... | ... | @@ -79,8 +79,11 @@ Usually you want to do something along these lines: |
|
|
|
|
|
- Make the crash happen repeatably, if possible. That means after
|
|
|
the exact same number of garbage collections and context switches
|
|
|
each time. Use `+RTS -C0` to turn off timer-based context switches. If
|
|
|
you're debugging heap profiling, add `-i0` too.
|
|
|
each time. Use `+RTS -C0` to turn off timer-based context switches.
|
|
|
Setting `+RTS -V0` disables the RTS timer entirely. Disabling the RTS
|
|
|
timer altogether can make the point at which the program crashes reproducible,
|
|
|
and is very helpful when using the "going back in time" method described below.
|
|
|
If you're debugging heap profiling, add `-i0` too.
|
|
|
|
|
|
- If you still haven't figured it out, you probably need to go back in
|
|
|
time a bit to find out what was happening just before the crash.
|
... | ... | |