From mboxrd@z Thu Jan 1 00:00:00 1970 X-Spam-Checker-Version: SpamAssassin 3.4.6 (2021-04-09) on ip-172-31-74-118.ec2.internal X-Spam-Level: X-Spam-Status: No, score=-1.9 required=3.0 tests=BAYES_00,FREEMAIL_FROM autolearn=ham autolearn_force=no version=3.4.6 X-Received: by 2002:a05:622a:1d3:: with SMTP id t19mr1879261qtw.128.1632941882700; Wed, 29 Sep 2021 11:58:02 -0700 (PDT) X-Received: by 2002:a5b:54e:: with SMTP id r14mr1632356ybp.457.1632941882516; Wed, 29 Sep 2021 11:58:02 -0700 (PDT) Path: eternal-september.org!reader02.eternal-september.org!news.misty.com!border2.nntp.dca1.giganews.com!nntp.giganews.com!news-out.google.com!nntp.google.com!postnews.google.com!google-groups.googlegroups.com!not-for-mail Newsgroups: comp.lang.ada Date: Wed, 29 Sep 2021 11:58:02 -0700 (PDT) In-Reply-To: Injection-Info: google-groups.googlegroups.com; posting-host=47.185.200.73; posting-account=zwxLlwoAAAChLBU7oraRzNDnqQYkYbpo NNTP-Posting-Host: 47.185.200.73 References: User-Agent: G2/1.0 MIME-Version: 1.0 Message-ID: Subject: Re: How to challenge a GCC patch? From: Andreas ZEURCHER Injection-Date: Wed, 29 Sep 2021 18:58:02 +0000 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable Xref: reader02.eternal-september.org comp.lang.ada:62869 List-Id: On Monday, September 27, 2021 at 9:20:04 AM UTC-5, J-P. Rosen wrote: > Le 27/09/2021 =C3=A0 14:48, Simon Wright a =C3=A9crit :=20 > > "J-P. Rosen" writes:=20 > >=20 > >> AdaCore has introduced a patch in FSF GCC to remove ASIS support.=20 > >>=20 > >> AdaCore is free to do what they want with their own version of=20 > >> GCC. However, removing a useful feature from the FSF version with the= =20 > >> goal to promote their own, in-house tool is clearly against the spirit= =20 > >> of free software.=20 > >>=20 > >> Does anybody know the procedures set by the FSF to challenge a patch?= =20 > >=20 > > It's not just the patch(es), it's any subsequent changes to affected=20 > > parts of the compiler.=20 > > > Right, if they want to contribute further patches, they'll have to keep= =20 > it ASIS compatible. That's not a reason to divert gcc to support their=20 > own private interest. > --=20 > J-P. Rosen=20 > Adalog=20 > 2 rue du Docteur Lombard, 92441 Issy-les-Moulineaux CEDEX=20 > Tel: +33 1 45 29 21 52=20 > https://www.adalog.fr There does seemingly exist a way to directly compete against AdaCore's incr= easing proprietarization of GNAT to herd Ada users into GNAT Pro licenses a= s a higher-revenue way of what was supposed (back in the 1990s) to have bee= n a business model for charging only for =E2=80=A2support=E2=80=A2 of free = open-source software (spelled out instead of FOSS to emphasize free and ope= nness of open source). Because neither FSF's GNAT nor FSF's GNAT Runtime L= ibrary are licensed under the GNU Affero Public License (GAPL), it seems th= at it would be possible to both 1) execute GPL-licensed proprietary-extende= d FSF GNAT in the cloud and then 2) link against the Runtime-Library-Except= ion(RLE)-licensed FSF-GNAT's runtime library in the cloud without divulging= the source code to the proprietary extensions to GNAT, as long as the copy= of the GNAT compiler that resides on VM instances in the cloud cannot be d= ownloaded or distributed. Then the R&D effort needed to extend GNAT in the= se proprietary ways would presumably (under regular GPL) not need to contri= buted back to AdaCore, because no distribution occurred, hence complete com= pliance (I suspect) with the terms of GPLv3. As I understand it, this woul= d change radically if AdaCore were to somehow switch FSF GNAT's licensing o= ver to the Affero license. But until AdaCore (with GNU Foundation's blessing) would switch FSF GNAT to= Affero license, for a third-party company who is aggrieved at the loss of = ASIS and its supporting constructs within GNAT, that aggrieved company coul= d conceivably raise funds to perform the expensive R&D to put ASIS back int= o a cloud-only proprietary-extension version of the GNAT compiler without g= iving their hard work away to AdaCore. Hence, that aggrieved company could= conceivably attract investors to the aggrieved company's business model by= having a barrier-to-entry-fortified value proposition in the cloud-only pr= oprietary-extension version of the GNAT compiler. Yes, of course, the downside would be that every compilation via the cloud-= only proprietary-extension version of the GNAT compiler would need an extan= t working Internet connection=E2=80=94no offline compilation capability in = this alternate version of GNAT. And yes, of course, this would raise the i= re of both AdaCore and GNU Foundation, who could then switch over to the Af= fero license for FSF GNAT, or even author some GPLv4 to address this loopho= le. And yes, of course, porting* any ASIS software to RLEv3.1-licensed lib= adalang would be fraught with far fewer legal complications, and in fact mi= ght be less software-development effort as well. * Porting here of course means practically rewriting from scratch, which is= not the usual definition of =E2=80=9Cporting=E2=80=9D. The posting above is merely a record of my personal musings. I am not a la= wyer. I haven't passed any bar exam nor been admitted to any bar in any ju= risdiction nor have any license to practice law. The letter & interpretati= on of copyright law might very well vary in your government's jurisdiction.= But even with the aid of an attorney, only you can form your own understa= nding of GPLv3 and GAPLv3; your understanding might reach different conclus= ions than anything indicated in this amateur posting.