Summary: | sys-devel/gcc-3.4.4 causes string corruption | ||
---|---|---|---|
Product: | Portage Development | Reporter: | Jakub Moc (RETIRED) <jakub> |
Component: | Tools | Assignee: | Gentoo Toolchain Maintainers <toolchain> |
Status: | RESOLVED UPSTREAM | ||
Severity: | minor | CC: | ka0ttic |
Priority: | High | ||
Version: | unspecified | ||
Hardware: | All | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Attachments: |
herdstat -Dm vlc output
failed tests output result of tests after gcc-3.4.4 recompile |
Description
Jakub Moc (RETIRED)
2005-07-13 11:30:31 UTC
hmm. I'm doing to need some kind of info to take this any further. All tests pass on all 3 of my boxes (nor can I reproduce). Post your emerge --info. Attach the output of 'herdstat -Dm vlc'. Attach the exact output from beginning of the tests to the end of the tests. (In reply to comment #1) > Post your emerge --info. Portage 2.0.51.22-r1 (default-linux/x86/2005.0, gcc-3.4.4, glibc-2.3.5-r0, 2.6.11-gentoo-r11 i686) ================================================================= System uname: 2.6.11-gentoo-r11 i686 AMD Sempron(tm) Gentoo Base System version 1.6.12 ccache version 2.4 [enabled] dev-lang/python: 2.3.5, 2.4.1-r1 sys-apps/sandbox: 1.2.10 sys-devel/autoconf: 2.13, 2.59-r6 sys-devel/automake: 1.4_p6, 1.5, 1.6.3, 1.7.9-r1, 1.8.5-r3, 1.9.5 sys-devel/binutils: 2.15.92.0.2-r10, 2.16.1 sys-devel/libtool: 1.5.18-r1 virtual/os-headers: 2.6.11-r2 ACCEPT_KEYWORDS="x86" AUTOCLEAN="yes" CBUILD="i686-pc-linux-gnu" CFLAGS="-march=athlon-xp -O2 -pipe -fomit-frame-pointer -fforce-addr" CHOST="i686-pc-linux-gnu" CONFIG_PROTECT="/etc /opt/glftpd/etc /opt/glftpd/ftp-data /usr/kde/2/share/config /usr/kde/3/share/config /usr/share/config /var/qmail/control" CONFIG_PROTECT_MASK="/etc/gconf /etc/terminfo /etc/env.d" CXXFLAGS="-march=athlon-xp -O2 -pipe -fomit-frame-pointer -fforce-addr" DISTDIR="/usr/portage/distfiles" FEATURES="autoconfig ccache collision-protect distlocks maketest sandbox sfperms strict test userpriv usersandbox" GENTOO_MIRRORS="ftp://ftp.sh.cvut.cz/MIRRORS/gentoo/gentoo ftp://ftp.fi.muni.cz/pub/linux/gentoo/" LANG="en_US.UTF-8" LINGUAS="en" MAKEOPTS="-j2" PKGDIR="/usr/portage/packages" PORTAGE_TMPDIR="/var/tmp" PORTDIR="/usr/portage" PORTDIR_OVERLAY="/usr/local/portage" SYNC="rsync://rsync.europe.gentoo.org/gentoo-portage" USE="x86 3dnow acpi apache2 apm bash-completion bcmath berkdb bzip2 calendar crypt curl curlwrappers exif fbcon fortran freetds ftp gd gdbm gif iconv imap innodb jpeg libg++ libwww maildir mcal memlimit mhash mime mmx mp3 mpeg mssql mysql mysqli ncurses nls nptl nptlonly odbc offensive pam pcntl pcre pdflib perl php png pnp posix python readline sasl session simplexml soap sockets spl sqlite sse ssl sysvipc tcpd tidy tiff tokenizer unicode wddx xml xml2 xmlrpc xsl zlib linguas_en userland_GNU kernel_linux elibc_glibc" Unset: ASFLAGS, CTARGET, LC_ALL, LDFLAGS > Attach the output of 'herdstat -Dm vlc'. > Attach the exact output from beginning of the tests to the end of the tests. See below.... Created attachment 63333 [details]
herdstat -Dm vlc output
Created attachment 63334 [details]
failed tests output
libs versions as requested: # emerge -pv libxml2 libxmlpp glibmm libsigc++ These are the packages that I would merge, in order: Calculating dependencies ...done! [ebuild R ] dev-libs/libxml2-2.6.19 -ipv6 +python +readline 0 kB [ebuild R ] dev-cpp/libxmlpp-2.10.0-r1 -debug -doc 0 kB [ebuild R ] dev-cpp/glibmm-2.6.1 -debug -doc 0 kB [ebuild R ] dev-libs/libsigc++-2.0.14 -debug -doc -test 0 kB OK, I tried on four almost identical boxes. It works on the two with gcc-3.3.5-20050130 (no broken tests, no bug) and is equally broken on those two with gcc-3.4.4. All use identical C[XX]FLAGS, the only major difference being gcc and glibc versions. I also tried to switch to the other binutils version on two gcc-3.4.4 boxes that have both installed but to no avail. Seems like a gcc bug to me. Just to make it clearer to the toolchain guys what Jakub is referring to. He was having these problems, and I noticed from the failed tests output that: port@/profiles/categories: No such file or directory Note that 'port@' should be 'portdir' (there's a test PORTDIR in ${S}/tests/portdir). I told him to try a different gcc version other than gcc-3.4.x as I had recieved weird string corruption results with 3.4.x in the past (that always goes away when using 3.3.x or 4). Works fine here with gcc-4.0.1 (amd64). Don't have 3.4.4 to test. it works on my amd64 with 3.4.4 :/ maybe it's related to the chk patch i just punted ? can someone who has this bug sync up, re-emerge gcc, re-emerge herdstat, and then see if the issue still exists ? (In reply to comment #9) > maybe it's related to the chk patch i just punted ? can someone who has this > bug sync up, re-emerge gcc, re-emerge herdstat, and then see if the issue still > exists ? Ouch! Seems much worse, now 8 out of 11 tests fail after emerge sync and recompile gcc-3.4.4 :( Created attachment 63601 [details]
result of tests after gcc-3.4.4 recompile
If its a regex failure, you might want to rebuild glibc among some other things as well with the new gcc ? (In reply to comment #12) > If its a regex failure, you might want to rebuild glibc among some other things > as well with the new gcc ? > OK, I did 'FEATURES="-maketest -test" emerge glibc libxml2 libxmlpp glibmm libsigc++ && emerge herdstat and it still fails 8 out of 11 tests. *shrug* OK, an update here: it passes all tests and works perfectly after upgrading the box to gcc-4.0.2-beta20050804 and glibc-2.3.5.20050722-r0 and recompiling herdstat and its dependencies with the new toolchain. So for me, gcc-3.4.4 is b0rked. Changed summary and re-assigned, not a herdstat bug. In a fresh chroot: # emerge -pv gcc These are the packages that I would merge, in order: Calculating dependencies ...done! [ebuild R ] sys-devel/gcc-3.4.4 (-altivec) -bootstrap -boundschecking -build +fortran -gcj +gtk -hardened -ip28 (-multilib) -multislot (-n32) (-n64) +nls -nocxx -nopie -nossp -objc -static +vanilla 0 kB FAIL: metadata-test.sh FAIL: metadata-regex-test.sh FAIL: dev-test.sh FAIL: herd-test.sh FAIL: stats-test.sh FAIL: away-test.sh 6 of 11 tests failed Well, I guess not a Gentoo problem then, should be fixed upstream. We'd need to identify exactly what is being generated wrong to send as a small testcase upstream. If someone can provide that, and the exact command that makes it fail, I'll push it upstream. |