[Python-Dev] Issue 13524: subprocess on Windows

[Terry Reedy]

On 12/4/2011 5:59 AM, Tim Golden wrote:
> http://bugs.python.org/issue13524
>
> Someone raised issue13524 yesterday to illustrate that a
> subprocess will crash immediately if an environment block is
> passed which does not contain a valid SystemRoot environment
> variable.
>
> Note that the calling (Python) process is unaffected; this
> isn't - strictly - a Python crash. The issue is essentially
> a Windows one where a fairly unusual cornercase -- passing
> an empty environment -- has unforseen effects.
>
> The smallest reproducible example is this:
>
> import os, sys
> import subprocess
> subprocess.Popen(
> [sys.executable],
> env={}
> )
>
> and it can be prevented like this:
>
> import os, sys
> import subprocess
> subprocess.Popen(
> [sys.executable],
> env={"SystemRoot" : os.environ['SystemRoot']}
> )
>
> There's a blog post here which gives a worked example:
>
>
> http://jpassing.com/2009/12/28/the-hidden-danger-of-forgetting-to-specify-systemroot-in-a-custom-environment-block/
>
>
> but as the author points out, nowhere on MSDN is there a warning
> that SystemRoot is mandatory. (And, in effect, it's not as it
> would just be possible to write code which had no need of it).
>
> So... what's our take on this? As I see it we could:
>
> 1) Do nothing: it's the caller's responsibility to understand the
> complications of the chosen Operating System.
>
> 2) Add a doc warning (ironically, considering the recent to-and-fro
> on doc warnings in this very module).
>
> 3) Add a check into the subprocess.Popen code which would raise some
> exception if the environment block is empty (or doesn't contain
> SystemRoot) on the grounds that this probably wasn't what the user
> thought they were doing.
>
> 4) Automatically add an entry for SystemRoot to the env block if it's
> not present already.
>
>
> It's tempting to opt for (1) and if we were exposing an API called
> CreateProcess which mimicked the underlying Windows API I would be
> inclined to go that way. But we're abstracting a little bit away
> from that and I think that that layer of abstraction carries its
> own responsibilities.
>
> Option (3) seems to give the best balance. It *is* a cornercase, but at
> the same time it's easy to misunderstand that the env block you're
> passing in *replaces* rather than *augments* that of the current
> process.
>
> Thoughts?

My inclination would be #4 on Windows, certainly for 3.3, unless there 
is a clear reason not to.

For 2.7/3.2, at least say (not warn, just say) in the doc that that a 
subprocess on Windows may require that SystemRoot be set.

The blog post says the problem is worse on Win 7. So it is not going away.

The blog post has a comment from Martin Loewis a year ago linking to
http://mail.python.org/pipermail/python-dev/2010-November/105866.html
That thread refers to a bug that was not posted on the tracker. This 
makes at least three (including #3440).

-- 
Terry Jan Reedy