Thomas Wouters [Tue, 23 Jan 2007 15:09:19 +0000 (15:09 +0000)]
Backport trunk revision 53527:
SF patch #1630975: Fix crash when replacing sys.stdout in sitecustomize
When running the interpreter in an environment that would cause it to set
stdout/stderr/stdin's encoding, having a sitecustomize that would replace
them with something other than PyFile objects would crash the interpreter.
Fix it by simply ignoring the encoding-setting for non-files.
This could do with a test, but I can think of no maintainable and portable
way to test this bug, short of adding a sitecustomize.py to the buildsystem
and have it always run with it (hmmm....)
Anthony Baxter [Tue, 10 Oct 2006 16:20:41 +0000 (16:20 +0000)]
Backport 50567
#1494314: Fix a regression with high-numbered sockets in 2.4.3. This
means that select() on sockets > FD_SETSIZE (typically 1024) work again.
The patch makes sockets use poll() internally where available.
Tim Peters [Mon, 9 Oct 2006 23:37:58 +0000 (23:37 +0000)]
Backport rev 51262 from trunk -- squashes a compiler warning on Windows
about truly wrong code.
Checkin comment from 51262:
Can't return NULL from a void function. If there is a memory error,
about the best we can do is call PyErr_WriteUnraisable and go on.
We won't be able to do the call below either, so verify delstr is valid.
Tim Peters [Mon, 9 Oct 2006 23:18:44 +0000 (23:18 +0000)]
Move fetching of encoding test files from the end of the
Windows builbot's "build" step to the start of its "test"
step.
This is poke-and-hope. The hope is that compilation failures
on Windows will become visible to the buildbot (bsddb has
apparently been failing to compile in 2.4 on Windows "for
some time" now, but the buildbots haven't noticed that).
Tim Peters [Mon, 9 Oct 2006 20:24:45 +0000 (20:24 +0000)]
Backport of the pieces of trunk rev 46589 relevant to
fixing an unlikely crash bug in dict resizing, SF
bug 1456209.
The rest of rev 46589 changes whether Python suppresses
exceptions during some dict-related comparisons. While I
think that's a good idea, it does change visible behavior at
times, and there was already some complaining about that on
the trunk. Not a good idea for backporting. The part of
46589 checked in here can at worst stop segfaults, and I doubt
anyone will gripe about that ;-)
Martin v. Löwis [Mon, 9 Oct 2006 19:29:06 +0000 (19:29 +0000)]
Backport r45505, r45573, r45576
- reset errno before calling confstr - use confstr() doc to simplify
checks afterwards
- Correct implementation and documentation of os.confstr. Add a simple
test case. I've yet to figure out how to provoke a None return I can test.
- Address issues brought up by MvL on python-checkins.
I tested this with valgrind on amd64.
The man pages I found for diff architectures are inconsistent on this.
I'm not entirely sure this change is correct for all architectures
either.
Perhaps we should just over-allocate and not worry about it?
The change to return None instead of "" in case of unconfigured
values has not been backported.
_ssl.c: under fail: self is DECREF'd, but it would have been NULL.
_csv.c: I'm not sure if lineterminator could have been anything other than
a string. However, other string method calls are checked, so check this
one too.
Fix three nits found by Coverity, adding null checks and comments.
[This commit only makes two changes. One change in the original patch
is just adding a comment, and another adds a 'base != NULL' check to
silence Coverity, but a comment adds that that base is never going to
be NULL. I didn't backport that change. --amk]
Ronald Oussoren [Sun, 8 Oct 2006 17:41:25 +0000 (17:41 +0000)]
Port of universal binary support for Mac OSX from python 2.5. This takes away
the need for the out-of-tree universal binary support that was used to build
the 2.4.3 installer.
Missing here relative to that tree are some changes to IDLE, IMHO those patches
aren't appropriate for the 2.4 branch and users are better of using 2.5's IDLE.
Prevent an invalid memory read from test_coding in case the done flag is set.
In that case, the loop isn't entered. I wonder if rather than setting
the done flag in the cases before the loop, if they should just exit early.
This code looks like it should be refactored.
Backport candidate (also the early break above if decoding_fgets fails)
[Backport r51222 | neal.norwitz -- if you hack the code to set r=NULL,
you find that Python does print "MemoryError". There's no traceback
and no indication of which line of code is responsible, but it's
better than a segfault.]
Handle NULL nodes while parsing. I'm not entirely sure this is correct.
There might be something else that needs to be done to setup the error.
[Backport r51221 | neal.norwitz -- the original commit message is wrong;
this code is only used if WITHOUT_COMPLEX is *not* defined, which is the
common case for Python builds.]
This code is actually not used unless WITHOUT_COMPLEX is defined.
However, there was no error checking that PyFloat_FromDouble returned
a valid pointer. I believe this change is correct as it seemed
to follow other code in the area.
v2 can be NULL if exception2 is NULL. I don't think that condition can happen,
but I'm not sure it can't either. Now the code will protect against either
being NULL.
[Backport r50783 | neal.norwitz. The bytes_left code is complicated,
but looks correct on a casual inspection and hasn't been modified
in the trunk. Does anyone want to review further?]
Ensure we don't write beyond errText. I think I got this right, but
it definitely could use some review to ensure I'm not off by one
and there's no possible overflow/wrap-around of bytes_left.
Reported by Klocwork #1.
Fix a problem if there is a failure allocating self->db.
Found with failmalloc.
[Partial backport of r51218 | neal.norwitz -- the changes to ast.c, symtable.c,
and _elementtree.c weren't applicable]
Klocwork made another run and found a bunch more problems.
This is the first batch of fixes that should be easy to verify based on context.
This fixes problem numbers: 220 (ast), 323-324 (symtable),
321-322 (structseq), 215 (array), 210 (hotshot), 182 (codecs), 209 (etree).
Cause a PyObject_Malloc() failure to trigger a MemoryError, and then
add 'if (PyErr_Occurred())' checks to various places so that NULL is
returned properly.
If _stat_float_times is false, we will try to INCREF ival which could be NULL.
Return early in that case. The caller checks for PyErr_Occurred so this
should be ok.
Try to handle a malloc failure. I'm not entirely sure this is correct.
There might be something else we need to do to handle the exception.
Klocwork # 212-213
[I think this needs more work; a malloc() failure will cause a match to fail,
but nothing raises MemoryError. I'll work on this on the trunk and backport.
--amk]
I'm not sure why this code allocates this string for the error message.
I think it would be better to always use snprintf and have the format
limit the size of the name appropriately (like %.200s).
Armin Rigo [Wed, 4 Oct 2006 10:13:32 +0000 (10:13 +0000)]
A review of overflow-detecting code in the 2.4 branch.
* unified the way intobject, longobject and mystrtoul handle
values around -sys.maxint-1.
* in general, trying to entierely avoid overflows in any computation
involving signed ints or longs is extremely involved. Fixed a few
simple cases where a compiler might be too clever (but that's all
guesswork).
* more overflow checks against bad data in marshal.c.
It's very unlikely, though possible that source is not a string. Verify
that PyString_AsString() returns a valid pointer. (The problem can
arise when zlib.decompress doesn't return a string.)
projects / thirdparty / Python / cpython.git / log
