I use GNU Fortran 95 (GCC 3.5.0 20040702 (experimental)) and compile with /home/eschnett/gcc/bin/gfortran -ftrapv -march=pentium4 -malign-double -g3 -Wall -Wshadow -c -o $current_wd/ih_3determinant.F90.o $current_wd/ih_3determinant.f90 and receive the error message /home/eschnett/Calpha/Cactus/configs/einstein-gcc-debug/build/IsolatedHorizon/ih_3determinant.f90:0: internal compiler error: Segmentation fault Please submit a full bug report, with preprocessed source if appropriate.
Can you provide the source?
Created attachment 6675 [details] Failing source code
Created attachment 6676 [details] necessary module source
hmm, we are trying to print out an error at this point. Here is the backtrace: #0 error_string (p=0x73 <Address 0x73 out of bounds>) at /home/gates/pinskia/src/gnu/gcc/src/ gcc/fortran/error.c:109 #1 0x0805b7b0 in error_print (type=0x8463fcd "Error:", format0=Variable "format0" is not available. ) at /home/gates/pinskia/src/gnu/gcc/src/gcc/fortran/error.c:425 #2 0x0805bbbe in gfc_error (format=0x8467948 "COMMON block '%s' at %C has already been USE- associated") at /home/gates/pinskia/src/gnu/gcc/src/gcc/fortran/error.c:596 #3 0x0806f1f6 in gfc_match_common () at /home/gates/pinskia/src/gnu/gcc/src/gcc/fortran/match. c:2203 #4 0x08075837 in match_word (str=Variable "str" is not available. ) at /home/gates/pinskia/src/gnu/gcc/src/gcc/fortran/parse.c:66 #5 0x08075f2e in decode_statement () at /home/gates/pinskia/src/gnu/gcc/src/gcc/fortran/parse.c: 190 #6 0x08076285 in next_statement () at /home/gates/pinskia/src/gnu/gcc/src/gcc/fortran/parse.c: 353 #7 0x08076bac in parse_spec (st=ST_DATA_DECL) at /home/gates/pinskia/src/gnu/gcc/src/gcc/ fortran/parse.c:1534 #8 0x08077a5c in parse_progunit (st=Variable "st" is not available. ) at /home/gates/pinskia/src/gnu/gcc/src/gcc/fortran/parse.c:2255 #9 0x08078066 in gfc_parse_file () at /home/gates/pinskia/src/gnu/gcc/src/gcc/fortran/parse.c: 2551 #10 0x0808d41a in gfc_be_parse_file (set_yydebug=0) at /home/gates/pinskia/src/gnu/gcc/src/gcc/ fortran/f95-lang.c:264 #11 0x08391098 in toplev_main (argc=115, argv=0xbffec304) at /home/gates/pinskia/src/gnu/gcc/ src/gcc/toplev.c:981 #12 0x420174d9 in __libc_start_main () from /lib/i686/libc.so.6 #13 0x0804b071 in _start ()
Here's a reduced example. It looks like the common COMMON block causes the trouble. =================================== MODULE bar INTEGER :: I COMMON /X/I END MODULE bar SUBROUTINE foo() USE bar INTEGER :: J COMMON /X/J END SUBROUTINE foo ===================================
In my code, the common block in the module is declared private, so the names should not clash.
Subject: Bug 16336 CVSROOT: /cvs/gcc Module name: gcc Changes by: tobi@gcc.gnu.org 2004-07-10 16:26:06 Modified files: gcc/fortran : ChangeLog match.c Log message: PR fortran/16336 * match.c (match_common): Fix error reporting for used common. Patches: http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/fortran/ChangeLog.diff?cvsroot=gcc&r1=1.109&r2=1.110 http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/fortran/match.c.diff?cvsroot=gcc&r1=1.15&r2=1.16
This is invalid code: PRIVATE applies only to entities, a common block is no entity. The segfault is the result of a bug in the call to the error handler. Patch here: http://gcc.gnu.org/ml/fortran/2004-07/msg00085.html
Section 11.2.1, "The USE statement and use association" of the draft of the Fortran 2003 standard (J3/04-007) says "The USE statement provides the means by which a scoping unit accesses [...]" and then gives a list of what is accessed. This list does not include common block names (except if common blocks are "named data objects", which I do not think). That means that a module should never export common block names. The standard speaks later in the same section of "accessible entities". If common blocks are not entities, then they should not become accessible.
Given the discussion which followed the patch, I changed this to rejects-valid.
Subject: Bug 16336 CVSROOT: /cvs/gcc Module name: gcc Changes by: tobi@gcc.gnu.org 2004-07-10 22:37:16 Modified files: gcc/fortran : ChangeLog decl.c gfortran.h match.c match.h module.c symbol.c trans-common.c Log message: PR fortran/16336 * decl.c (gfc_match_save): Use-associated common block doesn't collide. * gfortran.h (gfc_common_head): Add new field 'name'. Fix typo in comment after #endif. * match.c (gfc_get_common): Add new argument from_common, mangle name if flag is set, fill in new field in structure gfc_common_head. (match_common): Set new arg in call to gfc_get_common, use-associated common block doesn't collide. * match.h (gfc_get_common): Adapt prototype. * module.c (load_commons): Set new arg in call to gfc_get_common. * symbol.c (free_common_tree): New function. (gfc_free_namespace): Call new function. * trans-common.c (several functions): Remove argument 'name', use name from gfc_common_head instead. Patches: http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/fortran/ChangeLog.diff?cvsroot=gcc&r1=1.113&r2=1.114 http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/fortran/decl.c.diff?cvsroot=gcc&r1=1.17&r2=1.18 http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/fortran/gfortran.h.diff?cvsroot=gcc&r1=1.18&r2=1.19 http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/fortran/match.c.diff?cvsroot=gcc&r1=1.16&r2=1.17 http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/fortran/match.h.diff?cvsroot=gcc&r1=1.8&r2=1.9 http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/fortran/module.c.diff?cvsroot=gcc&r1=1.10&r2=1.11 http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/fortran/symbol.c.diff?cvsroot=gcc&r1=1.9&r2=1.10 http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/fortran/trans-common.c.diff?cvsroot=gcc&r1=1.11&r2=1.12
Fixed. I hope Erik agrees this time :-)
Subject: Bug 16336 CVSROOT: /cvs/gcc Module name: gcc Changes by: tobi@gcc.gnu.org 2004-07-10 22:59:58 Modified files: gcc/testsuite : ChangeLog Added files: gcc/testsuite/gfortran.fortran-torture/execute: common_2.f90 Log message: PR fortran/16336 * gfortran.fortran-torture/execute/common_2.f90: New test. Patches: http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/ChangeLog.diff?cvsroot=gcc&r1=1.3980&r2=1.3981 http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/gfortran.fortran-torture/execute/common_2.f90.diff?cvsroot=gcc&r1=NONE&r2=1.1
*** Bug 16479 has been marked as a duplicate of this bug. ***