While emerging PyQt I'll have errors complaining about operand being out of range: make[1]: Entering directory `/var/tmp/portage/PyQt-3.5/work/PyQt-x11-gpl-3.5/qt' g++ -c -pipe -w -O2 -D_REENTRANT -fPIC -DSIP_MAKE_MODULE_DLL -DQT_NO_DEBUG -DQT_THREAD_SUPPORT -I. -I/usr/in clude/python2.2 -I/usr/qt/3/include -I/usr/qt/3/mkspecs/linux-g++ -o qthuge.o qthuge.cpp {standard input}: Assembler messages: {standard input}:479752: Error: operand out of range (0000000000008000 not between -32768 and 32767) {standard input}:479755: Error: operand out of range (0000000000008004 not between -32768 and 32767) and so on ... Reproducible: Always Steps to Reproduce: 1. emerge PyQt Expected Results: Properly building and emerging PyQt Portage 2.0.46-r12 (default-ppc-1.0, gcc-2.95.3, glibc-2.2.5-r7) ================================================================= System uname: 2.4.19-ben0 ppc GENTOO_MIRRORS="http://gentoo.oregonstate.edu/ http://www.ibiblio.org/pub/Linux/distributions/gentoo" CONFIG_PROTECT="/etc /var/qmail/control /usr/kde/2/share/config /usr/kde/3/share/config /usr/kde/3.1/share/config /usr/share/config" CONFIG_PROTECT_MASK="/etc/gconf /etc/env.d" PORTDIR="/usr/portage" DISTDIR="/usr/portage/distfiles" PKGDIR="/usr/portage/packages" PORTAGE_TMPDIR="/var/tmp" PORTDIR_OVERLAY="" USE="oss dvd xv libwww jpeg nls mitshm gif gnome kde png truetype ppc gnome-libs gdbm berkdb slang readline arts ggi tcltk java guile X sdl gpm tcpd pam ssl perl python esd imlib oggvorbis gtk qt motif opengl mozilla" COMPILER="" CHOST="powerpc-unknown-linux-gnu" CFLAGS=" -O2 -pipe -mcpu=750 -mpowerpc-gfxopt -mmultiple -mstring " CXXFLAGS=" -O2 -pipe -mcpu=750 -mpowerpc-gfxopt -mmultiple -mstring " ACCEPT_KEYWORDS="ppc" MAKEOPTS="-j2" AUTOCLEAN="yes" SYNC="rsync://rsync.gentoo.org/gentoo-portage" FEATURES="ccache" I'm using an iBook2 with a 500 MHz G3 and 384 Mb of RAM with gentoo 1.2.
Created attachment 8680 [details] a shortened log of "emerge PyQt" Just removed the >45000 lines of: {standard input}:1817831: Error: operand out of range (0000000000036c8c not between -32768 and 32767)
Created attachment 8681 [details] a shortened log of "emerge PyQt" Just removed the >45000 lines of: {standard input}:1817831: Error: operand out of range (0000000000036c8c not between -32768 and 32767)
Created attachment 8682 [details] a shortened log of "emerge PyQt" Just removed the >45000 lines of: {standard input}:1817831: Error: operand out of range (0000000000036c8c not between -32768 and 32767)
Created attachment 8683 [details] a shortened log of "emerge PyQt" Just removed the >45000 lines of: {standard input}:1817831: Error: operand out of range (0000000000036c8c not between -32768 and 32767)
Sorry for the multiple attachement bur I got each time a: Status: 400 Bad request (malformed multipart POST) message ... yet another bug ? Cheerio, Guillaume
My first suggestion would be if you could test your system for bad RAM.. You can do that by emerging memtest86, and running that.. Thanks!
memtest86 on a ppc box, seems to be a joke. Anyway I doubt that it's a memory problem seems more like casting some long in an int ...
verwilst is a ppc n00b, dont mind him ;) anyways, what if you mess with the CFLAGS ? many times assembler messages like that are related to CFLAGS ... env CFLAGS="" emerge PyQt
Just tried it w/ CFLAGS="": same results.
There is a ppc memtester in tree, the name escapes me right now. It might still be good to test this :)
I've let memtest run overnight and it didn't report find a single error ... next suggestion ?
have you tried updating gcc? looks like you ahve a fairly old version of it installed (of the 2.95.3 branch). Try updating to the -r8 I think it is?
I just took a closer look to my 'emerge info' output and it look kinda weird: Portage 2.0.46-r12 (default-ppc-1.0, gcc-2.95.3, glibc-2.2.5-r7) ... Now if I take a look at the ggc version installed 'emerge -p gcc', I get gcc-2.95.3-r7 ... I'm going to update to gcc-2.95.3-r8 (BTW gcc-2.95.3-r7 is still the default).
I've emerged gcc-2.95.3-r8 and I still got the error ... I'm going to compile it from last cvs sources this WE !
Finally I got some time to proceed on tracing the bug. I've compiled w/o any problem the same version downloaded on the website with the following commands: $ python build.py -l qt $ make or: $ python build.py -l qt-mt $ make and there was no problems at all so I guess it's within the gentoo version ...
Is still that bug valid?
The bug is still valid, there was no update to the package and I can reproduce it on my system w/o problems.
I think I found the cause of the bug: the use of the -c switch for the build.py script causes problems when used with gcc-2.95.3. As I was on the process to upgrade to gentoo 1.4, I've tested the default ebuild and the compilation caused no error. I still have to check if removing the -c switch causes problems to gcc-3.2.2, if not I'll submit a patch by tomorrow night. Guillaume
Has anyone else been able to reproduce this? Knowing the solution is wonderful, but unless there are a few people with the problem, it's best not to risk breaking it by applying the fix to everyone
The real problem for reproducing it is that it seems to depend on the gcc version (here 2.95.3), and is probably a gcc bug. The next thing I do not really understand is why the ebuild use this -c switch ...
I've tried removing the -c switch and tested it with gcc-3.2.2 works fine. Does someone know why this -c option is there ? From the description I do not see the interest of using it. I think that if there is no real reason to be there it should be dropped or the package should be marked as not suitable for gentoo 1.2 for PPC.
seems that was a binutils issue. here emerged correctly the 3.6 closing it please reopen it if it is still an issue
