Conversation
1611faf to
fb345af
Compare
|
I think this test should be avoided on Windows using something like this. Also, we should probably add a Windows CI runner so that such failures are obvious. |
fb345af to
6c7f12d
Compare
test/unix/concurrent.ml
Outdated
| in | ||
| let pid = Unix.fork () in | ||
| if pid <> 0 then ( | ||
| Unix.sleepf 0.001; |
There was a problem hiding this comment.
this might be a bit fragile, but I guess that's good enough for now. We can restructure this later if we see that this the tests is not deterministic enough.
There was a problem hiding this comment.
Now that we have a Private functor abstraction, could we add a hook for the creation of the lock file to test this more consistently?
It would be nice to have these tests merged so that they're not a maintenance burden to you 🙂
There was a problem hiding this comment.
Thanks for the idea, I removed all calls to Unix.sleep.
2903fa0 to
08a9d79
Compare
08a9d79 to
d99e1b0
Compare
94337ed to
3fcfb43
Compare
41b8c7a to
99936e9
Compare
99936e9 to
26f51cd
Compare
8649a6f to
99c9a1c
Compare
|
I'm looking into the spurious test failures here. Will start with a rebase. |
|
Keeping this open as this seems to be a real synchronisation issue and need more investigation. |
|
@craigfe is planning to have a look at it at one point :-) |
99c9a1c to
8049736
Compare
feb15c4 to
db3fac7
Compare
No description provided.