Python 2.1b1 - pydoc shows nothing

[pfmoore]

BPO	406280
Nosy	@gvanrossum, @tim-one, @pfmoore

^{Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.}

Show more details

GitHub fields:

assignee = None
closed_at = <Date 2001-04-13.12:48:49.000>
created_at = <Date 2001-03-06.13:06:24.000>
labels = ['library']
title = 'Python 2.1b1 - pydoc shows nothing'
updated_at = <Date 2001-04-13.12:48:49.000>
user = 'https://github.com/pfmoore'

bugs.python.org fields:

activity = <Date 2001-04-13.12:48:49.000>
actor = 'ping'
assignee = 'ping'
closed = True
closed_date = None
closer = None
components = ['Library (Lib)']
creation = <Date 2001-03-06.13:06:24.000>
creator = 'paul.moore'
dependencies = []
files = []
hgrepos = []
issue_num = 406280
keywords = []
message_count = 5.0
messages = ['3716', '3717', '3718', '3719', '3720']
nosy_count = 4.0
nosy_names = ['gvanrossum', 'tim.peters', 'ping', 'paul.moore']
pr_nums = []
priority = 'low'
resolution = 'fixed'
stage = None
status = 'closed'
superseder = None
type = None
url = 'https://bugs.python.org/issue406280'
versions = []

[pfmoore]

Platform: Windows 2000, Python 2.1b1

The pydoc script works fine in "serve documents to a
browser" mode (python pydoc). Also, running it as a
command line application, as "python pydoc pydoc",
works fine when the environment variable PAGER is
unset.

However, when I have PAGER=less, I get no output at
all.

It looks like a bug in pydoc.pipepager(), which is the
result of a bug in os.popen(). I can work around the
bug by using pydoc.tempfilepager() in place of
pydoc.pipepager(), but I don't know what the
underlying popen() bug is.

To demonstrate the os.popen() bug, see the attached
interactive session:

C:\Data>python21
Python 2.1b1 (#11, Mar  2 2001, 11:23:29) [MSC 32 bit 
(Intel)] on win32
Type "copyright", "credits" or "license" for more 
information.
>>> import os
>>> a = os.popen("more", "w")
>>> a.write("Hello")
>>> a.close()

Run this, and note that the "More" program never
starts...

[gvanrossum]

Logged In: YES
user_id=6380

Assigned to Tim, because of that fine combination of
keywords:
popen and Windows.

[tim-one]

Logged In: YES
user_id=31435

Assigned to MarkH for popen insight.

"more" does start, but the output just vanishes under
Win98SE (and, I assume, W2K too). This I deduced via doing
os.popen("more > somefile.txt", "w") instead; the file is
created, and does contain the stuff written to the handle.
Besides, the std test_popen2 test uses "more" under Windows
too, and works fine.

Nothing unique about "more" here: tried a number of .exe
files, and it's all the same: the stdout of the popen'ed
program isn't displayed. I assume this is because the
original console's stdout doesn't manage to become the
popen'ed stdout, but I don't understand it in detail. Deep
or shallow?

In any case, trying to implement a pager via popen like
this goes beyond what C guarantees, so I reduced the
priority accordingly.

[tim-one]

Logged In: YES
user_id=31435

Reassigned to Ping since wrestling w/ popen() on Windows is
hopeless. Left the priority low because it's gotta be rare
for anyone to define a PAGER envar under Windows. Note
that Paul mailed some kind of related patch to Python-Dev
recently ... mumble, mumble, ... here:

http://mail.python.org/pipermail/python-dev/2001-
April/014070.html

[ping]

Logged In: YES
user_id=45338

Fixed by always using temp files in Windows.
New version is checked in.