OS : Mac OS x 10.6.2 (Snow Leopard) IDE tools : Xcode 3.2.2 Processor : Intel Core 2 Duo Memory : 1 GB Problem : the make step operates well until teh comparison phase between stages 2 and 3. Bootstrap matter ... I have saved the files you need but how to send them to you ? Yours sincerely, Laurent Delphin, Dunkerque (EU)
first can you give the output from the failure: i.e. which files have differences? ...the configuration line you are using. the output of autoconf --version automake --version m4 --version I should remind you of : http://gcc.gnu.org/install/prerequisites.html x86_64-apple-darwin10 will not build gcc properly with the auto* tools installed - you must ensure that the required versions are found.
MacPorts has accumulated a number of users who seem to run into this issue with gcc 4.5.0 (apparently always with libgomp)... https://svn.macports.org/ticket/24664 Peter O'Gorman also has a problem machine as a well (a Mac Mini) which randomly fails the bootstrap comparison. Oddly I haven't seen this on my MacPro.
If the comparison failure is for libgomp, this pr is a duplicate of pr43170. The origin of the problem can be seen with the following test: [macbook] f90/bug% grep -i tls /opt/gcc/omp_build_w_fail7/stage2-x86_64-apple-darwin10.3.0/libgomp/config.log | #define HAVE_TLS 1 | #define HAVE_TLS 1 | #define HAVE_TLS 1 | #define HAVE_TLS 1 gcc_cv_have_tls=yes #define HAVE_TLS 1 [macbook] f90/bug% grep -i tls /opt/gcc/omp_build_w_fail7/stage2-x86_64-apple-darwin10.3.0/i386/libgomp/config.log gcc_cv_have_tls=no [macbook] f90/bug% grep -i tls /opt/gcc/omp_build_w_fail7/stage3-x86_64-apple-darwin10.3.0/libgomp/config.log | #define HAVE_TLS 1 | #define HAVE_TLS 1 | #define HAVE_TLS 1 | #define HAVE_TLS 1 gcc_cv_have_tls=yes #define HAVE_TLS 1 [macbook] f90/bug% grep -i tls /opt/gcc/omp_build_w_fail7/stage3-x86_64-apple-darwin10.3.0/i386/libgomp/config.log | #define HAVE_TLS 1 | #define HAVE_TLS 1 | #define HAVE_TLS 1 | #define HAVE_TLS 1 gcc_cv_have_tls=yes #define HAVE_TLS 1 AFAICT "gcc_cv_have_tls=no" can occur at stage 2 or 3 and in the main lib or in the i386 one.
4.5-branch (as of r160013) has an error in config.gcc (which I just fixed on trunk yesterday, with r159979) in which several t-make fragments are included twice on x86_64-*-darwin*. If anyone is interested to try it out I could make you a patch for 4.5-branch to do the same fix. still, first we need a repeatable phenomenon - and to be sure that it's not tool-related.
Created attachment 20809 [details] Shell (bash) script to build gcc Files related to this bug : shell (bash) script, log files.
Created attachment 20810 [details] config.log
Created attachment 20811 [details] configure.log File generated by ./configure ... 2>&1 | tee configure.log
Created attachment 20812 [details] All the files
(In reply to comment #1) OK Thank you very much. I have verified the prerequisites before. Please note I just send a zip files with all log files. Thanks. ======================================================== > first can you give the output from the failure: > i.e. which files have differences? > ...the configuration line you are using. > the output of > autoconf --version > automake --version > m4 --version > I should remind you of : > http://gcc.gnu.org/install/prerequisites.html > x86_64-apple-darwin10 will not build gcc properly with the auto* tools > installed - you must ensure that the required versions are found.
(In reply to comment #9) > (In reply to comment #1) > OK Thank you very much. > > I have verified the prerequisites before. > Please note I just send a zip files with all log files. thanks for posting the files - see also : http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43170 It does look like this is the same issue. However, to confirm: Is this phenomenon deterministic - or a random occurrence depending on processor load? (as I look at your script, it would seem that you have not done a parallel build) if you remove all the libgomp directories and then re-do make - does it succeed ?
*** This bug has been marked as a duplicate of 43170 ***