* Paul Wise <p...@debian.org>, 2019-03-16, 18:43:
* corekeeper doesn't enforce this limit on its own either.
Should corekeeper use `ulimit -c` to retrieve the limit?
That would give you the limit for the core handler process, not for the
process that crashed.
You can use %c in core_pattern to get the soft limit of the crashing
process.
(Hmm, it's not documented what's the value you get when there's no
limit...)
Should corekeeper enforce the soft or hard limit?
Soft.
Is this an appropriate way to enforce the limit?
I think this should work:
head -c "$limit" > "/var/crash/$owner/$core"
(I'll leave wrapping this in su(1) as an exercise to reader. :-P)
You will also need special cases when the limit is 0, and when there's
no limit.
--
Jakub Wilk