Closed fhaust closed 5 years ago
Just noticed that the related pyDetach
function is documented as "does not work reliable". Maybe create the process in a Haskell thread and handle forking there?
Yeah.. Guess we're going to have to go the process route! I'll get to it later in the week. I wonder why they have an API call that's not reliable?
Things should work now. There's a race condition in how files a read. I changed python to not detach and to fork in Haskell instead. Let me know if that fixes things.
It works for me.
Hmm ... does not work for me. But honestly I am not sure if I got the latest version. Will check that out tomorrow maybe.
That could be. The stackage nightly and lts haven't been regenerated yet since I pushed this. should be another day for one and a week or os for the other. Let me know.
I actually pointed stack at the github repository. But as I said I am not sure that worked.
On 17425a831b254b4ef2cafdfe2edabf832058882d, I had to threadDelay 1
after onscreen
for the plot to appear. If main
is just onscreen ...
without the threadDelay 1
, the program exits without showing the plot.
Ok. I'll make a more permanent solution with a lockfile to eliminate these race conditions when I get some free time this weekend.
Hi folks, I'm going to close this as it's been a long time. Hopefully the ecosystem is more stable now and these problems have gone away. Feel free to reopen.
The
onscreen
way of displaying graphs does not work for me. Neither by calling it interactively fromghci
nor from a compiled program. The (Haskell) program just stops apparently waiting for the python subprocess to end.If edit the emitted script in the tmp folder manually and remove these lines:
I can call it manually but with them it just ends after a while.