Debugging under Linux: Difference between revisions
Jump to navigation
Jump to search
TomEdwards (talk | contribs) (this was posted on hlcoders) |
TomEdwards (talk | contribs) m (→Memory dumps) |
||
| Line 15: | Line 15: | ||
When you add -debug parameter to srcds_run the ulimit is called with 2000 kilo bytes (like my example). {{todo|Does this franglais mean that <code>-debug</code> allows you to skip the ulimit command?}} | When you add -debug parameter to srcds_run the ulimit is called with 2000 kilo bytes (like my example). {{todo|Does this franglais mean that <code>-debug</code> allows you to skip the ulimit command?}} | ||
When the server crashes, a new file appear, called "core". You can launch [sourceware.org/gdb/ gdb] on this coredump with this command: | When the server crashes, a new file appear, called "core". You can launch [http://sourceware.org/gdb/ gdb] on this coredump with this command: | ||
gdm ./srcds_i686 core | gdm ./srcds_i686 core | ||
Revision as of 13:21, 5 July 2008
Runtime debugging
TODO!
Memory dumps
Linux memory dumps are created and handled by coredump.
On many GNU/Linux distribution it is disabled by default, for security reasons. A simple command can be change this limitation :
ulimit -c 2000
2000 is in kilobytes. The command must be applied before the launch of the server.
When you add -debug parameter to srcds_run the ulimit is called with 2000 kilo bytes (like my example).
Todo: Does this franglais mean that
-debug allows you to skip the ulimit command?When the server crashes, a new file appear, called "core". You can launch gdb on this coredump with this command:
gdm ./srcds_i686 core
(Replace srcds_i686 by the executable really used on your system.)