linuxdcpp-team team mailing list archive

[Razzloss <razzloss@xxxxxxxxx>]

And you (or anyone else) hasn't modified the code? It's straight from
our release tarball?

That SIGPIPE shouldn't have been delivered to the program, but I guess
it was because it was run under gdb. So could you try to set 'ulimit -c
unlimited' and wait for the crash again. Then get the 'thread apply all
backtrace full' from the produced core file.

Of course you might as well download the latest trunk from bazaar and
see if it crashes. The files where the crash happens has changed quite
lot since the 1.0.3 release. So even if we figured out what happens
here, it would have to be reproduced with the latest development
version. (Instructions on how to download and compile the latest trunk
can be found from http://ubuntuforums.org/showthread.php?t=193984 . It
is for Ubuntu, but it should apply to Debian without modifications).

--RZ

** Changed in: linuxdcpp
       Status: New => Incomplete

-- 
Crashes with segmentation fault
https://bugs.launchpad.net/bugs/555594
You received this bug notification because you are a member of LinuxDC++
Team, which is subscribed to LinuxDC++.

Status in Linux DC++: Incomplete

Bug description:
When running linuxdcpp, it crashes with segmentation fault. Normally it takes over an hour between the start and crash, and within a day. But I have had periods where it worked for a week. (This means hashing, downloading, seeding etc. all works fine, until the crash).

I have newer succeeded to find out what the program is doing when it crashes, because I don't sit by the computer (server) all time.

I run Debian 5.0 i686, and compiled linuxdcpp 1.0.3 myself, guided by the manual. I had the same problem with 1.0.2.

An interesting thing was, that when running in gdb, the gui was still alive, but frozen, after the crash, while the whole program crashed at once when running linuxdcpp by it's own.