Re: Bison 3.7.90 (beta): test 4
Akim Demaille <[email protected]> Thu, 2 Sep 2021 07:20:51 +0200
| Newsgroups | gmane.comp.parsers.bison.bugs |
|---|---|
| Message-ID | <[email protected]> |
Hi Dagobert, > Le 29 août 2021 à 21:49, Dagobert Michelsen <[email protected]> a écrit : > > Hi Akim, > > Am 29.08.2021 um 10:08 schrieb Akim Demaille <[email protected]>: >> So it is indeed a problem with sed: neither sed.1.log nor sed.2.log have the expected NUL byte. >> >> So I'm installing the patch below. Could you please check this tarball: >> >> https://www.lrde.epita.fr/~akim/private/bison/bison-3.7.90.8-8556.tar.gz >> https://www.lrde.epita.fr/~akim/private/bison/bison-3.7.90.8-8556.tar.lz >> https://www.lrde.epita.fr/~akim/private/bison/bison-3.7.90.8-8556.tar.xz > > Sparc now compiles and tests cleanly, but x86 shows one error in the testsuite: > > Subject: [GNU Bison 3.7.90.8-8556] testsuite: 203 failed Wow... This is very very weird. > 203. report.at:1523: testing Multistart reports ... > ./report.at:1561: COLUMNS=1000; export COLUMNS; NO_TERM_HYPERLINKS=1; export NO_TERM_HYPERLINKS; bison --color=no -fno-caret -o input.cc -rall --graph=input.gv --xml input.yy > --- /dev/null 2021-08-29 20:48:39.000000000 +0200 > +++ /home/dam/mgar/pkg/bison/trunk/work/solaris10-i386/build-isa-pentium_pro/bison-3.7.90.8-8556/tests/testsuite.dir/at-groups/203/stderr 2021-08-29 20:48:44.601359080 +0200 > @@ -0,0 +1,3 @@ > +Assertion failed: !symbols_sorted, file src/symtab.c, line 903, function symbol_from_uniqstr > +gm4:stdin:14: ERROR: end of file in string > +/home/dam/mgar/pkg/bison/trunk/work/solaris10-i386/build-isa-pentium_pro/bison-3.7.90.8-8556/tests/bison: line 33: 8409 Abort (core dumped) $PREBISON "$abs_top_builddir/src/bison" ${1+"$@"} > ./report.at:1561: exit code was 134, expected 0 > 203. report.at:1523: 203. Multistart reports (report.at:1523): FAILED (report.at:1561) I do not understand how there could be portability issues in the area. Besides, none of the sanitizers I use found anything in this test. That being said, it is again a test about a feature I don't want to make public in 3.8, so I will disable this test, and I'll have to solve that issue afterwards. Many thanks!