[pacman-dev] Segfault at end of transaction?
I'm guess no-one here has seen this (or why have you not reported it yet!!!), but a couple of people on the Chakra forums are reporting a segfault at the end of the transaction - after releasing the databases. See http://chakra-project.org/bbs/viewtopic.php?id=10224 Waiting for more backtrace information. Allan
Allan McRae <allan@archlinux.org> on Thu, 2013/04/25 17:21:
I'm guess no-one here has seen this (or why have you not reported it yet!!!), but a couple of people on the Chakra forums are reporting a segfault at the end of the transaction - after releasing the databases.
See http://chakra-project.org/bbs/viewtopic.php?id=10224
Waiting for more backtrace information.
I saw this two or three times back when I ran git versions. But as it does no harm and I have no idea how to reproduce it... Did not report it yet. And I have not seen it for a long time, so no idea if this is still valid. -- main(a){char*c=/* Schoene Gruesse */"B?IJj;MEH" "CX:;",b;for(a/* Chris get my mail address: */=0;b=c[a++];) putchar(b-1/(/* gcc -o sig sig.c && ./sig */b/42*2-3)*42);}
On 04/25/13 09:21, Allan McRae wrote:
I'm guess no-one here has seen this (or why have you not reported it yet!!!), but a couple of people on the Chakra forums are reporting a segfault at the end of the transaction - after releasing the databases.
Just a guess, but what version of curl are they running? curl 7.29 has a bug that would cause it to segfault when cleaning and no operation/download was done. As that seems to be the case here, it could be curl segfaulting when libalpm calls curl's cleaning function. It was patched in Arch, and now we're at 7.30, which could explain why no one noticed it here. -j
Waiting for more backtrace information.
Allan
On 25/04/13 18:56, jjacky wrote:
On 04/25/13 09:21, Allan McRae wrote:
I'm guess no-one here has seen this (or why have you not reported it yet!!!), but a couple of people on the Chakra forums are reporting a segfault at the end of the transaction - after releasing the databases.
Just a guess, but what version of curl are they running? curl 7.29 has a bug that would cause it to segfault when cleaning and no operation/download was done.
As that seems to be the case here, it could be curl segfaulting when libalpm calls curl's cleaning function.
It was patched in Arch, and now we're at 7.30, which could explain why no one noticed it here.
http://www.chakra-project.org/packages/index.php?act=show&subdir=core/x86_64&sortby=date&file=curl-7.28.1-1-x86_64.pkg.tar.xz Was the bug in 7.28.1?
On 04/25/13 11:00, Allan McRae wrote:
On 25/04/13 18:56, jjacky wrote:
On 04/25/13 09:21, Allan McRae wrote:
I'm guess no-one here has seen this (or why have you not reported it yet!!!), but a couple of people on the Chakra forums are reporting a segfault at the end of the transaction - after releasing the databases.
Just a guess, but what version of curl are they running? curl 7.29 has a bug that would cause it to segfault when cleaning and no operation/download was done.
As that seems to be the case here, it could be curl segfaulting when libalpm calls curl's cleaning function.
It was patched in Arch, and now we're at 7.30, which could explain why no one noticed it here.
Was the bug in 7.28.1?
No, 7.29 only. Guess it's not that then.
On Thu, Apr 25, 2013 at 05:21:38PM +1000, Allan McRae wrote:
I'm guess no-one here has seen this (or why have you not reported it yet!!!), but a couple of people on the Chakra forums are reporting a segfault at the end of the transaction - after releasing the databases.
See http://chakra-project.org/bbs/viewtopic.php?id=10224
Waiting for more backtrace information.
Allan
We saw this with FTP only, and the debug log on that thread shows FTP mirrors as the first. I'm fairly sure that it's a bug with the FTP state machine in curl. d
participants (4)
-
Allan McRae
-
Christian Hesse
-
Dave Reisner
-
jjacky