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!mx02.eternal-september.org!news.eternal-september.org!.POSTED!not-for-mail From: Simon Wright Newsgroups: comp.lang.ada Subject: Re: ANN: GCC 4.9.0 (2014) for Max OS X Mavericks Date: Sun, 12 Oct 2014 10:05:42 +0100 Organization: A noiseless patient Spider Message-ID: References: <02796daf-0ab5-4722-a4ef-42155568e3c5@googlegroups.com> Mime-Version: 1.0 Content-Type: text/plain Injection-Info: mx02.eternal-september.org; posting-host="865ecc8333c04b7dc8d5f87fc772110b"; logging-data="31536"; mail-complaints-to="abuse@eternal-september.org"; posting-account="U2FsdGVkX1+/NNk4jWOdlq0rhe3URxzOAhTB1jYb1cA=" User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/24.3 (darwin) Cancel-Lock: sha1:8aBNeMe5EKqWR0nO8BygZYYeqoQ= sha1:4nzl0LhItYj0t/hk/bun7iRMjwA= Xref: news.eternal-september.org comp.lang.ada:22376 Date: 2014-10-12T10:05:42+01:00 List-Id: marioeroy@gmail.com writes: > On Sunday, June 1, 2014 5:06:04 AM UTC-4, Simon Wright wrote: >> >> ... >> >> GNAT GPL 2014 (for Mac) supports C, C++, Ada, and Objective C. My >> release additionally supports Fortran and Objective C++. In all >> cases, the compiler itself is built without change. I had to rebuild >> the command 'gnat', because the original build with the compiler >> failed with SIGSEGV; something to do with exceptions, I need to >> report this but I'm a bit too vague on the details. >> See https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61027 > I could not get the work around using gprbuild to work on Mavericks > 10.9.5. The gnat -v (specifying an invalid argument) still > crashes. After many attempts have finally tracked this down. Do you mean that the 'gnat' I provided doesn't work, or that you've tried to rebuild gnatcmd as I did and *that* didn't work? I'm still using gcc-4.9.0-x86_64-apple-darwin13-2014-1.tar.bz2 without problems (except that other gnat* commands still show the problem when used in ways which result in program exit via exception, such as 'gnatchop -h', or gnatmake when the library contains .ali files from a very old compiler release). I just rebuilt gnatcmd (Xcode 6.0.1 on Mac OS X 10.9.5, Darwin 13.4.0) and ./gnatcmd -v works fine. > One needs to pass --with-host-libstdcxx=-lstdc++ when configuring > GCC. Gnat will no longer SIGSEGV after make install. Thanks for the info. That must have been _some_ detective work! I see from my compiler build log that my gnat* commands were built with (a mammoth list of inputs, including) -B../../../x86_64-apple-darwin13/libstdc++-v3/src/.libs -B../../../x86_64-apple-darwin13/libstdc++-v3/libsupc++/.libs -L../../../x86_64-apple-darwin13/libstdc++-v3/src/.libs -L../../../x86_64-apple-darwin13/libstdc++-v3/libsupc++/.libs -static-libstdc++ which presumably means to use the newly-build libstc++ in GCC rather than the system one? Interesting, because the rebuilt gnatcmd doesn't appear to use libstdc++ at all (I eventually managed to get a build that used libgnat.dylib, of course that itself might have bound libstdc++ statically - but then how would my fix work?)