Bug#428876: [Fwd: Bug#428876: Testcase for 428876]
On Mon, Oct 15, 2007 at 08:14:13PM +, brian m. carlson wrote: On Thu, Oct 11, 2007 at 05:41:58PM +0200, Joachim Reichel wrote: I do not have acces to a sparc machine and I can't use the developer machines since I'm no DD (yet). I build the package in a current sid chroot on i386 and uploaded history.ps to http://www.joachim-reichel.de/tmp/history.ps . Can you try if this file triggers the problem? This file does trigger the bug. I am building gs-esp right now so that I can explore the problem further. fwiw building in -O0 let gs work. Though it probably hides the issue, and it's likely that it's not a compiler problem as it affects many archs. Though it could be used as a temporary work-around. -- ·O· Pierre Habouzit ··O[EMAIL PROTECTED] OOOhttp://www.madism.org pgpgRboyGn3IG.pgp Description: PGP signature
Bug#428876: [Fwd: Bug#428876: Testcase for 428876]
On Tue, Nov 27, 2007 at 05:00:20PM +0100, Pierre Habouzit wrote: On Mon, Oct 15, 2007 at 08:14:13PM +, brian m. carlson wrote: This file does trigger the bug. I am building gs-esp right now so that I can explore the problem further. fwiw building in -O0 let gs work. Though it probably hides the issue, and it's likely that it's not a compiler problem as it affects many archs. If it causes a SIGBUS on sparc, it is almost certainly an unaligned access; the only other alternatives are a misuse of mmap(2) or broken hardware. I'm not sure why -O0 works on sparc if it's an unaligned access, since the compiler should lay out data in the same way regardless of optimization. It's possible that the unaligned access only shows itself because the compiler only optimizes the access under those circumstances. I'm also surprised to see this not happening on alpha, since that's one of the architectures that is set up for unaligned traps (which on Debian buildds, cause a signal). When I looked at the problem last, it seemed to be pretty intricate, with some weird interactions between C and PostScript code. Though it could be used as a temporary work-around. It's entirely up to the maintainer, but this seems to be a difficult problem, so it might be a good idea. -- brian m. carlson / brian with sandals: Houston, Texas, US +1 713 440 7475 | http://crustytoothpaste.ath.cx/~bmc | My opinion only a typesetting engine: http://crustytoothpaste.ath.cx/~bmc/code/thwack OpenPGP: RSA v4 4096b 88AC E9B2 9196 305B A994 7552 F1BA 225C 0223 B187 signature.asc Description: Digital signature
Bug#428876: [Fwd: Bug#428876: Testcase for 428876]
On Thu, Oct 11, 2007 at 05:41:58PM +0200, Joachim Reichel wrote: I do not have acces to a sparc machine and I can't use the developer machines since I'm no DD (yet). I build the package in a current sid chroot on i386 and uploaded history.ps to http://www.joachim-reichel.de/tmp/history.ps . Can you try if this file triggers the problem? This file does trigger the bug. I am building gs-esp right now so that I can explore the problem further. -- brian m. carlson / brian with sandals: Houston, Texas, US +1 713 440 7475 | http://crustytoothpaste.ath.cx/~bmc | My opinion only a typesetting engine: http://crustytoothpaste.ath.cx/~bmc/code/thwack OpenPGP: RSA v4 4096b 88AC E9B2 9196 305B A994 7552 F1BA 225C 0223 B187 signature.asc Description: Digital signature
Bug#428876: Testcase for 428876
Hi, I do not have acces to a sparc machine and I can't use the developer machines since I'm no DD (yet). I build the package in a current sid chroot on i386 and uploaded history.ps to http://www.joachim-reichel.de/tmp/history.ps . Can you try if this file triggers the problem? Regards, Joachim -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#428876: Testcase for 428876
Can you send me a testcase that fails on sparc, so I can investigate? I'd build core++, but I really don't want to install texlive on my sparc machine if I don't have to (it's an old, hence slow, Ultra 5). I suspect the problem on sparc is an unaligned access, which should be fairly easy to fix. -- brian m. carlson / brian with sandals: Houston, Texas, US +1 713 440 7475 | http://crustytoothpaste.ath.cx/~bmc | My opinion only a typesetting engine: http://crustytoothpaste.ath.cx/~bmc/code/thwack OpenPGP: RSA v4 4096b 88AC E9B2 9196 305B A994 7552 F1BA 225C 0223 B187 signature.asc Description: Digital signature