#Blocking on subprocess pipes
1 messages · Page 1 of 1 (latest)
While I doubt that is helping, I'm actually surprised iterating over stdout isn't sufficient to keep the subprocess going. I guess a one byte buffer could be jamming something up in a codec somewhere, since the stream is in text mode.
Certainly worth trying the change, since the read is line-buffered anyway, so the small buffer is just going to be increasing interprocess comms overhead without reducing display latency.
fixed?
No one reported any issues so far
I couldn't replicate it with or without
Pushed an update yesterday
But I can tell that it feels more responsive now, compared to the previous implementation
So there is some truth to this