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!.POSTED!not-for-mail From: Simon Clubley Newsgroups: comp.lang.ada Subject: Re: Could you write a BSD like os in ADA? Date: Sun, 28 Aug 2016 08:36:09 -0000 (UTC) Organization: A noiseless patient Spider Message-ID: References: <79e591f0-3c3e-42b2-ad1f-3e59a031531e@googlegroups.com> <94756e03-7788-4032-a70b-3a0468fc3af9@googlegroups.com> <2b473012-b01e-4c46-add5-048889bf20df@googlegroups.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Injection-Date: Sun, 28 Aug 2016 08:36:09 -0000 (UTC) Injection-Info: mx02.eternal-september.org; posting-host="db214455796e2ab20a867b81483f4b1b"; logging-data="2299"; mail-complaints-to="abuse@eternal-september.org"; posting-account="U2FsdGVkX1+FxOjoTKh+8u3pbJYeatXkOF8LjeQ/vtQ=" User-Agent: slrn/0.9.9p1 (Linux) Cancel-Lock: sha1:Dfqmx640a8HsfDCezUhUMPfRHLs= Xref: news.eternal-september.org comp.lang.ada:31627 Date: 2016-08-28T08:36:09+00:00 List-Id: On 2016-08-28, Bob Broderbund wrote: > On 2016-08-28, Simon Clubley wrote: >> On 2016-08-24, Jeffrey R. Carter wrote: >>> >>> DEC contracted with AdaCore (then ACT) to provide the DEC Ada-95 compiler. This >>> is why GNAT recognizes all the DEC Ada(-83) pragmas, such as Passive. >>> >> >> That doesn't mean all the required bits have shown up in the FSF version >> of the GCC codebase (don't forget AdaCore maintain their own codebase). > > I think Mr. Carter probably knows the history pretty well ;-) > > But I don't think Adacore has much interest in non-Windows or non-Linux > platforms in the native target space. They got rid of their Solaris builds > over the years. Still, I remember a rumor that there is something going on > for the VMS Intel port at Adacore and it makes sense. It should not be that > difficult to port the Linux version to VMS _if_ a current gcc toolchain is > available. If VMS goes all the way with Clang/LLVM and doesn't get an up to > date gcc going it will be a lot harder. > I have the feeling (but no firm information) that VSI are looking at the same model that DEC used when DEC contracted AdaCore to supply an Ada 95 compiler for VMS Alpha. IOW, if true, AdaCore would do the port of gcc to VMS x86-64 while VSI continued with their own compiler toolchain for the other languages. This would match what happened with VMS Alpha with DEC using their own compiler technology for the other DEC supplied VMS Alpha compilers. >> I was able to use the cross compiler to build some C programs which ran >> on VMS, but Fortran code fell over at runtime either during the image >> activation process or very shortly afterwards (I can't remember which). > > If you post the results maybe I could help, or some of the guys in > c.l.f. would help. I'm a little surprised anyway. I though the front-end was > completely Ada and the back end was C/C++. I did not know there was any > Fortran involved in gnat. > The front end is purely Ada, and there is no Fortran code in GNAT. When I tried to build the FSF GCC cross compiler on Linux I added multiple front ends (Ada, Fortran, C, C++) to the --enable-languages option before reducing the list when the various front ends started falling over. Thanks for the offer BTW, but Fortran was a very low priority goal which I didn't spend any real time on when I tried this and things went wrong. My last attempt to build a VMS cross compiler is now getting on for a couple of years ago so I have moved into other projects now. >> Both the C++ and Ada frontends failed to build when building the >> cross compiler (GNAT fell over with an ICE during the build and C++ was >> some header problem IIRC). > > I suspect a current gcc on VMS is going to be essential. Much code that uses > gcc exploits every last gccism and does so increasingly all the time. > This was using the latest versions of gcc at the time. The problem here _appeared_ to be that not all the required bits to build gcc for a VMS target had been pushed from the AdaCore code base to the FSF code base. Simon. -- Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP Microsoft: Bringing you 1980s technology to a 21st century world