Geek asks and answers

Reading stderr subprocess at runtime

I would like to read what is written in stderr by a subprocess during its execution.

However, when I use this script that I wrote, stderr doesn't seem to have anything for me to read until the subprocess finishes.

#!/usr/bin/env python2

import sys
from subprocess import Popen, PIPE, STDOUT

if len(sys.argv) < 2:
    print "Please provide a command"
    sys.exit(1)

sub = Popen(sys.argv[1:], stdout=PIPE, stderr=STDOUT)

for i, line in enumerate(sub.stdout):
    sys.stdout.write("%d: %s" % (i, line))

change

OK, now I’ve got closer. If I specify the number of bytes read, it will overcome buffering.

#!/usr/bin/env python2

import sys 
from subprocess import Popen, PIPE, STDOUT

if len(sys.argv) < 2:
    print "Please provide a command"
    sys.exit(1)

sub = Popen(sys.argv[1:], stdout=PIPE, stderr=STDOUT)

i = 0 
while sub.poll() is None:
    line = sub.stdout.read(64)
    line.strip("\b")
    sys.stdout.write("%d: %s\n" % (i, line))
    i += 1

Output snippet:

58:  86 q=21.0 size= 4541841kB time=00:00:22.08 bitrate=1685014.2kbi
frame=  567 fps= 86 q=22.0 size= 4543667kB time=00:00:2
frame=  621 fps= 87 q=20.0 sizs/s    
frame= 4545352kB time=00:00:26.11 bitrate=1425939.2kbits/s    
62: =  686 fps= 90 q=12.0 size= 4546970kB time=00:00:28.89 bitrate=1
frame=  758 fps= 93 q=25.0 size= 4548534kB t
frame=  794 fps= 92 bitrate=1168185.5kbits/s    
65:  q=27.0 size= 4550901kB time=00:00:33.40 bitrate=1115897.0kbits/
frame=  827 fps= 91 q=27.0 size= 4552324kB time=00:00:34.7
frame=  857 fps= 89 q=26.0 size=  
frame=  254kB time=00:00:36.12 bitrate=1032874.9kbits/s    
69: 892 fps= 88 q=25.0 size= 4556598kB time=00:00:37.36 bitrate=9988
frame=  948 fps= 89 q=19.0 size= 4558565kB time=
frame= 1006 fps= 90 q=19937320.4kbits/s    
72: .0 size= 4560139kB time=00:00:42.16 bitrate=885880.0kbits/s    
73: frame= 1060 fps= 91 q=19.0 size= 4561958kB time=00:00:44.49 bitr
frame= 1122 fps= 93 q=18.0 size= 4563460
frame= 1173 fps=0:47.08 bitrate=793898.4kbits/s

My problem now seems to be that ffmpeg uses backspace characters or clutter-like ones with stdout. Not sure what is going on here.

+5
source share
3 answers

sh . , , , . docs.

sh, communicate Popen

+3

PIPE , communicate():

sub = Popen(sys.argv[1:], stdout=PIPE, stderr=PIPE)
output, error_output = sub.communicate()

print 'OUTPUT:'
print output

print 'ERROR:'
print error_output
+1

If the subprocess does not use block buffering , you can try:

for i, line in enumerate(iter(sub.stdout.readline, b"")):
    print i, line,
+1
source

More articles:


All Articles