[Bug c++/43036] [4.3/4.4/4.5 Regression] c++ compilation hang
--- Comment #20 from Roger dot Jeurninck at home dot nl 2010-02-16 10:36 --- great job, couldn't even hope that I got a solution that fast!! The patch did the trick for 4.4.2 on Solaris and the 'hanging' file builds without any issues. pls let me know if I can be of any help in further verification of this issue thanks a lot again! Roger -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43036
[Bug c++/43036] c++ compilation hang
--- Comment #7 from Roger dot Jeurninck at home dot nl 2010-02-12 14:14 --- Created an attachment (id=19850) -- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=19850action=view) test file -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43036
[Bug c++/43036] c++ compilation hang
--- Comment #8 from Roger dot Jeurninck at home dot nl 2010-02-12 14:15 --- I have attached a file on which I could reproduce the hang on: - sparc-sun-solaris2.8 / GCC 4.4.2 - sparc-sun-solaris2.10 / GCC 4.4.2 - sparc-sun-solaris2.10 / GCC 4.4.3 - i486-linux-gnu / GCC 4.4.2 I hope you can help me with the analysis of this issue. thanks, Roger -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43036
[Bug c++/43036] New: c++ compilation hang
Hi, compiling a specific C++ code keeps de GCC 4.4.2 compiler hanging forever (or at least 10 hours). Note that the c++ compiler works perfectly for the other c++ sources (1000) Unfortunately I cannot provide the C++ as-is because of company policy. But I did some anlysis myself, hope you can help me further with this info. These are the steps I took: - I generated a file with the preprocessed output and this file was identical with the one generated by my previous compiler (GCC 4.1.2/Solaris 8) - Compiling the preprocessed file with the 4.4.2 compiler did (of course) again fail with a hanging cc1plus process. Compiling the same preprocessed file with the 4.1.2 compiler did not cause any problems - I did try it to compile the preprocessed file with different optimizations with/without debug info but none of them were succesfull with the 4.4.2 compiler - Using different compilations of the compiler itself did not help either - I could reproduce it on bot solaris 8 10 - I could find out on which location in the preprocessed file the Compiler hangs. However if I for example reorder this file the compiler hangs on a different location, so this doesn't help the analysis either :-( - Next I forced a coredump of the hanging proces and I opened with GDB: Core was generated by `/sft/teamsdoc/sds_group/rjeurnin/gcc_cadenv/SunOS_5.8/gcc/gcc-4.4.2-B3/libexec/'. [New process 76177] #0 strip_array_types (type=0xfee91520) at /cws/rjeurnin/gcc/sources/gcc-4.4.2/gcc/tree.c:5814 5814 while (TREE_CODE (type) == ARRAY_TYPE) - A backtrace gave: #0 strip_array_types (type=0xfee91520) at /cws/rjeurnin/gcc/sources/gcc-4.4.2/gcc/tree.c:5814 #1 0x000dffb4 in cp_type_quals (type=0xfe1b3f48) at /cws/rjeurnin/gcc/sources/gcc-4.4.2/gcc/cp/typeck.c:7212 #2 0x0010e220 in cp_build_qualified_type_real (type=0xf865f588, type_quals=1, complain=value optimized out) at /cws/rjeurnin/gcc/sources/gcc-4.4.2/gcc/cp/tree.c:783 #3 0x0008111c in grokdeclarator (declarator=0x7a4398, declspecs=0x7a43bc, decl_context=PARM, initialized=0, attrlist=0x7a43fc) at /cws/rjeurnin/gcc/sources/gcc-4.4.2/gcc/cp/decl.c:8102 #4 0x000d1924 in cp_parser_parameter_declaration_list (parser=0xfef8d0a0, is_error=0xffbee42f ) at /cws/rjeurnin/gcc/sources/gcc-4.4.2/gcc/cp/parser.c:14063 #5 0x000d222c in cp_parser_declarator (parser=0xfef8d0a0, dcl_kind=CP_PARSER_DECLARATOR_NAMED, ctor_dtor_or_conv_p=0xffbee498, parenthesized_p=value optimized out, member_p=0 '\0') at /cws/rjeurnin/gcc/sources/gcc-4.4.2/gcc/cp/parser.c:13982 #6 0x000d7c2c in cp_parser_init_declarator (parser=0xfef8d0a0, decl_specifiers=0xffbee508, checks=0x0, function_definition_allowed_p=1 '\001', member_p=0 '\0', declares_class_or_enum=0, function_definition_p=0xffbee55f ) at /cws/rjeurnin/gcc/sources/gcc-4.4.2/gcc/cp/parser.c:12661 #7 0x000d8694 in cp_parser_simple_declaration (parser=0xfef8d0a0, function_definition_allowed_p=1 '\001') at /cws/rjeurnin/gcc/sources/gcc-4.4.2/gcc/cp/parser.c:8240 #8 0x000dca18 in cp_parser_declaration (parser=0xfef8d0a0) at /cws/rjeurnin/gcc/sources/gcc-4.4.2/gcc/cp/parser.c:8043 #9 0x000dc208 in cp_parser_declaration_seq_opt (parser=0xfef8d0a0) at /cws/rjeurnin/gcc/sources/gcc-4.4.2/gcc/cp/parser.c:7934 #10 0x000dc520 in c_parse_file () at /cws/rjeurnin/gcc/sources/gcc-4.4.2/gcc/cp/parser.c:3036 #11 0x001544a4 in c_common_parse_file (set_yydebug=0) at /cws/rjeurnin/gcc/sources/gcc-4.4.2/gcc/c-opts.c:1252 #12 0x0038fbc8 in toplev_main (argc=value optimized out, argv=value optimized out) at /cws/rjeurnin/gcc/sources/gcc-4.4.2/gcc/toplev.c:970 #13 0x00065a58 in _start () ... Triggered by a bug (id: 42803), solved with the 4.4.3 release, I did some more checks: - I could reproduce this bug on GCC 4.4.2/solaris 8 with the test file, uploaded in the bugreport: bilateral2.ii - With the new compiler (4.4.3) this bug was not reproducable anymore with the file bilateral2.ii and that problem is solved on Solaris 8 too - however my problem still exists... :-( Can someone help me on this one? -- Summary: c++ compilation hang Product: gcc Version: 4.4.2 Status: UNCONFIRMED Severity: normal Priority: P3 Component: c++ AssignedTo: unassigned at gcc dot gnu dot org ReportedBy: Roger dot Jeurninck at home dot nl GCC build triplet: sparc-sun-solaris2.8 GCC host triplet: sparc-sun-solaris2.8 GCC target triplet: sparc-sun-solaris2.8 http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43036
[Bug c++/43036] c++ compilation hang
--- Comment #1 from Roger dot Jeurninck at home dot nl 2010-02-11 15:30 --- Created an attachment (id=19844) -- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=19844action=view) config.log file the generated config.log file, from GCC's build dir -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43036
[Bug c++/43036] c++ compilation hang
--- Comment #2 from Roger dot Jeurninck at home dot nl 2010-02-11 15:34 --- gcc -v info: Using built-in specs. Target: sparc-sun-solaris2.8 Configured with: /cws/rjeurnin/gcc/sources/gcc-4.4.2/configure --prefix=/sft/teamsdoc/sds_group/rjeurnin/gcc_cadenv/SunOS_5.8/gcc/gcc-4.4.2-B3 --enable-languages=c,c++ --with-gmp-lib=/sft/teamsdoc/sds_group/rjeurnin/gcc_cadenv/SunOS_5.8/gmp/gmp-4.3.2/lib --with-gmp-include=/sft/teamsdoc/sds_group/rjeurnin/gcc_cadenv/SunOS_5.8/gmp/gmp-4.3.2/include --with-mpfr-lib=/sft/teamsdoc/sds_group/rjeurnin/gcc_cadenv/SunOS_5.8/mpfr/mpfr-2.4.2/lib --with-mpfr-include=/sft/teamsdoc/sds_group/rjeurnin/gcc_cadenv/SunOS_5.8/mpfr/mpfr-2.4.2/include --without-gnu-as --with-ld=/usr/ccs/bin/as --without-gnu-ld --with-ld=/usr/ccs/bin/ld --disable-nls Thread model: posix gcc version 4.4.2 (GCC) COLLECT_GCC_OPTIONS='-v' '-mcpu=v9' -- Roger dot Jeurninck at home dot nl changed: What|Removed |Added Summary| c++ compilation hang |c++ compilation hang http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43036
[Bug c++/43036] c++ compilation hang
--- Comment #4 from Roger dot Jeurninck at home dot nl 2010-02-11 15:57 --- Hi Paolo, thanks for the quick reply! Shortening the preprocessed file doesn't help as the problem won't occur anymore. I will try to obfuscate such that I'm allowed to upload it. Tomorrow I'll test the issue with the 4.4.2 linux compiler too. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43036
[Bug fortran/42723] Bootstrapping Fortran with GCC 4.4.2 fails on Solaris 5.10: f951: internal compiler error: Arithmetic Exception
--- Comment #5 from Roger dot Jeurninck at home dot nl 2010-01-15 12:47 --- I got GCC bootstrapped now! Instead of building the GMP and MPFR sources together with GCC I did a build of them in advice. Next I configured the GCC with the --with-gmp and --with-mpfr options (see below) and could successfully build GCC. Not sure why the first approach did fail because the MPFR and GMP code was indeed nicely build with the GCC compiler... Testresults are ok and almost match those of Rainer (http://gcc.gnu.org/ml/gcc-testresults/2009-11/msg01181.html). I will still look into the differences here. For now this issue can be closed/postponed as I have a workaround. thanks for your help! Roger (the configure options I used, note that the LD_LIBRARY_PATH needs to find the libs too) --- --prefix=../install \ --enable-languages=c,fortran,c++ \ --with-gmp-lib=/scratch/rjeurnin/gmp/gmp-4.3.2/install/lib \ --with-gmp-include=/scratch/rjeurnin/gmp/gmp-4.3.2/install/include \ --with-mpfr-lib=/scratch/rjeurnin/mpfr/mpfr-2.4.2/install/lib \ --with-mpfr-include=/scratch/rjeurnin/mpfr/mpfr-2.4.2/install/include \ --verbose -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=42723
[Bug fortran/42723] New: Bootstrapping Fortran with GCC 4.4.2 fails on Solaris 5.10: f951: internal compiler error: Arithmetic Exception
Building the GCC compiler for the language Fortran fails in bootstrapping mode. Strangely, I am able to build the Fortran compiler in native non-bootstrapping mode. If I do start the build in bootstrapping mode, it finally fails with the following error (I will attach the complete log): *** configure:14187: /scratch/rjeurnin/gcc/gcc-4.4.2-B10/build/./gcc/gfortran -B/scratch/rjeurnin/gcc/gcc-4.4.2-B10/build/./gcc/ -B/scratch/rjeurnin/gcc/gcc-4.4.2-B10/install/sparc-sun-solaris2.10/bin/ -B/scratch/rjeurnin/gcc/gcc-4.4.2-B10/install/sparc-sun-solaris2.10/lib/ -isystem /scratch/rjeurnin/gcc/gcc-4.4.2-B10/install/sparc-sun-solaris2.10/include -isystem /scratch/rjeurnin/gcc/gcc-4.4.2-B10/install/sparc-sun-solaris2.10/sys-include -c -g -O2 conftest.f 5 f951: internal compiler error: Arithmetic Exception Please submit a full bug report, with preprocessed source if appropriate. See http://gcc.gnu.org/bugs.html for instructions. configure:14193: $? = 1 configure: failed program was: | | program foo | real, parameter :: bar = sin (12.34 / 2.5) | end program foo *** Running the testsuite on the bootstrapped build will (obviously) give very poor results for the fortran part. This confirms that the fortran compiler is messed up seriously. Hope you can help me as I'm completely stuck on this one. thanks, Roger -- Summary: Bootstrapping Fortran with GCC 4.4.2 fails on Solaris 5.10: f951: internal compiler error: Arithmetic Exception Product: gcc Version: 4.4.2 Status: UNCONFIRMED Severity: normal Priority: P3 Component: fortran AssignedTo: unassigned at gcc dot gnu dot org ReportedBy: Roger dot Jeurninck at home dot nl GCC build triplet: sparc-sun-solaris2.10 GCC host triplet: sparc-sun-solaris2.10 GCC target triplet: sparc-sun-solaris2.10 http://gcc.gnu.org/bugzilla/show_bug.cgi?id=42723
[Bug fortran/42723] Bootstrapping Fortran with GCC 4.4.2 fails on Solaris 5.10: f951: internal compiler error: Arithmetic Exception
--- Comment #1 from Roger dot Jeurninck at home dot nl 2010-01-13 12:25 --- Created an attachment (id=19569) -- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=19569action=view) libgfortran/config.log file containing the log (libgfortran) of the failing bootstrap build -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=42723
[Bug fortran/42723] Bootstrapping Fortran with GCC 4.4.2 fails on Solaris 5.10: f951: internal compiler error: Arithmetic Exception
--- Comment #4 from Roger dot Jeurninck at home dot nl 2010-01-14 07:15 --- Thanks for the quick respond! I'm not building the compiler in the source tree. .. I will take a deeper look on the GMP and MPFR libs and keep you posted on the progress. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=42723