From mboxrd@z Thu Jan 1 00:00:00 1970 X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on polar.synack.me X-Spam-Level: X-Spam-Status: No, score=-1.9 required=5.0 tests=BAYES_00 autolearn=unavailable autolearn_force=no version=3.4.4 Path: eternal-september.org!reader01.eternal-september.org!reader02.eternal-september.org!news.eternal-september.org!news.eternal-september.org!news.eternal-september.org!.POSTED!not-for-mail From: Simon Clubley Newsgroups: comp.lang.ada Subject: Re: some questions on installing Ada on Linux Date: Fri, 4 Jul 2014 17:08:42 +0000 (UTC) Organization: A noiseless patient Spider Message-ID: References: <2871836a-3b86-4c5e-b7d9-7da3e9acf775@googlegroups.com> Injection-Date: Fri, 4 Jul 2014 17:08:42 +0000 (UTC) Injection-Info: mx05.eternal-september.org; posting-host="e458ff8b81bc0c159989eb0e36c6e372"; logging-data="15953"; mail-complaints-to="abuse@eternal-september.org"; posting-account="U2FsdGVkX1+D4vHCsG94PZRJ1rkgez33u1WzX+ZwhPk=" User-Agent: slrn/0.9.8.1 (VMS/Multinet) Cancel-Lock: sha1:hbcVRIbrMDEgSr/Np9Skq013vDo= Xref: news.eternal-september.org comp.lang.ada:20710 Date: 2014-07-04T17:08:42+00:00 List-Id: On 2014-07-04, Dan'l Miller wrote: > On Friday, July 4, 2014 7:22:37 AM UTC-5, Simon Clubley wrote: > >> > It is really confusing having Ada/GNAT in 2 different places. > > It is not done for reducing confusion. The AdaCore repository is done for > AdaCore to have a way to fund itself (which is better than the alternative: no > revenue). The FSF repository is done for legal reasons as FSF is the ultimate > owner of the rights to copy the GNAT compiler within GCC, not AdaCore. > Are you sure about that ? If the FSF has had a copyright assignment for the GNAT compiler from AdaCore, then surely it's only for the FSF branch of the gcc code. Please see below for my reasoning. > >> I'll let someone who installs the pre-packaged toolchains answer your >> other questions (I'm RedHat based and just build a new gcc/binutils >> toolchain from the FSF source when the need arises) but the core problem >> is that there are two distinct branches of the GNAT sources which cannot >> be merged due to licence conflicts. > > Is this actually factually correct? As I understand it, yes. If I am wrong about this, I would appreciate being corrected as this can be a bit subtle in areas. A third party cannot take the AdaCore GPL sources, import them into FSF GCC and then distribute the combined sources under the terms of the GMGPL. However, AdaCore, as the owner of the code in question, can choose, if it wishes, to add code to the FSF GCC codebase under more permissive GMGPL terms and then assign the rights to that copy of the added code to the FSF (if AdaCore indeed does do that). However, as I understand it, in the absence of any contracts/agreements granting additional rights to the customer, all AdaCore has to legally do to satisfy the GPL is to release it's source code, under the pure GPL, to the customers who purchase it's products or download the binaries from AdaCore's website. There's no requirement, unless it's part of some additional non-GPL contract, to release any source code under the GMGPL unless AdaCore choose to do so. Therefore, as I understand it, AdaCore could stop contributing code to the FSF gcc branch if it chooses, provided it makes the source code for the binaries it actually supplies to customers available under the GPL. > I was of the understanding that they > are substantially merged approximately once per year. What obstructs a > wholesale merge is the fact that AdaCore's GNAT lags behind 2 or 3 versions > from the latest stable back-end. FSF merges the vast majority of AdaCore's > evolution of GNAT that does not conflict with changes in the back-end. It is > bit-rot, not legalese, that naturally causes 2 source bases for GNAT. The > relationship is symbiotic: FSF depends on AdaCore for most of the evolution of > the Ada front-end, whereas AdaCore depends on FSF for most of the evolution of > the back-end. AdaCore merges in a newer (but still lagging) back-end version > approximately once per year to AdaCore's state-of-the-art front-end. > Separately, FSF merges in a newer (but still lagging) GNAT front-end version > approximately once per year to FSF's state-of-the-art back-end. Unless there's some contract I am unaware of, there's nothing to stop AdaCore saying that future releases to the FSF gcc codebase are done under the terms of the pure GPL. They have already done this once with another product. GtkAda used to have a GMGPL style licence up to around GtkAda 2.4, but the public version of GtkAda was converted (overnight) to a pure GPL licence. Existing releases of GtkAda stayed under the GMGPL; new releases were under the pure GPL. And once again, if I am wrong about this, please feel free to correct me. :-) Simon. PS: I have a polite request. Would it be possible for you to break up your paragraphs please ? They come across as a monolithic block of text which can be hard to read. Thanks. -- Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP Microsoft: Bringing you 1980s technology to a 21st century world