subprocess call acts differently than command line call?

J

Jim

Hello,

I need a program that will traverse a directory tree to ensure that
there
are unix-style line endings on every file in that tree that is a text
file.
To tell text files from others I want to use the unix "file" command
(Python's "mimetypes" is not so good for me). But I am stuck on
something about getting that output, and I'd greatly appreciate any
pointers.

Both the command line "file" and the python libmagic binding give the
same behavior, but I'll illustrate with "file". It sometimes acts
differently when run from the command line than when run using
the subprocess module (with the same user). For example, it
sometimes
gives output when run from the command line but no output when run as
a subprocess.

Below is a short program to demo. (I use this on a test file tree
that is at ftp://joshua.smcvt.edu/pub/hefferon/a.zip if anyone
is interested.)

.............................................
import subprocess
import glob

for fn in glob.glob('a*/*'):
cmd=['/usr/bin/file',fn]
cmdStr=" ".join(cmd)
try:

p=subprocess.Popen(cmdStr,shell=True,stdin=subprocess.PIPE,stdout=subprocess.PIPE,stderr=subprocess.PIPE,close_fds=True)
(child_stdin,
child_stdout,
child_stderr)=(p.stdin,p.stdout,p.stderr)
stdoutContent=child_stdout.read()
except Exception, err:
mesg=u"unable to execute %s" % (repr(cmdStr),)
raise StandardError, mesg+": "+str(err)
if (p.returncode):
mesg=u"trouble executing %s" % (repr(cmdStr),)
raise StandardError, mesg+": "+repr(p.returncode)
print "result: ",stdoutContent
print "done"
.............................................

I've put a transcript of what happens at the bottom of this message.
One file (in the test tree it is "eqchange.txt") gives no output from
the above program, but does give an output when I use "file" at the
command line.

Specifying "-m/usr/share/file/magic" in the "file" call doesn't
change
that the command line and subprocess calls act differently, so it is
not just a question of different environments causing the system to
use different "magic" files. Changing the PIPE's to files, then
closing
and reopening them also does not matter, I believe.

In short I expected subprocess to just mimic my typing it in. Is
there some reason "file" doesn't act this way, and is there some
way to make it do so?

I have Python 2.4.4 running on Ubuntu. Thank you for any suggestions,
Jim

-----transcript (edited to shorten)------------
$ python test.py
result: acrotex/readme.txt: ASCII English text, with CRLF line
terminators

result: acrotex/eq2db.ins: ASCII English text, with CRLF line
terminators

result: acrotex/eqchange.txt:

result: acrotex/exerquiz.dtx: ISO-8859 English text, with CRLF line
terminators
result: acrotex/doc: directory

done

$ file acrotex/eqchange.txt
acrotex/eqchange.txt: ISO-8859 English text, with CRLF line
terminators
$ file acrotex/eqchange.txt 1> test.out
$ cat test.out
acrotex/eqchange.txt: ISO-8859 English text, with CRLF line
terminators
$ file acrotex/eqchange.txt 2> test.out
acrotex/eqchange.txt: ISO-8859 English text, with CRLF line
terminators
$ cat test.out
$
 
R

Rob Wolfe

Jim said:
Hello,

I need a program that will traverse a directory tree to ensure that
there
are unix-style line endings on every file in that tree that is a text
file.
To tell text files from others I want to use the unix "file" command
(Python's "mimetypes" is not so good for me). But I am stuck on
something about getting that output, and I'd greatly appreciate any
pointers.

Both the command line "file" and the python libmagic binding give the
same behavior, but I'll illustrate with "file". It sometimes acts
differently when run from the command line than when run using
the subprocess module (with the same user). For example, it
sometimes
gives output when run from the command line but no output when run as
a subprocess.

Below is a short program to demo. (I use this on a test file tree
that is at ftp://joshua.smcvt.edu/pub/hefferon/a.zip if anyone
is interested.)
[...]

I've put a transcript of what happens at the bottom of this message.
One file (in the test tree it is "eqchange.txt") gives no output from
the above program, but does give an output when I use "file" at the
command line.
[...]

-----transcript (edited to shorten)------------
$ python test.py
result: acrotex/readme.txt: ASCII English text, with CRLF line
terminators

result: acrotex/eq2db.ins: ASCII English text, with CRLF line
terminators

result: acrotex/eqchange.txt:

result: acrotex/exerquiz.dtx: ISO-8859 English text, with CRLF line
terminators
result: acrotex/doc: directory

done

$ file acrotex/eqchange.txt
acrotex/eqchange.txt: ISO-8859 English text, with CRLF line

That's interesting. I've checked this eqchange.txt file from
your a.zip on my machine and the result is:

$ file acrotex/readme.txt
acrotex/readme.txt: ASCII English text, with CRLF line terminators
$ file acrotex/eqchange.txt
acrotex/eqchange.txt:
$ file acrotex/exerquiz.dtx
acrotex/exerquiz.dtx: ISO-8859 English text, with CRLF line terminators

$ file -v
file-4.17
magic file from /etc/magic:/usr/share/file/magic
$ uname -orvm
2.6.18-4-k7 #1 SMP Wed May 9 23:42:01 UTC 2007 i686 GNU/Linux

That's really strange. Have you got only *one* version of ``file``
program on your machine?
 
J

Jim

That's really strange. Have you got only *one* version of ``file``
program on your machine?
Thank you, Rob. I thought I had that kind of thing covered in
specifying the -m but obviously there was something I wasn't getting,
and you found it.

Hope I can return the favor sometime,
Jim
 

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.

Ask a Question

Members online

No members online now.

Forum statistics

Threads
473,744
Messages
2,569,483
Members
44,901
Latest member
Noble71S45

Latest Threads

Top