I run the following script:
with open("logfile.txt", 'w') as log_file:
cmd = path + '/somebinary'
log_file.write("Running the command: %s\n" % cmd)
subprocess.call(cmd, shell=True, stdout=log_file, stderr=log_file)
However, the cmd variable gets written to the end of the file and not to the beginning (which I was expecting / hoping for). Can somebody explain to me why and how to prevent this, please?
The operating system performs buffering which can cause output to occur in unexpected order. To force it, flush the file handle after writing.
with open("logfile.txt", 'w') as log_file:
cmd = path + '/somebinary'
log_file.write("Running the command: %s\n" % cmd)
log_file.flush()
subprocess.call(cmd, shell=True, stdout=log_file, stderr=log_file)
Demo: https://ideone.com/U8RPBy
As an aside, you generally want to avoid shell=True; try
cmd = [path + '/somebinary']
...
subprocess.call(cmd, stdout=log_file, stderr=log_file)
and perhaps make sure your PATH is correct instead of hardcoding a specific path.
shell=True. I will have to investigate. And I am not sure why you suggest to put the path into a list.
shell=True explains they why and the how of that. If you need a shell for some reason (like somebinary is not actually a binary at all) then of course that refactoring is not possible.