Cairo bindings and e-mail license virus bombs

Cairo bindings and e-mail license virus bombs

[David Botton]

I removed my modified Cairo bindings from Gnoga (Gnoga doesn't use Cairo internally) and have made it available as a separate archive if anyone would like to use it to generate pngs, pdfs and svgs. (Gnoga not required either).

The modification is simple, it just removes dependencies for GtkAda (cairo never required Gtk) and as per the example cairo_test you only need to link in -lcairo.

It is available in the Gnoga.com "marketplace" (although just a link there for now since I have not written the market place app yet).

I do not plan on using the Cairo bindings from AdaCore nor any of the AdaCore available bindings since finding this in the GNU/Debian distribution of gcc/ada.

/usr/share/doc/libgtkada2.24.4
[more before]
Comment: All the content available on libre.adacore.com is licensed
 under the terms of the pure GPL, despite the fact that AdaCore have
 not yet adjusted the licensing boilerplate in documentation.
 .
 This is stated for example at
 http://lists.adacore.com/pipermail/gtkada/2009-June/003789.html. The
 actual license is a decision of AdaCore. [more after]


If it is possible for a company to claim that they can change the licensing on their files (_even_ if the files say inside every one that they are GMGPL) _just_ because downloaded from a particular URL (or even if another file some place on a site) than this calls in to question not just the company's motivations but potential long term issues

 _EVEN FOR CUSTOMERS_

 since they will need to prove they have not updated or re-downloaded files from the repos or URLs under the virus hosts if their contracts lapse or do not have an anti-virus clause in the contract. (Remember that any one that received the exact same copies of the files from another "location" on the internet, say a customer or even a friend of a customer, would have the _legal right_ granted by the GMPLG to give you a copy with the GMGPL, etc.) 

I think it is important to make sure this becomes public because the long term damage to the entire free software community, not just Ada advocacy, is HUGE. If no one can trust the headers and specs to represent the license of a file when downloaded from source public repos, etc, how do you know that your copy is legal at all? This act, that an e-mail or simple statement at some place of questionable publicity changes all the licenses even though published on individual files, by AdaCore has potential to _tear down_ the entire Open Source world.

Does this mean that the FSF can one day just e-mail everyone that BTW everything you download from any FSF source is now no longer GPL or some new license that removes run time extensions after the fact that now requires proprietary software to open source their code if recompiled again on their own computers with the now e-mailed virus bomb? Does this mean that every library under the BSD or MIT license now require a notary public and contracts to insure that the website they were downloaded from them isn't a virus host?

AdaCore has the right to produce GPL only libraries and bindings. It is anti-Ada and doesn't help Free or Open Source either, but that is fine, their right.

What they don't have the right to do is spread false information and muddy the waters for their own corporate benefit.

Take this statement from the AdaCore website:
  (http://libre.adacore.com/tools/gnat-gpl-edition/faq/#other)

<<AdaCore has no control over these and cannot guarantee their quality or suitability for a particular purpose.>>

- Legitimate statement

<<Most important you should ascertain the license and IPR (Intellectual Property Rights) guarantees from its provider.

- Scare tactic (Using the term IPR is very anti-FSF... but whatever)

AdaCore offers top quality support and products. To harm Ada advocacy for corporate benefits, to harm Open Source efforts for corporate benefits, to use scare tactics to maintain a customer base, etc. These are _below_ the true quality of what AdaCore produces in code, the excellent top notch support offered, and dirties any one near Ada and AdaCore by extension.

If AdaCore wants to encumber and virus all their code do it, but don't muddy waters and use scare tactics that damage the Ada and Open Source communities.

The snooty attitude expressed by some should be replaced with a contrite attitude of service leaders who deserve respect from accomplishment that is more befitting the station AdaCore once had and should have for the hard work done on the compiler past and present.

David Botton

Re: Cairo bindings and e-mail license virus bombs

[David Botton]

The content of the e-mail link shout it change in the future is:

> Many of the GtkAda source files contain the text of the GMGPL  
> exception; some do not. Are you saying I should not rely on the  
> presence or absence of the exception in the individual source texts  
> obtained from this repository:

That's correct, headers on source file have no/little legal significance.
The main and simple message is that all the software available
via the libre site, including svn.eu.adacore.com/anonsvn is provided
under the pure GPL license.

Arno

Re: Cairo bindings and e-mail license virus bombs

[sbelmont700]

A license is not an inherent part of a file (any file); the license is something you and the vendor agree to, whether via a webpage, handshake, signed contract, or whatever.  Binary libraries, for example, don't have any licensing information 'in' them, yet whatever terms you and vendor come to is what's in effect, essentially forever, assuming that it's legal in the first place.  Nobody can change this after the fact.

As far as needing to know whether what you are given was the vendors to legally give you, regardless of the license, is like AdaCore says: your responsibility.  So yes, if FSF did something unscrupulous in the first place, then anything that uses FSF files is called into question.  But everything in life is like this, not just software; if a man on a street corner sells you a 10 carat diamond ring for $39.95, and it turns out to be stolen, you have to give the ring back to the owner (but, at the same time, have ground to sue the seller for recompense).

-sb

Re: Cairo bindings and e-mail license virus bombs

[David Botton]

On Wednesday, December 10, 2014 1:58:34 PM UTC-5, sbelm...@gmail.com wrote:
> A license is not an inherent part of a file (any file); the license is something you and the vendor agree to, whether via a webpage, handshake, signed contract, or whatever. 

I am not questioning that. The points given are:

1 The file states a license and then random statements raise questions on it.
2 The purposeful muddying clarity of license for corporate benefit at the expense of others

I doubt legally anyone using source files that state a license clearly as these do can be called to question that they are in violation of another license not stated if fulfilling the terms of the stated license in the file.

While it would be nice if AdaCore released their code libraries under GMGPL, I don't care if they do or do not, others will come along and just write new ones if need be. What needs to change is purposeful muddying of waters and clarity through out to stop damaging Ada advocacy (and open source) for one company's personal benefit.

David Botton

Re: Cairo bindings and e-mail license virus bombs

[sbelmont700]

On Wednesday, December 10, 2014 2:09:03 PM UTC-5, David Botton wrote:
> I doubt legally anyone using source files that state a license clearly as these do can be called to question that they are in violation of another license not stated if fulfilling the terms of the stated license in the file.
> 

Of course they can.  If I take the packages from, say, Green Hills, strip off whatever proprietary header they have, put on the GMGPL header, and start passing them around here for free, they aren't suddenly free for the whole world to use in perpetuity.  I get sued into oblivion for violating my license with Green Hills, and you get a threatening letter stating that you have to either stop using those stolen files or else buy a legitimate copy.  You don't get to keep stolen goods, even if you acquired them unknowingly and in good faith.

Re: Cairo bindings and e-mail license virus bombs

[Dmitry A. Kazakov]

On Wed, 10 Dec 2014 10:32:51 -0800 (PST), David Botton wrote:

> I do not plan on using the Cairo bindings from AdaCore nor any of the
> AdaCore available bindings since finding this in the GNU/Debian
> distribution of gcc/ada.
> 
> /usr/share/doc/libgtkada2.24.4

David, if you are OK with using outdated Cairo (GTK 2.x) then here are
bindings developed by Damien Carbonne:

http://sourceforge.net/projects/cairoada/

They were better designed than AdaCore's, higher-level, packing ugly
objects in controlled handle types.

-- 
Regards,
Dmitry A. Kazakov
http://www.dmitry-kazakov.de

Re: Cairo bindings and e-mail license virus bombs

[Simon Wright]

sbelmont700@gmail.com writes:

> A license is not an inherent part of a file (any file); the license is
> something you and the vendor agree to, whether via a webpage,
> handshake, signed contract, or whatever.  Binary libraries, for
> example, don't have any licensing information 'in' them, yet whatever
> terms you and vendor come to is what's in effect, essentially forever,
> assuming that it's legal in the first place.  Nobody can change this
> after the fact.

Binary libraries are irrelevant here.

Look at the FSF's page on "How to use GNU licenses for your own
software"[1]:

   "Whichever license you plan to use, the process involves adding two
   elements to each source file of your program: a copyright notice
   (such as "Copyright 1999 Terry Jones"), and a statement of copying
   permission, saying that the program is distributed under the terms of
   the GNU General Public License (or the Lesser GPL)."

And you _should_ include a copy of the licenses, but if you don't the
notice should say where to get them from.

So if AdaCore are releasing their code under the GPL they ought to
conform to the general policies of the FSF. And if the code they make
available contains the GCC Runtime Library Exception then it just
does. There are several remedies: one of those would be to take down the
publicly-visible anonymous repos.

They already recognise this in the GPL compiler suite by stripping the
exception out of the source files.

[1] http://www.gnu.org/licenses/gpl-howto.html

Re: Cairo bindings and e-mail license virus bombs

[Damien Carbonne]

Hello,

Thanks Dmitry for citing this binding.
I haven't written any Ada code for a long time (but I still follow CLA), 
and indeed, this Cairo binding is quite out dated. Most of it is 
independent of Gtk and can be used without Gtk (for png, svg, pdf, ...).
This binding would need an update from an API point of view.
I had done some work in that direction, but never uploaded it.
The Gtk part should be adapted to Gtk3, but that should not be too 
difficult.
If someone was interested by such an update, I could give it a try.
In my opinion, this binding had no chance against AdaCore. Would things 
have changed ?

Regards,
Damien Carbonne

Le 10/12/2014 21:23, Dmitry A. Kazakov a écrit :
> On Wed, 10 Dec 2014 10:32:51 -0800 (PST), David Botton wrote:
>
>> I do not plan on using the Cairo bindings from AdaCore nor any of the
>> AdaCore available bindings since finding this in the GNU/Debian
>> distribution of gcc/ada.
>>
>> /usr/share/doc/libgtkada2.24.4
>
> David, if you are OK with using outdated Cairo (GTK 2.x) then here are
> bindings developed by Damien Carbonne:
>
> http://sourceforge.net/projects/cairoada/
>
> They were better designed than AdaCore's, higher-level, packing ugly
> objects in controlled handle types.
>

Re: Cairo bindings and e-mail license virus bombs

[Ludovic Brenta]

David Botton writes on comp.lang.ada:
> The content of the e-mail link shout it change in the future is:
>
>> Many of the GtkAda source files contain the text of the GMGPL
>> exception; some do not. Are you saying I should not rely on the
>> presence or absence of the exception in the individual source texts
>> obtained from this repository:
>
> That's correct, headers on source file have no/little legal
> significance.  The main and simple message is that all the software
> available via the libre site, including svn.eu.adacore.com/anonsvn is
> provided under the pure GPL license.

AdaCore have changed this policy again very recently[1].  Now their
official statement is that files downloaded from their site are under
the license stated in the files.

Their previous policy ("everything is GPL") was in fact illegal; they
had embedded some non-GPL software into theirs (e.g. several years ago,
a copy of Berkeley DB was embedded in GPS) but they had, and still have,
no right to change the license of this third-party software.

So, I think their new policy is a big improvement.

[1] https://lists.debian.org/debian-ada/2014/10/msg00023.html

--
Ludovic Brenta.

Re: Cairo bindings and e-mail license virus bombs

[David Botton]

That was before the freeze for Jessie I think? Is there plans then on updating the Debian packages before full Jessie release? I think it is important that there not be "questions" of licenses floating around GNU and Free software.

> Their previous policy ("everything is GPL") was in fact illegal;

I don't know if Romain's e-mail really clears things up though completely, but ok.

David Botton

Re: Cairo bindings and e-mail license virus bombs

[Ludovic Brenta]

David Botton writes on comp.lang.ada:
> That was before the freeze for Jessie I think? Is there plans then on
> updating the Debian packages before full Jessie release? I think it is
> important that there not be "questions" of licenses floating around
> GNU and Free software.

Some packages were indeed updated just before the freeze for this
reason; libgtkada is one.

--
Ludovic Brenta.

ACT and the GPL (once again), was: Re: Cairo bindings and e-mail license virus bombs

[Simon Clubley]

On 2014-12-10, Ludovic Brenta <ludovic@ludovic-brenta.org> wrote:
> David Botton writes on comp.lang.ada:
>> The content of the e-mail link shout it change in the future is:
>>
>>> Many of the GtkAda source files contain the text of the GMGPL
>>> exception; some do not. Are you saying I should not rely on the
>>> presence or absence of the exception in the individual source texts
>>> obtained from this repository:
>>
>> That's correct, headers on source file have no/little legal
>> significance.  The main and simple message is that all the software
>> available via the libre site, including svn.eu.adacore.com/anonsvn is
>> provided under the pure GPL license.
>
> AdaCore have changed this policy again very recently[1].  Now their
> official statement is that files downloaded from their site are under
> the license stated in the files.
>
> Their previous policy ("everything is GPL") was in fact illegal; they
> had embedded some non-GPL software into theirs (e.g. several years ago,
> a copy of Berkeley DB was embedded in GPS) but they had, and still have,
> no right to change the license of this third-party software.
>
> So, I think their new policy is a big improvement.
>
> [1] https://lists.debian.org/debian-ada/2014/10/msg00023.html
>

Are we sure it's an official ACT statement and not just someone's opinion ?

That statement also doesn't match up with David's quote from ACT's email
above, which is presumably more recent than the 29-Oct-2014 Debian email
linked to above.

It's also clear from that quote above that GtkAda is considered to be
pure GPL as far as the distribution in Debian is concerned because that's
where the Debian notes say Debian obtains the original package from.

(This came up a couple of months ago when Jeffrey Carter thought the
Debian distribution was GMGPL and I looked into the situation to find out
what was going on.)

I've never forgotten that prior to GTK 2.4 ACT employees were describing
GtkAda as using a GMGPL type license, but that suddenly changed with
GTK 2.4 to pure GPL so I read _any_ ACT statements in this area very
carefully.

Sometimes ACT can be their own worst enemy when it comes to promoting
Ada. :-( How often do the C++ people have to worry about this stuff ?

(The C++ gtkmm binding is listed as LGPL; I just checked before posting.)

Simon.

-- 
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP
Microsoft: Bringing you 1980s technology to a 21st century world

Re: ACT and the GPL (once again), was: Re: Cairo bindings and e-mail license virus bombs

[Ludovic Brenta]

Simon Clubley writes on comp.lang.ada:
>> [1] https://lists.debian.org/debian-ada/2014/10/msg00023.html
>
> Are we sure it's an official ACT statement and not just someone's
> opinion ?

It is a statement by an AdaCore representative in response to an
explicit request by a Debian Developer, on a public mailing list.  I
don't think it can get any more official than that :)

> That statement also doesn't match up with David's quote from ACT's
> email above, which is presumably more recent than the 29-Oct-2014
> Debian email linked to above.

No, David is guilty here; he omitted context, URL, attribution and
timestamp in that post, leading to your confusion.  The URL was in his
previous post:

http://lists.adacore.com/pipermail/gtkada/2009-June/003789.html

> It's also clear from that quote above that GtkAda is considered to be
> pure GPL as far as the distribution in Debian is concerned because
> that's where the Debian notes say Debian obtains the original package
> from.

"is considered" is not a legal term and we Debian Developers prefer to
obtain sources from AdaCore's public Subversion repository, not the
yearly tarballs, whenever possible.

> Sometimes ACT can be their own worst enemy when it comes to promoting
> Ada. :-( How often do the C++ people have to worry about this stuff ?

Right.  Hopefully the latest statement by AdaCore sets things straight.

--
Ludovic Brenta.

Re: ACT and the GPL (once again), was: Re: Cairo bindings and e-mail license virus bombs

[David Botton]

> No, David is guilty here; he omitted context, URL, attribution and

Well guilty is a strong word :), as my second post said this is the text in case the archive goes away, the URL was in the original.

> timestamp in that post, leading to your confusion.  The URL was in his
> previous post:
> 
> http://lists.adacore.com/pipermail/gtkada/2009-June/003789.html

Ludovic, when will the change be reflected in the Debian Jessie repo? I pulled that text from Debian by my Jessie install still has the link of shame in it and says GtkAda on Debian is GPL.

David Botton

Re: ACT and the GPL (once again), was: Re: Cairo bindings and e-mail license virus bombs

[Ludovic Brenta]

David Botton writes on comp.lang.ada:
>> No, David is guilty here; he omitted context, URL, attribution and
>
> Well guilty is a strong word :), as my second post said this is the
> text in case the archive goes away, the URL was in the original.
>
>> timestamp in that post, leading to your confusion.  The URL was in his
>> previous post:
>> 
>> http://lists.adacore.com/pipermail/gtkada/2009-June/003789.html
>
> Ludovic, when will the change be reflected in the Debian Jessie repo?

Never.  Jessie is frozen and a change of license does not correct a
release-critical bug, so would not be accepted anyway.  Furthermore, it
I have not yet checked whether the sources in AdaCore's Subversion
repository contain the run-time library exception or not.

You can expect updates to libgtkada, reflecting whatever new license is
in force, in Debian unstable (Sid) after the freeze, probably early
2015.  These updates will be targeted at the next stable versions of
Debian: Debian 9 "Stretch" and Debian 10 "Buster".

--
Ludovic Brenta.

Re: ACT and the GPL (once again), was: Re: Cairo bindings and e-mail license virus bombs

[David Botton]

> Never.  Jessie is frozen and a change of license does not correct a
> release-critical bug, so would not be accepted anyway. 

Understood, score yet another point for anti-Ada advocacy (not your fault, the ones who created the mess are to blame) and Jessie will be around for a few years as is.

David Botton