Re: [parisc-linux] Re: Bug#342545: qt-x11-free FTBFS
That would be wonderful if you, or another hppa porter, could track down where the bug lies. libgcc2 is almost certainly the wrong package, since nothing should be *using* libgcc2 in a fresh build of qt-x11-free; it may be a bug in libgcc4 instead, but I think that's yet to be determined. In the meantime, I think it's best to reassign this back to qt-x11-free. This might be a nan bug. There is one GCC nan fix that's only installed on the trunk: 2006-05-24 John David Anglin [EMAIL PROTECTED] PR target/27627 * pa/pa-modes.def: Use mips_single_format, mips_double_format and mips_quad_format formats instead of ieee_single_format, ieee_double_format and ieee_quad_format formats, respectively. However, I think the real bug is here: 0x40cb2150 in negNan () at tools/qlocale.cpp:131 131 *((const double *) le_neg_nan_bytes)); PA-RISC requires strict alignment and it's highly likely that the pointer le_neg_nan_bytes isn't aligned to an eight byte boundary. You could see the faulting insn by disassembling around 0x40cb2150 to be sure. The nan problem fixed by the above change would cause a SIGFPE instead of a SIGBUS. I'm fairly certain we have a bug in handling unaligned fixups for doubles in the kernel. This caused a problem for libffi. This depends on whether the kernel is 32/64 bits. As Kyle pointed out, unaligned fixups by the kernel are expensive and they should be avoided. tools/qlocale.cpp appears to be a qt-x11-free routine, so I agree that the reassignment was correct. Dave -- J. David Anglin [EMAIL PROTECTED] National Research Council of Canada (613) 990-0752 (FAX: 952-6602) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [parisc-linux] Re: Bug#342545: qt-x11-free FTBFS
On Thu, Aug 24, 2006 at 08:56:40PM -0400, John David Anglin wrote: I'm fairly certain we have a bug in handling unaligned fixups for doubles in the kernel. This caused a problem for libffi. This depends on whether the kernel is 32/64 bits. I'll try to come up with some testcases. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [parisc-linux] Re: Bug#342545: qt-x11-free FTBFS
On Thu, Aug 24, 2006 at 09:50:23PM -0400, John David Anglin wrote: This might be a nan bug. There is one GCC nan fix that's only installed on the trunk: 2006-05-24 John David Anglin [EMAIL PROTECTED] PR target/27627 * pa/pa-modes.def: Use mips_single_format, mips_double_format and mips_quad_format formats instead of ieee_single_format, ieee_double_format and ieee_quad_format formats, respectively. Just saw your patch. Watch out, there are at least two different representations for nans. In GCC, they are called mips and ieee. However, as far as I can tell, PA-RISC used the mips format before mips. Both formats are complient with the original IEEE standard, so it's also a bit of a misnomer to call the other format the IEEE format. Uh, my patch only attempts to correct for the bad alignment assumptions in the existing code that cause build failures on hppa; invalid assumptions about the byte representations of -NaN on particular platforms can be someone else's problem if and when it becomes an issue. ;) -- Steve Langasek Give me a lever long enough and a Free OS Debian Developer to set it on, and I can move the world. [EMAIL PROTECTED] http://www.debian.org/ signature.asc Description: Digital signature
Re: [parisc-linux] Re: Bug#342545: qt-x11-free FTBFS
Steve, This might be a nan bug. There is one GCC nan fix that's only installed on the trunk: 2006-05-24 John David Anglin [EMAIL PROTECTED] PR target/27627 * pa/pa-modes.def: Use mips_single_format, mips_double_format and mips_quad_format formats instead of ieee_single_format, ieee_double_format and ieee_quad_format formats, respectively. Just saw your patch. Watch out, there are at least two different representations for nans. In GCC, they are called mips and ieee. However, as far as I can tell, PA-RISC used the mips format before mips. Both formats are complient with the original IEEE standard, so it's also a bit of a misnomer to call the other format the IEEE format. Dave -- J. David Anglin [EMAIL PROTECTED] National Research Council of Canada (613) 990-0752 (FAX: 952-6602) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Bug#342545: qt-x11-free FTBFS
The qt-x11-free package builds fine with a standard Debian setup. Building with prctl --unaligned=signal makes the bug reproducible. Christopher Martin writes: reassign 342545 libgcc2 stop On Thursday 10 August 2006 00:25, Steve Langasek wrote: It hasn't been, because I can't see any way that libglu1-mesa could have anything to do with the failure in question. libglu1-mesa should not be a dependency of the tool that's failing with SIGBUS in the build log. I would suggest that someone should investigate this further and get a clear answer on the nature of the bug, because I really don't buy that libgcc skew is to blame. Fair enough, but before I take off for the weekend, I'm sending this report back to libgcc2, since it seems to have been established long ago that this isn't a Qt bug, and it really should be assigned to something in the toolchain. I note that, for a time, the problem was thought to be in glibc, so perhaps the glibc team would again be worth consulting. Cheers, Christopher Martin -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Bug#342545: qt-x11-free FTBFS
On Wed, Aug 23, 2006 at 10:39:04AM +0200, Matthias Klose wrote: The qt-x11-free package builds fine with a standard Debian setup. Building with prctl --unaligned=signal makes the bug reproducible. Right. The buildd is set up to deliver SIGBUS on unaligned accesses. This is configurable, and not the default kernel behaviour. It is, however, a Good Thing(tm). Unaligned accesses are quite costly, so catching and fixing them on the buildd is ideal. However, in most cases they are nontrivial to fix and should be rebuilt and uploaded and a bug filed with upstream... When I initially added prctl for parisc, iirc there were three options logging an unaligned access, delivering a SIGBUS, or silently catching it and continuing on. I have a machine that can be used to do these rebuilds, if need be. Cheers, Kyle -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Bug#342545: qt-x11-free FTBFS
reassign 342545 qt-x11-free thanks On Wed, Aug 23, 2006 at 07:52:48AM -0400, Kyle McMartin wrote: On Wed, Aug 23, 2006 at 10:39:04AM +0200, Matthias Klose wrote: The qt-x11-free package builds fine with a standard Debian setup. Building with prctl --unaligned=signal makes the bug reproducible. Right. The buildd is set up to deliver SIGBUS on unaligned accesses. This is configurable, and not the default kernel behaviour. It is, however, a Good Thing(tm). Unaligned accesses are quite costly, so catching and fixing them on the buildd is ideal. However, in most cases they are nontrivial to fix and should be rebuilt and uploaded and a bug filed with upstream... When I initially added prctl for parisc, iirc there were three options logging an unaligned access, delivering a SIGBUS, or silently catching it and continuing on. I have a machine that can be used to do these rebuilds, if need be. That would be wonderful if you, or another hppa porter, could track down where the bug lies. libgcc2 is almost certainly the wrong package, since nothing should be *using* libgcc2 in a fresh build of qt-x11-free; it may be a bug in libgcc4 instead, but I think that's yet to be determined. In the meantime, I think it's best to reassign this back to qt-x11-free. Thanks, -- Steve Langasek Give me a lever long enough and a Free OS Debian Developer to set it on, and I can move the world. [EMAIL PROTECTED] http://www.debian.org/ signature.asc Description: Digital signature
Re: Bug#342545: qt-x11-free FTBFS
On Wed, Aug 23, 2006 at 05:22:45PM -0700, Steve Langasek wrote: That would be wonderful if you, or another hppa porter, could track down where the bug lies. libgcc2 is almost certainly the wrong package, since nothing should be *using* libgcc2 in a fresh build of qt-x11-free; it may be a bug in libgcc4 instead, but I think that's yet to be determined. In the meantime, I think it's best to reassign this back to qt-x11-free. Agreed, probably the best course of action. I'll kick off a build in my sid chroot with prctl set to dump core and we'll see what happens. Cheers, Kyle -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Bug#342545: qt-x11-free FTBFS on hppa (again)
reassign 342545 qt-x11-free thanks please make sure, that qt-x11-free is built using g++-4.1 on hppa. The binary packages still depend on libgcc2 in some way. Note, that (before the release) we need to rebuild all binaries depending on libgcc2 on hppa, so that the dependency is replaced by libgcc4. Matthias Christopher Martin writes: reopen 342545 2.3.6-15 severity 342545 grave stop Unfortunately, the qt-x11-free FTBFS on hppa has re-occurred: http://buildd.debian.org/fetch.php?pkg=qt-x11-freever=3%3A3.3.6-3arch=hppastamp=1153689954file=logas=raw The error is exactly the same as before: /build/buildd/qt-x11-free-3.3.6/bin/uic -L /build/buildd/qt-x11-free-3.3.6/plugins pixmapfunction.ui -o pixmapfunction.h make[4]: *** [pixmapfunction.h] Bus error Cheers, Christopher Martin -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
