Changes to sre.c after the application of patch #726869 have increased
stack usage on FreeBSD, requiring the recursion limit to be lowered
further. Building with gcc 2.95 (the standard compiler on FreeBSD 4.x)
is now also affected.
The underlying issue is that FreeBSD's pthreads implementation has a
hard-coded 1MB stack size for the initial (or "primary") thread, which
can not be changed without rebuilding libc_r. Exhausting this stack
results in a bus error.
Building without pthreads (configure --without-threads), or linking
with the port of the Linux pthreads library (aka Linuxthreads) instead
of libc_r, avoids this limitation.
On OS/2, only gcc 3.2 is affected and the stack size is controllable,
so the special handling has been removed.
reorganise modules, so that most standard extensions are built into
the Python DLL. this slightly improves performance; reduces
fragmentation of address spaces and slightly reduces memory footprint.
reorganise modules, so that most standard extensions are built into
the Python DLL. this slightly improves performance; reduces
fragmentation of address spaces and slightly reduces memory footprint.
- urllib2.py now knows how to order proxy classes, so the user doesn't
have to insert it in front of other classes, nor do dirty tricks like
inserting a "dummy" HTTPHandler after a ProxyHandler when building an
opener with proxy support.
Kurt B. Kaiser [Fri, 6 Jun 2003 21:58:38 +0000 (21:58 +0000)]
1. Update WindowList registry when filename changes so that
Window menu updates.
2. Display Python Shell window in Window menu
3. Remove some dead code in FileList.py
Jack Jansen [Sun, 1 Jun 2003 20:57:12 +0000 (20:57 +0000)]
Added two keys to database format: User-install-skips is an array of pathname
prefixes, any file that is skipped during a per-user install that matches
this set is *not* an error; Systemwide-only is a boolean that says the
package cannot be installer per-user.
Jack Jansen [Sun, 1 Jun 2003 20:03:43 +0000 (20:03 +0000)]
Use splitlines() in stead of split() to split lines, and added a method
shortdescription() so the code to split off the first line of the
description isn't all over the place.
Kurt B. Kaiser [Sun, 1 Jun 2003 01:08:32 +0000 (01:08 +0000)]
Set the default for delete-exitfunc in config-main.def to True to handle
abnormal exit situations cleanly, especially stuck user threads. Future
plan is to intercept the user's atexit functions and run them under
IDLE's control.
Greg Ward [Thu, 29 May 2003 21:53:06 +0000 (21:53 +0000)]
Completely change the interface to setparameters():
* it no longer takes ssize, which served no purpose apart from
scolding you if you got it wrong
* changed the order of the three remaining required arguments
to (format, channels, rate) to match the order in which they
must be set
* replaced the optional argument 'emulate' with 'strict': if strict
true, and the audio device does not accept the requested sampling
parameters, raise OSSAudioError
* return a tuple (format, channels, rate) reflecting the sampling
parameters that were actually set
Change the canonical name of ossaudiodev.error to
ossaudiodev.OSSAudioError (keep an alias for backwards compatibility).
Remove 'audio_types' list and 'n_audio_types' (no longer needed now that
setparameters() no longer has an 'ssize' argument to police).
Barry Warsaw [Thu, 29 May 2003 19:39:33 +0000 (19:39 +0000)]
_make_boundary(): Fix for SF bug #745478, broken boundary calculation
in some locales. This code simplifies the boundary algorithm to use
randint() which is what we wanted anyway.
Guido van Rossum [Thu, 29 May 2003 14:36:57 +0000 (14:36 +0000)]
I saw errors from _fileobject.__del__ about missing self._sock. This
can happen if __init__ doesn't complete. Fix it by adding a
try/except to __del__.
Greg Ward [Thu, 29 May 2003 01:39:32 +0000 (01:39 +0000)]
Rewrite the description of setparameters() so it actually reflects the
code (which has also been recently overhauled, so that it can be
documented without embarassment).
Jack Jansen [Wed, 28 May 2003 18:56:30 +0000 (18:56 +0000)]
- Added a scheme whereby pimp can update itself, by importing a module
pimp_update if it exists. Upped the version number to indicate this.
Fixes #731626.
- Added -V (print version) and -u (specify database URL) options when run
as a command line tool.
Gregory P. Smith [Wed, 28 May 2003 16:20:03 +0000 (16:20 +0000)]
Remove the note about the BerkeleyDB license. Barry indicates that
sleepycat has told him that since python is OSS and berkeleydb is
being distributed with python for the bsddb library all is fine.
Gregory P. Smith [Wed, 28 May 2003 07:56:45 +0000 (07:56 +0000)]
Include a link to the pybsddb web documentation for the modern object
oriented DbEnv & Db object BerkeleyDB interface.
Include a note about Sleepycat's BerkeleyDB license with regards to
distributing it within non-opensource applications and include a link
to their website for license details.
Document that 'r' is the default flag for the legacy bsddb btopen(),
hashopen(), and rnopen() functions. This is apparently different than the
dbm libraries in other languages according to Bug #732951. Changing
the default from 'r' to 'c' would break backwards compatibility with
legacy bsddb applications; documenting the default should suffice.
Greg Ward [Tue, 27 May 2003 01:57:21 +0000 (01:57 +0000)]
Oops, move the GIL release/reacquire from oss_sync() to _do_ioctl_0():
that way it applies to *only* the ioctl() call, and also happens for the
other blocking ioctls (POST, RESET).
Greg Ward [Mon, 26 May 2003 22:47:30 +0000 (22:47 +0000)]
Release the GIL in two more methods:
* sync(), because it waits for hardware buffers to flush, which
can take several seconds depending on cirumstances (according
to the OSS docs)
* close(), because it does an implicit sync()