the
current state of the patch (we are unable to load the license file if
command line flag -nostdinc is used, as we rely on the patch to the
runtime library to load it).

Any suggestions as to how to load a file from adainclude even when
-nostdinc is active?

Thomas.

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From charlet@ACT-Europe.FR  Tue Nov 14 11:11:25 2006
Date: Tue, 14 Nov 2006 17:11:24 +0100
From: Arnaud Charlet <charlet@adacore.com>
To: Thomas Quinot <quinot@adacore.com>
Cc: report@adacore.com
Message-ID: <20061114161124.GC79089@adacore.com>
References: <20061113133517.GC28678@melamine.cuivre.fr.eu.org> <20061114160831.GB88168@melamine.cuivre.fr.eu.org>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <20061114160831.GB88168@melamine.cuivre.fr.eu.org>
Subject: Re: F831-014 HEADS-UP: about to activate license file verification
Status: O
Content-Length: 441
Lines: 14

> Any suggestions as to how to load a file from adainclude even when
> -nostdinc is active?

Not really practical I'd say. You may want to disable the check when
-nostdinc is used.

Also, please keep in mind that the change should be easy to disable (not
sync) at the FSF. Currently the references to gnatlic.txt in Makefile.in
are already at the border line (i.e. too many place to patch/replace at
the FSF version of Makefile.in).

Arno



From quinot@adacore.com  Tue Nov 14 11:21:36 2006
Date: Tue, 14 Nov 2006 17:21:35 +0100
From: Thomas Quinot <quinot@adacore.com>
To: Arnaud Charlet <charlet@adacore.com>
Cc: report@adacore.com
Message-ID: <20061114162135.GA88428@melamine.cuivre.fr.eu.org>
References: <20061113133517.GC28678@melamine.cuivre.fr.eu.org> <20061114160831.GB88168@melamine.cuivre.fr.eu.org> <20061114161124.GC79089@adacore.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <20061114161124.GC79089@adacore.com>
Subject: Re: F831-014 HEADS-UP: about to activate license file verification
Status: O
Content-Length: 965
Lines: 28

* Arnaud Charlet, 2006-11-14 :

> > Any suggestions as to how to load a file from adainclude even when
> > -nostdinc is active?
> 
> Not really practical I'd say. You may want to disable the check when
> -nostdinc is used.

That's an option indeed.
 
> Also, please keep in mind that the change should be easy to disable (not
> sync) at the FSF. Currently the references to gnatlic.txt in Makefile.in
> are already at the border line (i.e. too many place to patch/replace at
> the FSF version of Makefile.in).

I don't think you need to remove these references, all you need is to
have a gnatlic.txt file at the FSF that just contains the GPL (or
whatever license text is appropriate). In the code, the checkin stuff
will remain isolated in instpar, and a single line will be added to
gnat1drv to call the check routine.

Thomas.

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From charlet@ACT-Europe.FR  Tue Nov 14 11:24:35 2006
Date: Tue, 14 Nov 2006 17:24:34 +0100
From: Arnaud Charlet <charlet@adacore.com>
To: Thomas Quinot <quinot@adacore.com>
Cc: report@adacore.com
Message-ID: <20061114162434.GC84961@adacore.com>
References: <20061113133517.GC28678@melamine.cuivre.fr.eu.org> <20061114160831.GB88168@melamine.cuivre.fr.eu.org> <20061114161124.GC79089@adacore.com> <20061114162135.GA88428@melamine.cuivre.fr.eu.org>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <20061114162135.GA88428@melamine.cuivre.fr.eu.org>
Subject: Re: F831-014 HEADS-UP: about to activate license file verification
Status: O
Content-Length: 674
Lines: 22

> I don't think you need to remove these references, all you need is to
> have a gnatlic.txt file at the FSF that just contains the GPL (or
> whatever license text is appropriate).

I do not want to do that because:

- this will make this very visible, and people will undoubetly question this
  change

- I do not know 'whatever license text is appropriate' for the FSF, nor do I
  want to take such responsibility or open this can of worms.

> will remain isolated in instpar, and a single line will be added to
> gnat1drv to call the check routine.

I guess two lines (a 'with' and a procedure call), so this part is fine.

The Makefile is still an issue though.

Arno



From quinot@adacore.com  Tue Nov 14 11:24:55 2006
Date: Tue, 14 Nov 2006 17:24:52 +0100
From: Thomas Quinot <quinot@adacore.com>
To: Arnaud Charlet <charlet@adacore.com>
Cc: report@adacore.com
Message-ID: <20061114162452.GB88428@melamine.cuivre.fr.eu.org>
References: <20061113133517.GC28678@melamine.cuivre.fr.eu.org> <20061114160831.GB88168@melamine.cuivre.fr.eu.org> <20061114161124.GC79089@adacore.com> <20061114162135.GA88428@melamine.cuivre.fr.eu.org>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <20061114162135.GA88428@melamine.cuivre.fr.eu.org>
Subject: Re: F831-014 HEADS-UP: about to activate license file verification
Status: O
Content-Length: 334
Lines: 12

* Thomas Quinot, 2006-11-14 :

> That's an option indeed.

(However that means that people can work around the restriction by using
-nostdinc + appropriate -I flags without rebuilding the compiler).

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From quinot@adacore.com  Tue Nov 14 11:25:22 2006
Date: Tue, 14 Nov 2006 17:25:20 +0100
From: Thomas Quinot <quinot@adacore.com>
To: Arnaud Charlet <charlet@adacore.com>
Cc: report@adacore.com
Message-ID: <20061114162520.GC88428@melamine.cuivre.fr.eu.org>
References: <20061113133517.GC28678@melamine.cuivre.fr.eu.org> <20061114160831.GB88168@melamine.cuivre.fr.eu.org> <20061114161124.GC79089@adacore.com> <20061114162135.GA88428@melamine.cuivre.fr.eu.org> <20061114162434.GC84961@adacore.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <20061114162434.GC84961@adacore.com>
Subject: Re: F831-014 HEADS-UP: about to activate license file verification
Status: O
Content-Length: 333
Lines: 15

* Arnaud Charlet, 2006-11-14 :

> I guess two lines (a 'with' and a procedure call), so this part is fine.

Correct.
 
> The Makefile is still an issue though.

I'm open to alternative suggestions.

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From charlet@ACT-Europe.FR  Tue Nov 14 11:26:05 2006
Date: Tue, 14 Nov 2006 17:26:04 +0100
From: Arnaud Charlet <charlet@adacore.com>
To: Thomas Quinot <quinot@adacore.com>
Cc: report@adacore.com
Message-ID: <20061114162604.GD84961@adacore.com>
References: <20061113133517.GC28678@melamine.cuivre.fr.eu.org> <20061114160831.GB88168@melamine.cuivre.fr.eu.org> <20061114161124.GC79089@adacore.com> <20061114162135.GA88428@melamine.cuivre.fr.eu.org> <20061114162452.GB88428@melamine.cuivre.fr.eu.org>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <20061114162452.GB88428@melamine.cuivre.fr.eu.org>
Subject: Re: F831-014 HEADS-UP: about to activate license file verification
Status: O
Content-Length: 266
Lines: 9

> (However that means that people can work around the restriction by using
> -nostdinc + appropriate -I flags without rebuilding the compiler).

Seems fine to me. People smart enough to read the sources and find this
hole are smart enough to rebuild anyway.

Arno



From charlet@ACT-Europe.FR  Tue Nov 14 11:26:59 2006
Date: Tue, 14 Nov 2006 17:26:54 +0100
From: Arnaud Charlet <charlet@adacore.com>
To: Thomas Quinot <quinot@adacore.com>
Cc: report@adacore.com
Message-ID: <20061114162654.GE84961@adacore.com>
References: <20061113133517.GC28678@melamine.cuivre.fr.eu.org> <20061114160831.GB88168@melamine.cuivre.fr.eu.org> <20061114161124.GC79089@adacore.com> <20061114162135.GA88428@melamine.cuivre.fr.eu.org> <20061114162434.GC84961@adacore.com> <20061114162520.GC88428@melamine.cuivre.fr.eu.org>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <20061114162520.GC88428@melamine.cuivre.fr.eu.org>
Subject: Re: F831-014 HEADS-UP: about to activate license file verification
Status: O
Content-Length: 151
Lines: 8

> I'm open to alternative suggestions.

I'm suggesting that you keep this issue in mind and think about a solution
(by possibly opening a TN).

Arno



From quinot@adacore.com  Tue Nov 14 11:35:18 2006
Date: Tue, 14 Nov 2006 17:35:17 +0100
From: Thomas Quinot <quinot@adacore.com>
To: Arnaud Charlet <charlet@adacore.com>
Cc: report@adacore.com
Message-ID: <20061114163517.GF88428@melamine.cuivre.fr.eu.org>
References: <20061113133517.GC28678@melamine.cuivre.fr.eu.org> <20061114160831.GB88168@melamine.cuivre.fr.eu.org> <20061114161124.GC79089@adacore.com> <20061114162135.GA88428@melamine.cuivre.fr.eu.org> <20061114162434.GC84961@adacore.com> <20061114162520.GC88428@melamine.cuivre.fr.eu.org> <20061114162654.GE84961@adacore.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <20061114162654.GE84961@adacore.com>
Subject: Re: F831-014 HEADS-UP: about to activate license file verification
Status: O
Content-Length: 427
Lines: 15

* Arnaud Charlet, 2006-11-14 :

> I'm suggesting that you keep this issue in mind and think about a solution
> (by possibly opening a TN).

How about:
 1. renaming gnatlic.txt to gnatlic.adp (or whatever letter)
 2. replacing all occurrences of (gnatlic.txt + *.ads + *.adb)
    by (*.ad?) ?

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From dewar@adacore.com  Tue Nov 14 11:37:08 2006
Message-ID: <4559F0AC.7040508@adacore.com>
Date: Tue, 14 Nov 2006 11:37:00 -0500
From: Robert Dewar <dewar@adacore.com>
MIME-Version: 1.0
To: Arnaud Charlet <charlet@adacore.com>
CC: Thomas Quinot <quinot@adacore.com>,  report@adacore.com
References: <20061113133517.GC28678@melamine.cuivre.fr.eu.org> <20061114160831.GB88168@melamine.cuivre.fr.eu.org> <20061114161124.GC79089@adacore.com> <20061114162135.GA88428@melamine.cuivre.fr.eu.org> <20061114162434.GC84961@adacore.com>
In-Reply-To: <20061114162434.GC84961@adacore.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: F831-014 HEADS-UP: about to activate license file verification
Status: O
Content-Length: 519
Lines: 16

Arnaud Charlet wrote:
>> I don't think you need to remove these references, all you need is to
>> have a gnatlic.txt file at the FSF that just contains the GPL (or
>> whatever license text is appropriate).
> 
> I do not want to do that because:
> 
> - this will make this very visible, and people will undoubetly question this
>   change
> 
> - I do not know 'whatever license text is appropriate' for the FSF, nor do I
>   want to take such responsibility or open this can of worms.

I strongly agree with Arno here



From charlet@ACT-Europe.FR  Tue Nov 14 11:56:32 2006
Date: Tue, 14 Nov 2006 17:56:31 +0100
From: Arnaud Charlet <charlet@adacore.com>
To: Thomas Quinot <quinot@adacore.com>
Cc: report@adacore.com
Message-ID: <20061114165631.GA96349@adacore.com>
References: <20061113133517.GC28678@melamine.cuivre.fr.eu.org> <20061114160831.GB88168@melamine.cuivre.fr.eu.org> <20061114161124.GC79089@adacore.com> <20061114162135.GA88428@melamine.cuivre.fr.eu.org> <20061114162434.GC84961@adacore.com> <20061114162520.GC88428@melamine.cuivre.fr.eu.org> <20061114162654.GE84961@adacore.com> <20061114163517.GF88428@melamine.cuivre.fr.eu.org>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <20061114163517.GF88428@melamine.cuivre.fr.eu.org>
Subject: Re: F831-014 HEADS-UP: about to activate license file verification
Status: O
Content-Length: 474
Lines: 17

> How about:
>  1. renaming gnatlic.txt to gnatlic.adp (or whatever letter)

.adp is the old GLIDe extension. Maybe .adl (license)

>  2. replacing all occurrences of (gnatlic.txt + *.ads + *.adb)
>     by (*.ad?) ?

That would take care of the install-gnatlib target, that's good.
There would be LIBGNAT_SRCS = and ../stamp-gnatlib2 left.

If we can get rid of stamp-gnatlib2, we can propbably live with
a single diff between FSF and AdaCore on LIBGNAT_SRCS = line

Arno



From quinot@adacore.com  Tue Nov 14 11:58:23 2006
Date: Tue, 14 Nov 2006 17:58:21 +0100
From: Thomas Quinot <quinot@adacore.com>
To: Arnaud Charlet <charlet@adacore.com>
Cc: report@adacore.com
Message-ID: <20061114165821.GA90383@melamine.cuivre.fr.eu.org>
References: <20061113133517.GC28678@melamine.cuivre.fr.eu.org> <20061114160831.GB88168@melamine.cuivre.fr.eu.org> <20061114161124.GC79089@adacore.com> <20061114162135.GA88428@melamine.cuivre.fr.eu.org> <20061114162434.GC84961@adacore.com> <20061114162520.GC88428@melamine.cuivre.fr.eu.org> <20061114162654.GE84961@adacore.com> <20061114163517.GF88428@melamine.cuivre.fr.eu.org> <20061114165631.GA96349@adacore.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <20061114165631.GA96349@adacore.com>
Subject: Re: F831-014 HEADS-UP: about to activate license file verification
Status: O
Content-Length: 648
Lines: 21

* Arnaud Charlet, 2006-11-14 :

> >  1. renaming gnatlic.txt to gnatlic.adp (or whatever letter)
> .adp is the old GLIDe extension. Maybe .adl (license)

OK

> That would take care of the install-gnatlib target, that's good.
> There would be LIBGNAT_SRCS = and ../stamp-gnatlib2 left.
> 
> If we can get rid of stamp-gnatlib2, we can propbably live with
> a single diff between FSF and AdaCore on LIBGNAT_SRCS = line

We might just want to get rid of the stamp-gnatlib2 line altogether, I'm
not sure it's useful.

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From quinot@gnat.com  Tue Nov 14 12:21:09 2006
To: file-ci@gnat.com
Message-Id: <20061114172108.EB6B348CC59@nile.gnat.com>
Date: Tue, 14 Nov 2006 12:21:08 -0500 (EST)
From: quinot@gnat.com (Thomas Quinot)
Subject: [gnat] gnatlic.txt (1.36) checked in gnat
Status: O
Content-Length: 60
Lines: 6


Remove gnatlic.txt, renamed to gnatlic.adl
For F831-014




From quinot@gnat.com  Tue Nov 14 12:21:30 2006
To: file-ci@gnat.com
Message-Id: <20061114172126.049DD48CDB5@nile.gnat.com>
Date: Tue, 14 Nov 2006 12:21:26 -0500 (EST)
From: quinot@gnat.com (Thomas Quinot)
Subject: [gnat] gnatlic.adl (1.36) checked in gnat
Status: O
Content-Length: 60
Lines: 7


Forced commit
Repo-copied from gnatlic.txt
For F831-014




From quinot@gnat.com  Tue Nov 14 12:22:34 2006
To: file-ci@gnat.com
Message-Id: <20061114172233.47FE748CBB1@nile.gnat.com>
Date: Tue, 14 Nov 2006 12:22:33 -0500 (EST)
From: quinot@gnat.com (Thomas Quinot)
Subject: [gnat] Makefile.in (1.2030) checked in gnat
Status: O
Content-Length: 69
Lines: 6


Account for renaming of gnatlic.txt to gnatlic.adl.
For F831-014




From quinot@gnat.com  Tue Nov 14 12:23:56 2006
To: file-ci@gnat.com
Message-Id: <20061114172348.035F548D029@nile.gnat.com>
Date: Tue, 14 Nov 2006 12:23:48 -0500 (EST)
From: quinot@gnat.com (Thomas Quinot)
Subject: [gnat] instpar.adb (1.11) checked in gnat
Status: O
Content-Length: 191
Lines: 10


Account for renaming of gnatlic.txt into gnatlic.adl.

No check when using -nostdinc as in that case we do not have practical
means of location the installation parameters.

For F831-014




From quinot@gnat.com  Tue Nov 14 12:24:21 2006
To: file-ci@gnat.com
Message-Id: <20061114172421.40BF248CC19@nile.gnat.com>
Date: Tue, 14 Nov 2006 12:24:21 -0500 (EST)
From: quinot@gnat.com (Thomas Quinot)
Subject: [gnat] MANIFEST.GNAT (1.307) checked in gnat
Status: O
Content-Length: 52
Lines: 6


Rename gnatlic.txt to gnatlic.adl.
For F831-014




From quinot@gnat.com  Tue Nov 14 12:25:32 2006
To: file-ci@gnat.com
Message-Id: <20061114172532.3417748CBB1@nile.gnat.com>
Date: Tue, 14 Nov 2006 12:25:32 -0500 (EST)
From: quinot@gnat.com (Thomas Quinot)
Subject: [gnat] Makefile.hie (1.84) checked in gnat
Status: O
Content-Length: 51
Lines: 6


Rename gnatlic.txt to gnatlic.adl
For F831-014




From quinot@gnat.com  Tue Nov 14 12:27:04 2006
To: file-ci@gnat.com
Message-Id: <20061114172703.B8BCC48CE3B@nile.gnat.com>
Date: Tue, 14 Nov 2006 12:27:03 -0500 (EST)
From: quinot@gnat.com (Thomas Quinot)
Subject: [ada] Makefile.in (1.1679.2.223:gcc-head) checked in ada
Status: O
Content-Length: 98
Lines: 7


Merge rev. 1.2030 from HEAD:
Account for renaming of gnatlic.txt to gnatlic.adl.
For F831-014




From quinot@gnat.com  Tue Nov 14 12:27:41 2006
To: file-ci@gnat.com
Message-Id: <20061114172740.1A8AB48CC19@nile.gnat.com>
Date: Tue, 14 Nov 2006 12:27:40 -0500 (EST)
From: quinot@gnat.com (Thomas Quinot)
Subject: [gnat] Makefile.in (1.1679.2.206.2.18:gcc-4_1) checked in gnat
Status: O
Content-Length: 98
Lines: 7


Merge rev. 1.2030 from HEAD:
Account for renaming of gnatlic.txt to gnatlic.adl.
For F831-014




From quinot@gnat.com  Tue Nov 14 12:30:32 2006
To: file-ci@gnat.com
Message-Id: <20061114173032.83D5748CC0D@nile.gnat.com>
Date: Tue, 14 Nov 2006 12:30:32 -0500 (EST)
From: quinot@gnat.com (Thomas Quinot)
Subject: [scripts] make-bin-5 (1.145) checked in scripts/nightly
Status: O
Content-Length: 51
Lines: 6


Rename gnatlic.txt to gnatlic.adl
For F831-014




From quinot@gnat.com  Tue Nov 14 12:30:34 2006
To: file-ci@gnat.com
Message-Id: <20061114173032.D156E48CFF2@nile.gnat.com>
Date: Tue, 14 Nov 2006 12:30:32 -0500 (EST)
From: quinot@gnat.com (Thomas Quinot)
Subject: [scripts] daemon (1.134) checked in scripts/nightly
Status: O
Content-Length: 51
Lines: 6


Rename gnatlic.txt to gnatlic.adl
For F831-014




From quinot@gnat.com  Tue Nov 14 12:30:33 2006
To: file-ci@gnat.com
Message-Id: <20061114173032.B110F48CE48@nile.gnat.com>
Date: Tue, 14 Nov 2006 12:30:32 -0500 (EST)
From: quinot@gnat.com (Thomas Quinot)
Subject: [scripts] cosrc (1.167) checked in scripts/nightly
Status: O
Content-Length: 51
Lines: 6


Rename gnatlic.txt to gnatlic.adl
For F831-014




From charlet@ACT-Europe.FR  Wed Nov 15 04:10:34 2006
Date: Wed, 15 Nov 2006 10:10:33 +0100
From: Arnaud Charlet <charlet@adacore.com>
To: Thomas Quinot <quinot@adacore.com>
Cc: report@adacore.com
Message-ID: <20061115091033.GA84954@adacore.com>
References: <20061113133517.GC28678@melamine.cuivre.fr.eu.org> <20061114160831.GB88168@melamine.cuivre.fr.eu.org> <20061114161124.GC79089@adacore.com> <20061114162135.GA88428@melamine.cuivre.fr.eu.org> <20061114162434.GC84961@adacore.com> <20061114162520.GC88428@melamine.cuivre.fr.eu.org> <20061114162654.GE84961@adacore.com> <20061114163517.GF88428@melamine.cuivre.fr.eu.org> <20061114165631.GA96349@adacore.com> <20061114165821.GA90383@melamine.cuivre.fr.eu.org>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <20061114165821.GA90383@melamine.cuivre.fr.eu.org>
Subject: Re: F831-014 HEADS-UP: about to activate license file verification
Status: O
Content-Length: 161
Lines: 8

> We might just want to get rid of the stamp-gnatlib2 line altogether, I'm
> not sure it's useful.

If it can be removed, that's certainly fine with me.

Arno



From quinot@gnat.com  Wed Nov 15 04:19:43 2006
To: file-ci@gnat.com
Message-Id: <20061115091943.091DB48D05D@nile.gnat.com>
Date: Wed, 15 Nov 2006 04:19:43 -0500 (EST)
From: quinot@gnat.com (Thomas Quinot)
Subject: [scripts] cosrc (1.168) checked in scripts/nightly
Status: O
Content-Length: 66
Lines: 6


Finish up renaming of gnatlic.txt to gnatlic.adl
For F831-014




From quinot@adacore.com  Wed Nov 15 05:03:55 2006
Date: Wed, 15 Nov 2006 11:03:53 +0100
From: Thomas Quinot <quinot@adacore.com>
To: Arnaud Charlet <charlet@adacore.com>
Cc: file@adacore.com
Message-ID: <20061115100353.GA24083@melamine.cuivre.fr.eu.org>
References: <20061114172740.43CC448CE3B@nile.gnat.com> <20061115095042.GB123@adacore.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <20061115095042.GB123@adacore.com>
Subject: Re: [gnat] Makefile.in (1.1679.2.206.2.18:gcc-4_1) checked in gnat F831-014
Status: O
Content-Length: 347
Lines: 12

* Arnaud Charlet, 2006-11-15 :

> I assume you'll also update the gcc-head version, right ?

Not clear to me why you are asking, this change has already been merged
on the gcc-head branch, see rev. 1.1679.2.223.

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From charlet@ACT-Europe.FR  Wed Nov 15 05:07:55 2006
Date: Wed, 15 Nov 2006 11:07:53 +0100
From: Arnaud Charlet <charlet@adacore.com>
To: Thomas Quinot <quinot@adacore.com>
Cc: file@adacore.com, Arnaud Charlet <charlet@adacore.com>
Message-ID: <20061115100753.GA6638@adacore.com>
References: <20061114172740.43CC448CE3B@nile.gnat.com> <20061115095042.GB123@adacore.com> <20061115100353.GA24083@melamine.cuivre.fr.eu.org>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <20061115100353.GA24083@melamine.cuivre.fr.eu.org>
Subject: Re: [gnat] Makefile.in (1.1679.2.206.2.18:gcc-4_1) checked in gnat - F831-014
Status: O
Content-Length: 336
Lines: 13

> > I assume you'll also update the gcc-head version, right ?
> 
> Not clear to me why you are asking, this change has already been merged
> on the gcc-head branch, see rev. 1.1679.2.223.

Because I didn't receive the corresponding email while I received the emails
for gcc-4_1 and HEAD branches.

Thanks for the clarification.

Arno



From quinot@adacore.com  Wed Nov 15 05:18:25 2006
Date: Wed, 15 Nov 2006 11:18:21 +0100
From: Thomas Quinot <quinot@adacore.com>
To: Arnaud Charlet <charlet@adacore.com>
Cc: file@adacore.com
Message-ID: <20061115101821.GC24083@melamine.cuivre.fr.eu.org>
References: <20061114172740.43CC448CE3B@nile.gnat.com> <20061115095042.GB123@adacore.com> <20061115100353.GA24083@melamine.cuivre.fr.eu.org> <20061115100753.GA6638@adacore.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <20061115100753.GA6638@adacore.com>
Subject: Re: [gnat] Makefile.in (1.1679.2.206.2.18:gcc-4_1) checked in gnat - F831-014
Status: O
Content-Length: 945
Lines: 20

* Arnaud Charlet, 2006-11-15 :

> Because I didn't receive the corresponding email while I received the emails
> for gcc-4_1 and HEAD branches.

Well that's a bit surprising, because all three messages were sent to
gci@gnat.com, and the one for the gcc-head checkin was delivered to your
account:

Nov 14 18:27:09 province postfix/cleanup[15313]: 3E9BB4ACE4: message-id=<20061114172703.BF9E848CE48@nile.gnat.com>
Nov 14 18:27:09 province postfix/qmgr[12758]: 3E9BB4ACE4: from=<report-internal@adacore.com>, size=3611, nrcpt=12 (queue active)
Nov 14 18:27:20 province postfix/local[16945]: 3E9BB4ACE4: to=<charlet@EU.adacore.com>, relay=local, delay=11, delays=0.16/0.01/0/11, dsn=2.0.0, status=sent (delivered to command: /usr/local/libexec/check-home "$USER" /usr/local/bin/procmail -a "$EXTENSION")

Thomas.

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From charlet@ACT-Europe.FR  Wed Nov 15 05:20:00 2006
Date: Wed, 15 Nov 2006 11:19:59 +0100
From: Arnaud Charlet <charlet@adacore.com>
To: Thomas Quinot <quinot@adacore.com>
Cc: file@adacore.com, Arnaud Charlet <charlet@adacore.com>
Message-ID: <20061115101959.GA12213@adacore.com>
References: <20061114172740.43CC448CE3B@nile.gnat.com> <20061115095042.GB123@adacore.com> <20061115100353.GA24083@melamine.cuivre.fr.eu.org> <20061115100753.GA6638@adacore.com> <20061115101821.GC24083@melamine.cuivre.fr.eu.org>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <20061115101821.GC24083@melamine.cuivre.fr.eu.org>
Subject: Re: [gnat] Makefile.in (1.1679.2.206.2.18:gcc-4_1) checked in gnat - F831-014
Status: O
Content-Length: 315
Lines: 10

> Well that's a bit surprising, because all three messages were sent to
> gci@gnat.com, and the one for the gcc-head checkin was delivered to your
> account:

Thanks for checking. I can now find this email, so that's my fault.
Do not know how I missed it the first time, I guess I hit the wrong "sort" key.

Arno



From quinot@malevil.act-europe.fr  Wed Nov 15 05:20:31 2006
Date: Wed, 15 Nov 2006 11:20:45 +0100 (CET)
From: Thomas Quinot <quinot@malevil.act-europe.fr>
Message-Id: <200611151020.kAFAKjnR018463@malevil.act-europe.fr>
To: file-ci@gnat.com
Subject: [cvs_script] mailfilter.pl (1.49) checked in cvs_script
Status: O
Content-Length: 124
Lines: 6


Limit size of diffs sent to mailing lists. Fixes FA31-007
Account for rename of gnatlic.txt to gnatlic.adl for F831-014




From quinot@adacore.com  Wed Nov 15 05:23:41 2006
Date: Wed, 15 Nov 2006 11:23:39 +0100
From: Thomas Quinot <quinot@adacore.com>
To: Arnaud Charlet <charlet@adacore.com>
Cc: file@adacore.com
Message-ID: <20061115102339.GE24083@melamine.cuivre.fr.eu.org>
References: <20061114172740.43CC448CE3B@nile.gnat.com> <20061115095042.GB123@adacore.com> <20061115100353.GA24083@melamine.cuivre.fr.eu.org> <20061115100753.GA6638@adacore.com> <20061115101821.GC24083@melamine.cuivre.fr.eu.org> <20061115101959.GA12213@adacore.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <20061115101959.GA12213@adacore.com>
Subject: Re: [gnat] Makefile.in (1.1679.2.206.2.18:gcc-4_1) checked in gnat - F831-014
Status: O
Content-Length: 617
Lines: 17

* Arnaud Charlet, 2006-11-15 :

> Thanks for checking. I can now find this email, so that's my fault.
> Do not know how I missed it the first time, I guess I hit the wrong "sort" key.

No problem, good to know that the message wasn't silently lost after
all. This might have been related to my checkin being tagged as 'ada' rather
than 'gnat' like the other two; I've now changed that in my checkout so
that all of my checkins come with the same module name in the future.

Thomas.

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From quinot@malevil.act-europe.fr  Thu Nov 16 04:20:27 2006
Date: Thu, 16 Nov 2006 10:20:42 +0100 (CET)
From: Thomas Quinot <quinot@malevil.act-europe.fr>
Message-Id: <200611160920.kAG9KgMk002638@malevil.act-europe.fr>
To: file@gnat.com
Subject: [F831-014] MANIFEST.GNAT(1.261.2.1) RH change
Status: O
Content-Length: 105
Lines: 6

Add gnatlic.txt in MANIFEST.GNAT
----
date: 2006/11/16 09:20:38;  author: quinot;
This is for F831-014



From quinot@gnat.com  Thu Nov 16 04:25:10 2006
To: file-ci@gnat.com
Message-Id: <20061116092510.1A2AB48CBBF@nile.gnat.com>
Date: Thu, 16 Nov 2006 04:25:10 -0500 (EST)
From: quinot@gnat.com (Thomas Quinot)
Subject: [gnat] MANIFEST.GNAT (1.261.2.2:gcc-3_2) checked in gnat
Status: O
Content-Length: 151
Lines: 8


Merge rev. 1.307 from HEAD onto gcc-3_2 branch:
Rename gnatlic.txt to gnatlic.adl, for F831-014

Move gnatlic.adl to the 'branched files' section.




From quinot@adacore.com  Thu Nov 16 04:27:04 2006
Date: Thu, 16 Nov 2006 10:27:03 +0100
From: Thomas Quinot <quinot@adacore.com>
To: Nicolas Roche <roche@adacore.com>
Cc: berrendo@adacore.com, charlet@gnat.com, ochem@adacore.com,
	schenker@gnat.com, villoing@adacore.com, file@adacore.com
Message-ID: <20061116092703.GA70938@melamine.cuivre.fr.eu.org>
References: <20061116020825.7BAA548CBAF@nile.gnat.com> <455C23B6.4080400@adacore.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <455C23B6.4080400@adacore.com>
Subject: Re: gnat (gcc-3_2) checkout failure [nile 6.0.0w 20061115] F831-014
Status: O
Content-Length: 462
Lines: 16

* Nicolas Roche, 2006-11-16 :

> Thomas can you take care of the fix.

Done. The reason why this did not show up earlier is because you had
merged the addition of gnatlic.txt on the gcc-3_2 branch, which I was
completely unaware of (the TN was not mentioned in the RH so there was
no filed reference to that checkin).

Thomas.

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From quinot@gnat.com  Tue Nov 21 08:39:08 2006
To: file-ci@gnat.com
Message-Id: <20061121133907.ECA7248CDC1@nile.gnat.com>
Date: Tue, 21 Nov 2006 08:39:07 -0500 (EST)
From: quinot@gnat.com (Thomas Quinot)
Subject: [ada] instpar.adb (1.13) checked in ada
Status: O
Content-Length: 48
Lines: 6


Add missing dependency on Opt.
For F831-014




From quinot@gnat.com  Tue Nov 21 08:39:48 2006
To: file-ci@gnat.com
Message-Id: <20061121133948.3572B48CC34@nile.gnat.com>
Date: Tue, 21 Nov 2006 08:39:48 -0500 (EST)
From: quinot@gnat.com (Thomas Quinot)
Subject: [ada] Make-lang.in (1.285) checked in ada
Status: O
Content-Length: 64
Lines: 6


Add instpar.o to the list of objects for gnat1
For F831-014




From quinot@gnat.com  Tue Nov 21 08:40:44 2006
To: file-ci@gnat.com
Message-Id: <20061121134043.C0F8548CFA9@nile.gnat.com>
Date: Tue, 21 Nov 2006 08:40:43 -0500 (EST)
From: quinot@gnat.com (Thomas Quinot)
Subject: [ada] gnat1drv.adb (1.189) checked in ada
Status: O
Content-Length: 148
Lines: 8


Enable verification of the installation parameters file (gnatlic.adl)
in gnat1drv.
For F831-014
This change is not to be propagated to the FSF.




From quinot@adacore.com  Tue Nov 21 08:42:55 2006
Date: Tue, 21 Nov 2006 14:42:53 +0100
From: Thomas Quinot <quinot@adacore.com>
To: report@adacore.com
Message-ID: <20061121134253.GC18014@melamine.cuivre.fr.eu.org>
References: <20061113133517.GC28678@melamine.cuivre.fr.eu.org>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <20061113133517.GC28678@melamine.cuivre.fr.eu.org>
Subject: F831-014 HEADS-UP: license file verification ENABLED
Status: O
Content-Length: 890
Lines: 26

* Thomas Quinot, 2006-11-13 :

> I am about to enable verification of the license file in the compiler.
> A valid gnatlic.txt file is always available for the current wavefront
> in CVS; this file is updated nightly at the same time as gnatvsn.ads.
> It is installed automatically in the GNAT runtime directory
> (adainclude).

This is now activated (gnat1drv.adb rev. 1.189).

Note that gnatlic.txt is now named gnatlic.adl.

Again, if this is blocking your work, all you need to get a compiler
without license file verification is to comment out the call to
Check_Installation_Parameters (and the dependency on Instpar) in
gnat1drv.adb. If you need to do that, I'd appreciate if you could let me
know so I can fix the problem properly.

Thanks,
Thomas.

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From quinot@adacore.com  Tue Nov 21 08:43:39 2006
Date: Tue, 21 Nov 2006 14:43:37 +0100
From: Thomas Quinot <quinot@adacore.com>
To: friess@adacore.com, gasperoni@adacore.com, file@adacore.com
Message-ID: <20061121134337.GD18014@melamine.cuivre.fr.eu.org>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Subject: F831-014 [quinot@gnat.com: [ada] gnat1drv.adb (1.189) checked in ada]
Status: O
Content-Length: 1485
Lines: 55

Verification of the license file is now enabled in gnat1.

----- Forwarded message from Thomas Quinot <quinot@gnat.com> -----

Old-To: gci@gnat.com
Subject: [ada] gnat1drv.adb (1.189) checked in ada
X-ACT-Mailfilter: check-in in ada
X-ACT-GTI: http://www.adacore.com/gti/tn.html?View_TN=F831-014
Date: Tue, 21 Nov 2006 08:40:43 -0500 (EST)
From: Thomas Quinot <quinot@gnat.com>
X-ACT-List: gci
X-ACT-TN: 
Reply-To: quinot@gnat.com
To: gci-do-not-reply@adacore.com


Enable verification of the installation parameters file (gnatlic.adl)
in gnat1drv.
For F831-014
This change is not to be propagated to the FSF.


Index: ada/gnat1drv.adb
===================================================================
RCS file: gnat1drv.adb,v
retrieving revision 1.188
retrieving revision 1.189
diff -u -p -r1.188 -r1.189
--- gnat1drv.adb	2006/11/10 19:33:54	1.188
+++ gnat1drv.adb	2006/11/21 13:40:43	1.189
@@ -38,6 +38,7 @@ with Frontend;
 with Gnatvsn;  use Gnatvsn;
 with Hostparm;
 with Inline;
+with Instpar;  use Instpar;
 with Lib;      use Lib;
 with Lib.Writ; use Lib.Writ;
 with Lib.Xref;
@@ -119,6 +120,8 @@ begin
       Sem_Eval.Initialize;
       Sem_Type.Init_Interp_Tables;
 
+      Check_Installation_Parameters;
+
       --  Acquire target parameters from system.ads (source of package System)
 
       declare

----- End forwarded message -----

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From charlet@ACT-Europe.FR  Tue Nov 21 09:00:45 2006
Date: Tue, 21 Nov 2006 15:00:28 +0100
From: Arnaud Charlet <charlet@adacore.com>
To: Thomas Quinot <quinot@gnat.com>
Cc: Arnaud Charlet <charlet@adacore.com>, file@adacore.com
Message-ID: <20061121140028.GA82142@adacore.com>
References: <20061121133948.3BBBA48CDC1@nile.gnat.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <20061121133948.3BBBA48CDC1@nile.gnat.com>
Subject: Re: [ada] Make-lang.in (1.285) checked in ada - F831-014
Status: O
Content-Length: 336
Lines: 10

You will need to update also the gcc-3_4 and gcc-head branches.

Also note that this is another annoying differences that cannot be merged
at the FSF which will likely cause conflicts in the future, so would be good to
find another way to add this object file, thanks.

> Add instpar.o to the list of objects for gnat1
> For F831-014



From quinot@adacore.com  Tue Nov 21 09:25:24 2006
Date: Tue, 21 Nov 2006 15:25:23 +0100
From: Thomas Quinot <quinot@adacore.com>
To: Arnaud Charlet <charlet@adacore.com>
Cc: report@adacore.com
Message-ID: <20061121142523.GB20170@melamine.cuivre.fr.eu.org>
References: <20061121133948.3BBBA48CDC1@nile.gnat.com> <20061121140028.GA82142@adacore.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <20061121140028.GA82142@adacore.com>
Subject: Re: [ada] Make-lang.in (1.285) checked in ada - F831-014
Status: O
Content-Length: 543
Lines: 15

* Arnaud Charlet, 2006-11-21 :

> Also note that this is another annoying differences that cannot be merged
> at the FSF which will likely cause conflicts in the future, so would be good to
> find another way to add this object file, thanks.

Well let's ask everyone for ideas, because I for one do not have the
faintest idea as to how to include an object file in gnat1 without
mentioning it to the linker.

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From quinot@gnat.com  Tue Nov 21 09:26:23 2006
To: file-ci@gnat.com
Message-Id: <20061121142622.E6BB248CFCB@nile.gnat.com>
Date: Tue, 21 Nov 2006 09:26:22 -0500 (EST)
From: quinot@gnat.com (Thomas Quinot)
Subject: [gnat] Make-lang.in (1.131.2.186:gcc-head) checked in gnat
Status: O
Content-Length: 92
Lines: 7


Merge rev. 1.285 from HEAD:
Add instpar.o to the list of objects for gnat1
For F831-014




From quinot@gnat.com  Tue Nov 21 09:28:32 2006
To: file-ci@gnat.com
Message-Id: <20061121142832.1966B48CCB6@nile.gnat.com>
Date: Tue, 21 Nov 2006 09:28:32 -0500 (EST)
From: quinot@gnat.com (Thomas Quinot)
Subject: [gnat] Make-lang.in (1.149.2.11:gcc-3_4) checked in gnat
Status: O
Content-Length: 92
Lines: 7


Merge rev. 1.285 from HEAD:
Add instpar.o to the list of objects for gnat1
For F831-014




From kenner@vlsi1.ultra.nyu.edu  Tue Nov 21 09:30:06 2006
From: kenner@vlsi1.ultra.nyu.edu (Richard Kenner)
Message-Id: <10611211434.AA24118@vlsi1.ultra.nyu.edu>
Date: Tue, 21 Nov 2006 09:34:30 EST
To: quinot@adacore.com
Cc: charlet@adacore.com, report@adacore.com
In-Reply-To: <20061121142523.GB20170@melamine.cuivre.fr.eu.org>
References: <20061121133948.3BBBA48CDC1@nile.gnat.com>
    <20061121140028.GA82142@adacore.com>
    <20061121142523.GB20170@melamine.cuivre.fr.eu.org>
Subject: Re: [ada] Make-lang.in (1.285) checked in ada - F831-014
Status: O
Content-Length: 362
Lines: 9

> Well let's ask everyone for ideas, because I for one do not have the
> faintest idea as to how to include an object file in gnat1 without
> mentioning it to the linker.

Why do we have a separate object file?  I thought that the idea was that
we already had one file (I forget which) that's different than at the
FSF and we'd put everything needed in there.



From charlet@ACT-Europe.FR  Tue Nov 21 09:36:11 2006
Date: Tue, 21 Nov 2006 15:36:06 +0100
From: Arnaud Charlet <charlet@adacore.com>
To: Richard Kenner <kenner@vlsi1.ultra.nyu.edu>
Cc: quinot@adacore.com, report@adacore.com
Message-ID: <20061121143606.GA93622@adacore.com>
References: <20061121133948.3BBBA48CDC1@nile.gnat.com> <20061121140028.GA82142@adacore.com> <20061121142523.GB20170@melamine.cuivre.fr.eu.org> <10611211434.AA24118@vlsi1.ultra.nyu.edu>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <10611211434.AA24118@vlsi1.ultra.nyu.edu>
Subject: Re: [ada] Make-lang.in (1.285) checked in ada - F831-014
Status: O
Content-Length: 286
Lines: 10

> Why do we have a separate object file?  I thought that the idea was that
> we already had one file (I forget which) that's different than at the
> FSF and we'd put everything needed in there.

Indeed, it's gnatvsn, and it makes lots of sense to put the check in this
package.

Arno



From quinot@adacore.com  Tue Nov 21 09:39:27 2006
Date: Tue, 21 Nov 2006 15:39:25 +0100
From: Thomas Quinot <quinot@adacore.com>
To: Richard Kenner <kenner@vlsi1.ultra.nyu.edu>
Cc: charlet@adacore.com, report@adacore.com
Message-ID: <20061121143925.GA20724@melamine.cuivre.fr.eu.org>
References: <20061121133948.3BBBA48CDC1@nile.gnat.com> <20061121140028.GA82142@adacore.com> <20061121142523.GB20170@melamine.cuivre.fr.eu.org> <10611211434.AA24118@vlsi1.ultra.nyu.edu>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <10611211434.AA24118@vlsi1.ultra.nyu.edu>
Subject: Re: [ada] Make-lang.in (1.285) checked in ada - F831-014
Status: O
Content-Length: 433
Lines: 15

* Richard Kenner, 2006-11-21 :

> Why do we have a separate object file?  I thought that the idea was that
> we already had one file (I forget which) that's different than at the
> FSF and we'd put everything needed in there.

No, it's a package that's *not present* at the FSF (instpar).

Thomas.

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From quinot@adacore.com  Tue Nov 21 09:40:21 2006
Date: Tue, 21 Nov 2006 15:40:20 +0100
From: Thomas Quinot <quinot@adacore.com>
To: Arnaud Charlet <charlet@adacore.com>
Cc: Richard Kenner <kenner@vlsi1.ultra.nyu.edu>, report@adacore.com
Message-ID: <20061121144020.GB20724@melamine.cuivre.fr.eu.org>
References: <20061121133948.3BBBA48CDC1@nile.gnat.com> <20061121140028.GA82142@adacore.com> <20061121142523.GB20170@melamine.cuivre.fr.eu.org> <10611211434.AA24118@vlsi1.ultra.nyu.edu> <20061121143606.GA93622@adacore.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <20061121143606.GA93622@adacore.com>
Subject: Re: [ada] Make-lang.in (1.285) checked in ada - F831-014
Status: O
Content-Length: 383
Lines: 15

* Arnaud Charlet, 2006-11-21 :

> Indeed, it's gnatvsn, and it makes lots of sense to put the check in this
> package.

Hmmm, interesting idea, thanks!

So the suggestion would be to move all of instpar into our internal
version of gnatvsn, right?

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From charlet@ACT-Europe.FR  Tue Nov 21 13:27:12 2006
Date: Tue, 21 Nov 2006 19:27:11 +0100
From: Arnaud Charlet <charlet@adacore.com>
To: Thomas Quinot <quinot@adacore.com>
Cc: Richard Kenner <kenner@vlsi1.ultra.nyu.edu>, report@adacore.com
Message-ID: <20061121182711.GA88574@adacore.com>
References: <20061121133948.3BBBA48CDC1@nile.gnat.com> <20061121140028.GA82142@adacore.com> <20061121142523.GB20170@melamine.cuivre.fr.eu.org> <10611211434.AA24118@vlsi1.ultra.nyu.edu> <20061121143606.GA93622@adacore.com> <20061121144020.GB20724@melamine.cuivre.fr.eu.org>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <20061121144020.GB20724@melamine.cuivre.fr.eu.org>
Subject: Re: [ada] Make-lang.in (1.285) checked in ada - F831-014
Status: O
Content-Length: 115
Lines: 8

> So the suggestion would be to move all of instpar into our internal
> version of gnatvsn, right?

Right.

Arno



From quinot@adacore.com  Tue Nov 21 17:21:50 2006
Date: Tue, 21 Nov 2006 23:21:44 +0100
From: Thomas Quinot <quinot@adacore.com>
To: Arnaud Charlet <charlet@adacore.com>
Cc: Richard Kenner <kenner@vlsi1.ultra.nyu.edu>, report@adacore.com
Message-ID: <20061121222144.GB37311@melamine.cuivre.fr.eu.org>
References: <20061121133948.3BBBA48CDC1@nile.gnat.com> <20061121140028.GA82142@adacore.com> <20061121142523.GB20170@melamine.cuivre.fr.eu.org> <10611211434.AA24118@vlsi1.ultra.nyu.edu> <20061121143606.GA93622@adacore.com> <20061121144020.GB20724@melamine.cuivre.fr.eu.org> <20061121182711.GA88574@adacore.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <20061121182711.GA88574@adacore.com>
Subject: Re: [ada] Make-lang.in (1.285) checked in ada - F831-014
Status: O
Content-Length: 536
Lines: 16

* Arnaud Charlet, 2006-11-21 :

> > So the suggestion would be to move all of instpar into our internal
> > version of gnatvsn, right?
> Right.

OK, as discussed live this looks like a very good solution as this file
had already different versions in our repo and the FSF one, and so this
will avoid having a unit that is present in one version and not the
other. I'll implement this change tomorrow.

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From quinot@adacore.com  Wed Nov 22 04:26:27 2006
Date: Wed, 22 Nov 2006 10:26:25 +0100
From: Thomas Quinot <quinot@adacore.com>
To: Arnaud Charlet <charlet@adacore.com>
Cc: dewar@adacore.com, Richard Kenner <kenner@vlsi1.ultra.nyu.edu>,
	report@adacore.com
Message-ID: <20061122092625.GA58050@melamine.cuivre.fr.eu.org>
References: <20061121133948.3BBBA48CDC1@nile.gnat.com> <20061121140028.GA82142@adacore.com> <20061121142523.GB20170@melamine.cuivre.fr.eu.org> <10611211434.AA24118@vlsi1.ultra.nyu.edu> <20061121143606.GA93622@adacore.com> <20061121144020.GB20724@melamine.cuivre.fr.eu.org> <20061121182711.GA88574@adacore.com> <20061121222144.GB37311@melamine.cuivre.fr.eu.org>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <20061121222144.GB37311@melamine.cuivre.fr.eu.org>
Subject: Re: [ada] Make-lang.in (1.285) checked in ada - F831-014
Status: O
Content-Length: 769
Lines: 22

* Thomas Quinot, 2006-11-21 :

> OK, as discussed live this looks like a very good solution as this file
> had already different versions in our repo and the FSF one, and so this
> will avoid having a unit that is present in one version and not the
> other. I'll implement this change tomorrow.

This is proving more complicated than expected:

gnatvsn.adb:37:06: warning: license of withed unit "Sdefault" is incompatible
gnatvsn.adb:39:12: warning: license of withed unit "L" is incompatible

Is there a fundamental reason why we would want to keep Gnatvsn a GMGPL
unit? If so, can we make Sdefault and Sinput.L GMGPL too?

Thomas.

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From comar@adacore.com  Wed Nov 22 06:18:30 2006
Message-ID: <45643206.8030500@adacore.com>
Date: Wed, 22 Nov 2006 12:18:30 +0100
From: Cyrille Comar <comar@adacore.com>
MIME-Version: 1.0
To: Thomas Quinot <quinot@adacore.com>
CC: Arnaud Charlet <charlet@adacore.com>,  dewar@adacore.com, 
 Richard Kenner <kenner@vlsi1.ultra.nyu.edu>,
  report@adacore.com
References: <20061121133948.3BBBA48CDC1@nile.gnat.com> <20061121140028.GA82142@adacore.com> <20061121142523.GB20170@melamine.cuivre.fr.eu.org> <10611211434.AA24118@vlsi1.ultra.nyu.edu> <20061121143606.GA93622@adacore.com> <20061121144020.GB20724@melamine.cuivre.fr.eu.org> <20061121182711.GA88574@adacore.com> <20061121222144.GB37311@melamine.cuivre.fr.eu.org> <20061122092625.GA58050@melamine.cuivre.fr.eu.org>
In-Reply-To: <20061122092625.GA58050@melamine.cuivre.fr.eu.org>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Subject: Re: [ada] Make-lang.in (1.285) checked in ada - F831-014
Status: O
Content-Length: 457
Lines: 13

Thomas Quinot a crit :
> Is there a fundamental reason why we would want to keep Gnatvsn a GMGPL
> unit? 

The fundamental reason, if any, would be that it is with'ed by a unit 
that needs to be GMGPL because of ASIS (such as opt.ad? for instance)

I wonder why we don't play more with license switching here. Couldn't we 
have all the frontend sources pure GPL in the context of the compiler 
and all the required sources GMGPL in the context of ASIS?




From charlet@ACT-Europe.FR  Wed Nov 22 09:44:46 2006
Date: Wed, 22 Nov 2006 15:44:44 +0100
From: Arnaud Charlet <charlet@adacore.com>
To: Cyrille Comar <comar@adacore.com>
Cc: Thomas Quinot <quinot@adacore.com>, dewar@adacore.com,
	Richard Kenner <kenner@vlsi1.ultra.nyu.edu>, report@adacore.com
Message-ID: <20061122144444.GB96936@adacore.com>
References: <20061121133948.3BBBA48CDC1@nile.gnat.com> <20061121140028.GA82142@adacore.com> <20061121142523.GB20170@melamine.cuivre.fr.eu.org> <10611211434.AA24118@vlsi1.ultra.nyu.edu> <20061121143606.GA93622@adacore.com> <20061121144020.GB20724@melamine.cuivre.fr.eu.org> <20061121182711.GA88574@adacore.com> <20061121222144.GB37311@melamine.cuivre.fr.eu.org> <20061122092625.GA58050@melamine.cuivre.fr.eu.org> <45643206.8030500@adacore.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <45643206.8030500@adacore.com>
Subject: Re: [ada] Make-lang.in (1.285) checked in ada - F831-014
Status: O
Content-Length: 257
Lines: 9

> I wonder why we don't play more with license switching here. Couldn't we 
> have all the frontend sources pure GPL in the context of the compiler 
> and all the required sources GMGPL in the context of ASIS?

Right, that would seem cleaner to me.

Arno



From quinot@adacore.com  Thu Nov 23 04:45:03 2006
Date: Thu, 23 Nov 2006 10:45:02 +0100
From: Thomas Quinot <quinot@adacore.com>
To: Cyrille Comar <comar@adacore.com>
Cc: Arnaud Charlet <charlet@adacore.com>, dewar@adacore.com,
	Richard Kenner <kenner@vlsi1.ultra.nyu.edu>, report@adacore.com
Message-ID: <20061123094502.GB9174@melamine.cuivre.fr.eu.org>
References: <20061121133948.3BBBA48CDC1@nile.gnat.com> <20061121140028.GA82142@adacore.com> <20061121142523.GB20170@melamine.cuivre.fr.eu.org> <10611211434.AA24118@vlsi1.ultra.nyu.edu> <20061121143606.GA93622@adacore.com> <20061121144020.GB20724@melamine.cuivre.fr.eu.org> <20061121182711.GA88574@adacore.com> <20061121222144.GB37311@melamine.cuivre.fr.eu.org> <20061122092625.GA58050@melamine.cuivre.fr.eu.org> <45643206.8030500@adacore.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <45643206.8030500@adacore.com>
Subject: Re: [ada] Make-lang.in (1.285) checked in ada - F831-014
Status: O
Content-Length: 472
Lines: 14

* Cyrille Comar, 2006-11-22 :

> I wonder why we don't play more with license switching here. Couldn't we 
> have all the frontend sources pure GPL in the context of the compiler 
> and all the required sources GMGPL in the context of ASIS?

That looks like a significant amount of work so I'd rather find a short
term solution for now.

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From quinot@adacore.com  Thu Nov 23 10:18:20 2006
Date: Thu, 23 Nov 2006 16:18:19 +0100
From: Thomas Quinot <quinot@adacore.com>
To: Robert Dewar <dewar@adacore.com>
Cc: Eric Botcazou <ebotcazou@adacore.com>,
	Vincent Celier <celier@adacore.com>,
	Automatic Filer <file@adacore.com>
Message-ID: <20061123151819.GA21983@melamine.cuivre.fr.eu.org>
References: <20061123055227.GA8663@adacore.com> <200611231547.40722.ebotcazou@adacore.com> <20061123144709.GB19956@melamine.cuivre.fr.eu.org> <200611231556.17928.ebotcazou@adacore.com> <20061123151319.GB21675@melamine.cuivre.fr.eu.org> <4565BAF5.9060708@adacore.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <4565BAF5.9060708@adacore.com>
Subject: Re: fixed+noms testsuite [red 6.0.0w 20061122-41 x86_64-pc-linux-gnu] (DB26-008) F831-014
Status: O
Content-Length: 323
Lines: 12

* Robert Dewar, 2006-11-23 :

> this has always been the case till now. I think I would load system.ads
> before the license file to avoid any difficulties.

Sounds easy, mailserving now.

-- 
Thomas Quinot, Ph.D. ** quinot@adacore.com ** Senior Software Engineer
               AdaCore -- Paris, France -- New York, USA



From quinot@gnat.com  Thu Nov 23 12:17:04 2006
To: file-ci@gnat.com
Message-Id: <20061123171704.7D13348CC01@nile.gnat.com>
Date: Thu, 23 Nov 2006 12:17:04 -0500 (EST)
From: quinot@gnat.com (Thomas Quinot)
Subject: [gnat] gnat1drv.adb (1.190) checked in gnat
Status: O
Content-Length: 318
Lines: 10


Check installation parameters file only after loading target parameters
from system.ads.
Fixes DB26-008 pseudo-regression caused by F831-014 change
Also ensures that when adainclude is missing altogether, we get the usual
message about an installation problem rather than an installation
parameters check failure.




From quinot@gnat.com  Mon Nov 27 05:59:11 2006
To: file-ci@gnat.com
Message-Id: <20061127105911.0B3F148CC43@nile.gnat.com>
Date: Mon, 27 Nov 2006 05:59:11 -0500 (EST)
From: quinot@gnat.com (Thomas Quinot)
Subject: [scripts] make-eval-gnatlic (1.1) checked in scripts/nightly
Status: O
Content-Length: 79
Lines: 6


New script to generate a GNAT Pro eval license file manually.
For F831-014




From quinot@gnat.com  Mon Nov 27 12:12:17 2006
To: file-ci@gnat.com
Message-Id: <20061127171152.D2A5848CC4E@nile.gnat.com>
Date: Mon, 27 Nov 2006 12:11:52 -0500 (EST)
From: quinot@gnat.com (Thomas Quinot)
Subject: [scripts] make-eval-gnatlic (1.2) checked in scripts/nightly
Status: O
Content-Length: 250
Lines: 10


Redirect tar output to /dev/null so that we get exactly the new license
file on stdout.
Account for the fact that a GNAT Pro distribution may contain several
(identical) copies of the default gnatlic.adl file (one for each
runtime).
For F831-014




From quinot@gnat.com  Thu Dec 14 11:42:55 2006
From: quinot@gnat.com
To: file@gnat.com
Message-Id: <20061214164254.4AB7B48CD93@nile.gnat.com>
Date: Thu, 14 Dec 2006 11:42:54 -0500 (EST)
Subject: [bugtool:F831-014] Moved to dead
Status: O
Content-Length: 9
Lines: 4

Why: 




