p5pRT
commented
23 years ago From pvhp@forte.com
`make test` dies on a -des Configured default build of perl on OS/390 (where $usedl='undef' is still the default and DPERL_EXTERNAL_GLOB is used simply as a trick) with the following:
Making re (static) make[2]: Entering directory `/SYSTEM/tmp/ttt/perl/ext/re' make[2]: Leaving directory `/SYSTEM/tmp/ttt/perl/ext/re' make[2]: Entering directory `/SYSTEM/tmp/ttt/perl/ext/re' make[2]: Leaving directory `/SYSTEM/tmp/ttt/perl/ext/re' cd t && (rm -f ./perl; /bin/ln -s .././perl ./perl) make[1]: Leaving directory `/SYSTEM/tmp/ttt/perl' make: *** No rule to make target `lib/auto/File/Glob/Glob.none'\, needed by `test'. Stop.
There appears to be a new bug in Makefile.SH\, I note the following in my extracted Makefile:
$ grep EXTGLOB Makefile EXTGLOB = lib/auto/File/Glob/Glob$(DLSUFFIX) TESTREQEXT = $(EXTGLOB)
A workaround for `make test` appears to be to touch the requisite file (why no suych make rule in Makefile?):
$ ls -F lib/auto/File/Glob Glob.a* autosplit.ix extralibs.ld $ touch lib/auto/File/Glob/Glob.none $ make test
some test failures and # comments:
op/append.............Malformed UTF-8 character (unexpected continuation byte 0x62) in pattern match (m//) at op/append.t line 46. Wide character in print at op/append.t line 46. Malformed UTF-8 character (unexpected continuation byte 0x62) in pattern match (m//) at op/append.t line 50. Wide character in print at op/append.t line 50. Malformed UTF-8 character (unexpected continuation byte 0x62) in pattern match (m//) at op/append.t line 54. Wide character in print at op/append.t line 54. Malformed UTF-8 character (unexpected continuation byte 0x62) in pattern match (m//) at op/append.t line 58. Wide character in print at op/append.t line 58. FAILED at test 7
op/pat................CEE5213S The signal SIGPIPE was received. FAILED at test 234 # UTF-EBCDIC bug in perl regular expressions?
op/tr.................CEE5213S The signal SIGPIPE was received. FAILED at test 50
pragma/locale.........Unmatched [ before HERE mark in regex m/[ \<\< HERE / at pragma/locale.t line 710. FAILED at test 99 # known seen before - memory or C RTL setlocale() bug
pragma/warnings.......PROG: -W
# perl.c no warnings 'once' ; $x = 3 ; use warnings 'once' ; $z = 3 ; EXPECTED: Name "main::z" used only once: possible typo at - line 6. Name "main::x" used only once: possible typo at - line 4. GOT: Name "main::x" used only once: possible typo at - line 4. Name "main::z" used only once: possible typo at - line 6. CEE5213S The signal SIGPIPE was received. FAILED at test 74 # seen before\, possible patch # that would need s/72/73/ modification # is at:
http://www.xray.mpe.mpg.de/mailing-lists/perl-mvs/2001-03/msg00099.html
lib/b.................CEE5213S The signal SIGPIPE was received. FAILED at test 17 # new failure (IIRC)
lib/glob-basic........FAILED at test 11 # new failure ?
lib/io_unix...........Can't call method "getline" on an undefined value at lib/io_unix.t line 65. FAILED at test 3 # known seen before\, fork problem?
lib/mimeqp............FAILED at test 2 # known seen before\, an EBCDIC workaround exists but needs developement
Failed 9 test scripts out of 324\, 97.22% okay. u=18.23 s=6.07 cu=405.26 cs=135.08 scripts=324 tests=23104
One completely untested "hack around" might be to:
Inline Patch
```diff --- Makefile.SH;1 Fri Apr 6 14:07:00 2001 +++ Makefile.SH Fri Apr 6 16:43:51 2001 @@ -135,7 +135,12 @@ case "$DPERL_EXTERNAL_GLOB" in "") ;; -*) testreqext="$testreqext \$(EXTGLOB)" ;; +*) + case "${osname}${osvers}" in + os390*) ;; + *) testreqext="$testreqext \$(EXTGLOB)" ;; + esac + ;; esac echo "Extracting Makefile (with variable substitutions)" ```End of "hack:"
I am sorry - I actually do not like that and it is untested. I am primarily trying to point out the trouble spot. I'd appreciate it if a Makefile.SH expert took a crack at it. Thank you. (Running out of the office for the weekend now).
Migrated from rt.perl.org#6760 (status was 'resolved')
Searchable as RT6760$