Race condition in tests? about ptyprocess HOT 4 CLOSED

It seems to happen in about half of the time in test_spawn_sh and so far once in test_spawn_unicode_sh.

from ptyprocess.

I would imagine that the race is between the subprocess responding to the EOF and the parent process calling read(). If the first read doesn't get everything that the subprocess wrote, then the second read will succeed. Does putting a tiny sleep after p.sendeof() improve matters? I'd like to find a better solution than adding sleeps, but it should check if my guess about the problem is correct.

from ptyprocess.

Yes, that's indeed the case. The first read sometimes just gets exit and not the \r\n bit thereafter.

from ptyprocess.

What about calling p.readline() there instead of p.read()?

from ptyprocess.