Path: csiph.com!v102.xanadu-bbs.net!xanadu-bbs.net!feeder.erje.net!eu.feeder.erje.net!newsfeed.xs4all.nl!newsfeed4.news.xs4all.nl!xs4all!newsgate.cistron.nl!newsgate.news.xs4all.nl!post.news.xs4all.nl!not-for-mail
MIME-Version: 1.0
In-Reply-To: <53c66ba8$0$9505$c3e8da3$5496439d@news.astraweb.com>
References: <mailman.10416.1401304999.18130.python-list@python.org> <57ajo9poljjre4c4ig0n0ss8kph8k78lp0@4ax.com> <5389cb53$0$29978$c3e8da3$5496439d@news.astraweb.com> <99b7b2a2-7521-42d7-a5a0-1a35d4d5b922@googlegroups.com> <53C4A454.9010600@gmail.com> <mailman.11826.1405398039.18130.python-list@python.org> <lq365f$4c9$1@speranza.aioe.org> <mailman.11836.1405429260.18130.python-list@python.org> <lq3bvu$ob4$1@dont-email.me> <mailman.11838.1405434699.18130.python-list@python.org> <87zjga4j4v.fsf@elektro.pacujo.net> <53c57bae$0$9505$c3e8da3$5496439d@news.astraweb.com> <87iomy4ciy.fsf@elektro.pacujo.net> <53c5f6dc$0$9505$c3e8da3$5496439d@news.astraweb.com> <87egxl4zq8.fsf@elektro.pacujo.net> <53c62e7f$0$29897$c3e8da3$5496439d@news.astraweb.com> <871ttlfune.fsf@elektro.pacujo.net> <53c66ba8$0$9505$c3e8da3$5496439d@news.astraweb.com>
Date: Wed, 16 Jul 2014 22:55:54 +1000
Subject: Re: Python 3 is killing Python
From: Chris Angelico <rosuav@gmail.com>
Cc: "python-list@python.org" <python-list@python.org>
Content-Type: text/plain; charset=UTF-8
Precedence: list
Newsgroups: comp.lang.python
Message-ID: <mailman.11876.1405515780.18130.python-list@python.org>
Lines: 65
NNTP-Posting-Host: 2001:888:2000:d::a6
Xref: csiph.com comp.lang.python:74550

On Wed, Jul 16, 2014 at 10:10 PM, Steven D'Aprano
<steve+comp.lang.python@pearwood.info> wrote:
> Linux, like all Unixes, is primarily a text-based platform. With a few
> exceptions, /etc is filled with text files, not binary files, and half
> the executables on the system are text (Python, Perl, bash, sh, awk,
> etc.).

An interesting assertion. I know "half" is not meant to be an actual
estimate, but out of curiosity, I whipped up a quick script to figure
out just how many of my executables are text and how many aren't.

#!/usr/bin/env python3
import os, subprocess
text = binary = unknown = unreadable = 0
for path in os.environ["PATH"].split(":"):
    for file in os.listdir(path):
        fn = os.path.join(path, file)
        try:
            t = subprocess.check_output(["file", "-L", fn])
        except subprocess.CalledProcessError:
            print("Unreadable: %s" % fn)
            unreadable += 1
            continue
        if isinstance(t, bytes): t = t.decode("ascii")
        # Now to try to figure out what's text and what's binary.
        if "text" in t:
            # Most Unixes follow the convention of having "text" in
            # the output of all files that can be safely blatted to
            # a terminal - for instance, "ASCII text executable" is
            # used to describe most shell scripts etc; this file is
            # a "Python script, ASCII text executable". If I put in
            # a non-ASCII char, the 'file' descr becomes changes to
            # "Python script, UTF-8 Unicode text executable".
            text += 1
        elif "directory" in t:
            # Ignore directories.
            pass
        elif "LSB executable" in t or "LSB shared object" in t:
            binary += 1
        else:
            print(t.strip())
            unknown += 1
print("%d text, %d binary" % (text, binary))
if unknown: print("Also %d unknowns, which are probably binary." % unknown)
if unreadable: print("Plus %d files that couldn't be read." % unreadable)


On my system, it says:
rosuav@sikorsky:~$ python3 exectypes.py
/usr/local/bin/youtube-dl: data
Unreadable: /usr/bin/wine-safe
/usr/bin/mptopdf: LaTeX auxiliary file,
/usr/bin/gvfs-less: Palm OS dynamic library data "#!/bin/sh"
Unreadable: /usr/bin/gserialver
1140 text, 2060 binary
Also 3 unknowns, which are probably binary.
Plus 2 files that couldn't be read.

So a bit more than a third of my executables are text. That's a pretty
high proportion, and not very far off the rough guesstimate of half.
(And I tried this on three other Linuxes I have around the house,
getting broadly the same proportion, although the numbers are quite
different.)

ChrisA