https://gcc.gnu.org/bugzilla/show_bug.cgi?id=124616
--- Comment #6 from Steve Kargl <kargl at gcc dot gnu.org> ---
Hacking the Makefile to find the xml2 headers and fixing the non-posix
linuxism in open.cc leads to
=== cobol Summary ===
# of expected passes 6256
# of unexpected failures 624
# of expected failures 78
# of unresolved testcases 120
/home/kargl/gcc/obj/gcc/gcobol version 16.0.1 20260324 (experimental) (GCC)
% grep "Segmentation fault" gcc/testsuite/cobol/cobol.log | wc -l
342
The first segfault is
FAIL: cobol.dg/group1/check_88.cob -O0 (internal compiler error:
Segmentation fault)
FAIL: cobol.dg/group1/check_88.cob -O0 (test for excess errors)
Excess errors:
cobol1: internal compiler error: Segmentation fault
0x26eb9db internal_error(char const*, ...)
../../gcc/gcc/diagnostic-global-context.cc:787
0x11c2257 crash_signal
../../gcc/gcc/toplev.cc:325
0x82866797b handle_signal
/usr/src/lib/libthr/thread/thr_sig.c:317
0x828666f8a thr_sighandler
/usr/src/lib/libthr/thread/thr_sig.c:261
0xb50f84 charmap_t::charmap_t(cbl_encoding_t)
/home/kargl/gcc/gcc/libstdc++-v3/libsupc++/../../libgcobol/charmaps.h:353
0xb4f967 __gg__get_charmap(cbl_encoding_t)
cobol/charmaps.cc:1760
0xb4fc5f __gg__iconverter(cbl_encoding_t, cbl_encoding_t, void const*, unsigned
long, unsigned long*, unsigned long*)
cobol/charmaps.cc:1532
0xa61a74 parser_symbol_add(cbl_field_t*)
../../gcc/gcc/cobol/genapi.cc:17350
0xb2a227 symbols_update(unsigned long, bool)
../../gcc/gcc/cobol/symbols.cc:1993
0xa990a2 data_division_ready
../../gcc/gcc/cobol/parse_ante.h:3359
0xac756f yyparse()
../../gcc/gcc/cobol/parse.y:1548
0xb40b1c parse_file
../../gcc/gcc/cobol/util.cc:3373
0xb40b1c cobol_parse_files(int, char const**)
../../gcc/gcc/cobol/util.cc:3427
/home/kargl/gcc/obj/gcc/testsuite/cobol/../../cobol1
/home/kargl/gcc/gcc/gcc/testsuite/cobol.dg/group1/check_88.cob -quiet -dumpdir
./ -dumpbase check_88.cob -dumpbase-ext .cob
-main=/home/kargl/gcc/gcc/gcc/testsuite/cobol.dg/group1/check_88.cob
-mtune=generic -march=x86-64 -O0 -fdiagnostics-color=never
-fdiagnostics-urls=never -fdiagnostics-path-format=separate-events
-fdiagnostics-text-art-charset=none -fno-diagnostics-show-caret
-fno-diagnostics-show-line-numbers -fdiagnostics-path-format=separate-events
-fdiagnostics-text-art-charset=none -fno-diagnostics-show-event-links
-fno-diagnostics-show-nesting -o /tmp/ccvgruWY.s
with
% grep "thr_sig.c:261" gcc/testsuite/cobol/cobol.log | wc -l
228
which suggest that threading model with gcobol may needs some attention.