The latest ebuild for Perl (perl-5.6.1-r2.ebuild) does not build properly. I get the following failure: make[1]: Leaving directory `/var/tmp/portage/perl-5.6.1-r2/work/perl-5.6.1/ext/re' Making Errno (nonxs) make[1]: Entering directory `/var/tmp/portage/perl-5.6.1-r2/work/perl-5.6.1/ext/Errno' make[1]: Leaving directory `/var/tmp/portage/perl-5.6.1-r2/work/perl-5.6.1/ext/Errno' make[1]: Entering directory `/var/tmp/portage/perl-5.6.1-r2/work/perl-5.6.1/ext/Errno' make[1]: Leaving directory `/var/tmp/portage/perl-5.6.1-r2/work/perl-5.6.1/ext/Errno' cd t && (rm -f perl; /bin/ln -s ../perl perl) if (true </dev/tty) >/dev/null 2>&1; then \ cd t && ./perl TEST </dev/tty; \ else \ cd t && PERL_SKIP_TTY_TEST=1 ./perl TEST; \ fi base/cond............ok base/if..............ok base/lex.............FAILED at test 12 Failed a basic test--cannot continue. make: *** [test] Error 29 !!! ERROR: The ebuild did not complete successfully. !!! Function src_compile, Line 76, Exitcode 2 !!! (no error message) !!! emerge aborting on /usr/portage/sys-devel/perl/perl-5.6.1-r2.ebuild . I am using a pretty up-to-date system. Here is what I get for emerge --pretend --world update (if it's useful): These are the packages that I would merge, in order. Calculating system dependencies.................................................................................................................................................................................................................................................. !!! Warning: couldn't find match for virtual/x11-web-browser in update (likely old /var/db/pkg entry) .............................................. done! [ebuild U] sys-devel/perl-5.6.1-r2 to / [ebuild U] dev-perl/PDL-2.3.2 to / [ebuild U] x11-base/xfree-4.2.0-r5 to / [ebuild U] gnome-base/gnome-print-0.35 to / [ebuild N ] media-libs/glut-3.7-r1 to / [ebuild U] sys-kernel/linux-sources-2.4.17-r3 to / [ebuild U] dev-java/blackdown-jdk-1.3.1-r2 to / Here is the "USE" in make.conf that I'm using: USE="slang readline gpm berkdb gdbm tcpd pam libwww ssl alsa nls mitshm perl python esd gif sdl vorbis ogg gnome gtk X qt kde motif opengl mozilla objprelink cups mmx 3dnow java guile mysql tex bonobo libg++ tcltk lame" I'm also using: CHOST="i586-pc-linux-gnu" CFLAGS="-mcpu=k6 -march=k6 -O3 -pipe" CXXFLAGS="-mcpu=k6 -march=k6 -O3 -pipe" Let me know if there is any other information that is needed. Thanks!
Hmmm, and this build does not fail in the same place twice!!! When I repeat the emerge command, it fails at a different place each time. Example: CCCMD = cc -DPERL_CORE -c -fno-strict-aliasing -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64 -mcpu=k6 -march=k6 -O3 -pipe `sh cflags libperl.a perlio.o` perlio.c CCCMD = cc -DPERL_CORE -c -fno-strict-aliasing -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64 -mcpu=k6 -march=k6 -O3 -pipe perlio.c: In function `PerlIO_setpos': perlio.c:498: incompatible type for argument 2 of `fseeko' make: *** [perlio.o] Error 1 !!! ERROR: The ebuild did not complete successfully. !!! Function src_compile, Line 74, Exitcode 2 !!! (no error message) !!! emerge aborting on /usr/portage/sys-devel/perl/perl-5.6.1-r2.ebuild . So in this case, it doesn't even make it as far as the tests! I've had build problems with other packages (and submitter bugs for those as well, see bugs 417 and 332. In going back to the Mozilla bug, it ALSO fails in different places! (I'll make a note of that in the Mozilla bug as well). I don't think my build environment is totally hosed though, because I just successfully built all the KDE pacakages and it works great. Mystified...
Fourth time is the charm. It built this time around. Very strange. Any clues as to what could be causing this? I don have any problems with systems crashing or stuff like that. Just the builds...
could you check if perl-5.6.1-r1 has similar problems?
The original build (when I did the install for Gentoo) went without a hitch. To try it now, I did the following (so I wouldn't have to unmerge my current (r2) install of perl: ebuild perl-5.6.1-r1.ebuild clean unpack compile This compiled without any trouble. So I would guess that the r1 build did not have this problem. Very strange. It doesn't look like all that much has changed in the script except for some configure options... it's the same perl source code....
The new r3 build works fine. It built the first time through.