]> granicus.if.org Git - python/log
python
24 years agoIntroduction to rich comparisons:
Guido van Rossum [Wed, 17 Jan 2001 15:20:39 +0000 (15:20 +0000)]
Introduction to rich comparisons:

- Removed the nb_add slot from the PyNumberMethods struct.

- Renamed Py_TPFLAGS_NEWSTYLENUMBER to Py_TPFLAGS_CHECKTYPES.

- Added typedef richcmpfunc.

- Added tp_richcompare slot to PyTypeObject (replacing spare tp_xxx7).

- Added APIs PyObject_RichCompare() and PyObject_RichCompareBool().

- Added rich comparison operators Py_LT through Py_GE.

24 years agoPatch #103279: sysconfig.py always looks for versions of files in
Andrew M. Kuchling [Wed, 17 Jan 2001 15:16:52 +0000 (15:16 +0000)]
Patch #103279: sysconfig.py always looks for versions of files in
sys.prefix + 'config/Makefile'. When building Python for the first
time, these files aren't there, so the files from the build tree have
to be used instead; this file adds an entry point for specifying that
the build tree files should be used.  (Perhaps 'set_python_build' should
should be preceded with an underscore?)

24 years agoFix a bizarre typo in the helper class ComparableException: the
Guido van Rossum [Wed, 17 Jan 2001 15:08:37 +0000 (15:08 +0000)]
Fix a bizarre typo in the helper class ComparableException: the
__getattr__() method, which clearly (like the other methods) was
intended to pass the __getattr__() call on to the self.err object,
mistakenly returned getattr(self, self.err) rather than
getattr(self.err, attr).  Since self.err is not a string, this always
raises a TypeError.  Apparently that doesn't bother for the one
attribute for which __getattr__() is actually called ('__coerce__'),
but it broke the rich comparisons stuff that I'm trying to get into
shape, so I'm fixing this now.  (I could also simply remove the
__getattr__() method, but fixing it seems more in the spirit of what
the ComparableException class is trying to do.)

24 years agoChanged name of codec to full path name. This allows importing
Marc-André Lemburg [Wed, 17 Jan 2001 15:07:00 +0000 (15:07 +0000)]
Changed name of codec to full path name. This allows importing
the test_charmapcodec test via the test package.

24 years agoBump version to 2.1a1. (To be released Friday.)
Guido van Rossum [Wed, 17 Jan 2001 14:12:33 +0000 (14:12 +0000)]
Bump version to 2.1a1.  (To be released Friday.)

24 years agoFixed typo: Using --with-dbm had no effect because due to a typo it
Sjoerd Mullender [Wed, 17 Jan 2001 09:42:21 +0000 (09:42 +0000)]
Fixed typo: Using --with-dbm had no effect because due to a typo it
was never enabled.

24 years agoHand repair of cases where reindent changed lines of the form
Tim Peters [Wed, 17 Jan 2001 09:13:33 +0000 (09:13 +0000)]
Hand repair of cases where reindent changed lines of the form

\t\t\t\t\treal code
##\t\t\t\t\tunused code
\t\t\t\t\treal code

via untabifying and shifting the real code left.  Semantically the
same but made the intent of the commented-out-in-column-0 unused code
unclear.  The exact same unused code appears to have gotten copied from
file to file over the years.

24 years agoWindows installer: update dialogs, program groups, etc, to 2.1 alpha 1.
Tim Peters [Wed, 17 Jan 2001 08:54:19 +0000 (08:54 +0000)]
Windows installer:  update dialogs, program groups, etc, to 2.1 alpha 1.

24 years agoWhitespace normalization.
Tim Peters [Wed, 17 Jan 2001 08:48:39 +0000 (08:48 +0000)]
Whitespace normalization.

24 years agoEric the half-a-wit, driven to berserk rage after repeatedly doing
Eric S. Raymond [Wed, 17 Jan 2001 08:25:11 +0000 (08:25 +0000)]
Eric the half-a-wit, driven to berserk rage after repeatedly doing
builds during which he forgot to uncomment crucial library lines in
Setup, walks into Guido's East End nightclub with a tactical nuclear
weapon on his shoulder.  Said nuclear weapon is promptly deployed
exactly where it will do the most good, right in the middle of
configure.in.

With this patch, the set of libraries autoconfigured in is extended to
include ndbm, gdbm, and crypt.  This essentially eliminates any need to
tweak Setup for a normal Linux build.

"'E was a fair man.  Cruel, but fair."

24 years agoAdd a missing newline in an example; caught by Chris Ryland
Fred Drake [Wed, 17 Jan 2001 05:12:13 +0000 (05:12 +0000)]
Add a missing newline in an example; caught by Chris Ryland
<cpr@emsoftware.com>.

24 years agoUse __name__ instead of "test_regex" as the module name in the
Guido van Rossum [Wed, 17 Jan 2001 03:12:01 +0000 (03:12 +0000)]
Use __name__ instead of "test_regex" as the module name in the
warnings.filterwarnings() call.  This suppresses the warning when the
module is imported with its full name (test.test_regex) too.

24 years agoDocument xreadlines() method. (Forgot to check this in before!)
Guido van Rossum [Wed, 17 Jan 2001 01:18:00 +0000 (01:18 +0000)]
Document xreadlines() method.  (Forgot to check this in before!)

24 years agoRationalizing the fallback code for portable fseek -- this is all much
Guido van Rossum [Tue, 16 Jan 2001 20:53:31 +0000 (20:53 +0000)]
Rationalizing the fallback code for portable fseek -- this is all much
simpler if we use fgetpos and fsetpos, rather than trying to mess with
platform-specific TELL64 alternatives.

Of course, this hasn't been tested on a 64-bit platform, so I may have
to withdraw this -- but I'm hopeful, and Trent Mick supports this
patch!

24 years agoFix a few small markup/consistency nits.
Fred Drake [Tue, 16 Jan 2001 20:52:41 +0000 (20:52 +0000)]
Fix a few small markup/consistency nits.

24 years agoFix bugs with integer-valued variables when parsing Makefiles. Values
Andrew M. Kuchling [Tue, 16 Jan 2001 16:33:28 +0000 (16:33 +0000)]
Fix bugs with integer-valued variables when parsing Makefiles.  Values
for done[n] can be integers as well as strings, but the code
concatenates them with strings (fixed by adding a str()) and calls
string.strip() on them (fixed by rearranging the logic)

(Presumably this wasn't noticed previously because parse_makefile()
was only called on Modules/Makefile, which contains no integer-valued
variables.)

24 years agoModified version of a patch from Jeremy Kloth, to make .get_outputs()
Andrew M. Kuchling [Tue, 16 Jan 2001 16:16:03 +0000 (16:16 +0000)]
Modified version of a patch from Jeremy Kloth, to make .get_outputs()
produce a list of unique filenames:
    "While attempting to build an RPM using distutils on Python 2.0,
    rpm complained about duplicate files.  The following patch fixed
    that problem.

24 years agoTrigger keep-console-open on GUSISIOUX_STATE_UNKNOWN. Better than the previous compli...
Jack Jansen [Tue, 16 Jan 2001 15:54:58 +0000 (15:54 +0000)]
Trigger keep-console-open on GUSISIOUX_STATE_UNKNOWN. Better than the previous complicated expression.

24 years agoConfig file for standalone carbon python.
Jack Jansen [Tue, 16 Jan 2001 15:51:24 +0000 (15:51 +0000)]
Config file for standalone carbon python.

24 years agoAdded Carbon interpreter.
Jack Jansen [Tue, 16 Jan 2001 15:50:48 +0000 (15:50 +0000)]
Added Carbon interpreter.

24 years agoCommit version of config.h.in that covers the TERMIOS test.
Eric S. Raymond [Tue, 16 Jan 2001 15:32:53 +0000 (15:32 +0000)]
Commit version of config.h.in that covers the TERMIOS test.

24 years agoCheck in a version with the termios test in place.
Eric S. Raymond [Tue, 16 Jan 2001 15:26:34 +0000 (15:26 +0000)]
Check in a version with the termios test in place.

24 years agoMake pop_source and push_source available, as documented.
Eric S. Raymond [Tue, 16 Jan 2001 15:19:13 +0000 (15:19 +0000)]
Make pop_source and push_source available, as documented.

24 years agoMake HAVE_TERMIOS_H and associated config.sh shell variable available
Eric S. Raymond [Tue, 16 Jan 2001 15:01:26 +0000 (15:01 +0000)]
Make HAVE_TERMIOS_H and associated config.sh shell variable available
when configure detects the presence of termios.h; later we'll use this
for correct configuration of edline/readline.

Also, fix a bug in acconfig.h -- somebody forgot to add an undef to
cover the LIBNDBM configure symbol, which was preventing autoheader
from working properly.

24 years agoConverted to CW Pro 6 and new naming scheme.
Jack Jansen [Tue, 16 Jan 2001 14:24:56 +0000 (14:24 +0000)]
Converted to CW Pro 6 and new naming scheme.

24 years agopush_source() and pop_source() entry points for shlex instances.
Eric S. Raymond [Tue, 16 Jan 2001 14:18:55 +0000 (14:18 +0000)]
push_source() and pop_source() entry points for shlex instances.
These basically just make available to the user what userhook()
does to the source stack.  Documentation included.

24 years agoAdded the 'carb' resource to the carbon interpreter.
Jack Jansen [Tue, 16 Jan 2001 13:01:48 +0000 (13:01 +0000)]
Added the 'carb' resource to the carbon interpreter.

24 years agoReplaced the single .exp file by two files, one for classic and one for carbon.
Jack Jansen [Tue, 16 Jan 2001 13:01:11 +0000 (13:01 +0000)]
Replaced the single .exp file by two files, one for classic and one for carbon.

24 years agoAdded checks to prevent PyUnicode_Count() from dumping core
Marc-André Lemburg [Tue, 16 Jan 2001 11:54:12 +0000 (11:54 +0000)]
Added checks to prevent PyUnicode_Count() from dumping core
in case the parameters are out of bounds and fixes error handling
for .count(), .startswith() and .endswith() for the case of
mixed string/Unicode objects.

This patch adds Python style index semantics to PyUnicode_Count()
indices (including the special handling of negative indices).

The patch is an extended version of patch #103249 submitted
by Michael Hudson (mwh) on SF. It also includes new test cases.

24 years agobumped SRE version number to 2.1. cleaned up and added 1.5.2
Fredrik Lundh [Tue, 16 Jan 2001 07:37:30 +0000 (07:37 +0000)]
bumped SRE version number to 2.1.  cleaned up and added 1.5.2
compatibility patches.

24 years agoWhitespace normalization.
Tim Peters [Tue, 16 Jan 2001 07:12:46 +0000 (07:12 +0000)]
Whitespace normalization.

24 years agodoctest-- The Little Module That Could --finally makes it to the Big Show <wink>.
Tim Peters [Tue, 16 Jan 2001 07:10:57 +0000 (07:10 +0000)]
doctest-- The Little Module That Could --finally makes it to the Big Show <wink>.

24 years agofixed a memory leak in pattern cleanup (patch #103248 by cgw)
Fredrik Lundh [Tue, 16 Jan 2001 07:05:29 +0000 (07:05 +0000)]
fixed a memory leak in pattern cleanup (patch #103248 by cgw)

24 years agoFix for SF bug #123625: some newsservers need 'authinfo' *before* 'mode
Thomas Wouters [Tue, 16 Jan 2001 06:35:14 +0000 (06:35 +0000)]
Fix for SF bug #123625: some newsservers need 'authinfo' *before* 'mode
readers', others *after*. (Netscape Collabra for the first category,
INN-which-forks-nnrpd for the second.)

24 years agoPlug memory leak.
Neil Schemenauer [Tue, 16 Jan 2001 04:27:47 +0000 (04:27 +0000)]
Plug memory leak.

24 years agoAdd strip_dir argument to the single call to .object_filenames(), to
Andrew M. Kuchling [Tue, 16 Jan 2001 03:10:43 +0000 (03:10 +0000)]
Add strip_dir argument to the single call to .object_filenames(), to
    prevent creating files such as build/lib.whatever/Modules/foo.o
    when given a source filename such as Modules/foo.c.

24 years agoVariant of Skip's patch 103246 (Remove unneeded string exception compat from Queue).
Tim Peters [Mon, 15 Jan 2001 22:53:46 +0000 (22:53 +0000)]
Variant of Skip's patch 103246 (Remove unneeded string exception compat from Queue).

24 years agoUgh. Sorry. Checked in the wrong file. Please ignore revision 1.3;
Ka-Ping Yee [Mon, 15 Jan 2001 22:27:06 +0000 (22:27 +0000)]
Ugh.  Sorry.  Checked in the wrong file.  Please ignore revision 1.3;
it anticipates another patch i was about to propose.

24 years agobetter format names and error messages
Ka-Ping Yee [Mon, 15 Jan 2001 22:21:39 +0000 (22:21 +0000)]
better format names and error messages

24 years agoThis patch makes sure that the function name always appears in the error
Ka-Ping Yee [Mon, 15 Jan 2001 22:14:16 +0000 (22:14 +0000)]
This patch makes sure that the function name always appears in the error
message, and tries to make the messages more consistent and helpful when
the wrong number of arguments or duplicate keyword arguments are supplied.
Comes with more tests for test_extcall.py and and an update to an error
message in test/output/test_pyexpat.

24 years agoAdd tokenizer support and tests for u'', U"", uR'', Ur"", etc.
Ka-Ping Yee [Mon, 15 Jan 2001 22:04:30 +0000 (22:04 +0000)]
Add tokenizer support and tests for u'', U"", uR'', Ur"", etc.

24 years agoAdd a test case suggested by Guido, where a method is created with the
Barry Warsaw [Mon, 15 Jan 2001 21:00:02 +0000 (21:00 +0000)]
Add a test case suggested by Guido, where a method is created with the
new module.

24 years agoeffbot caught a typo!
Barry Warsaw [Mon, 15 Jan 2001 20:51:40 +0000 (20:51 +0000)]
effbot caught a typo!

24 years agoAdd a NEWS item about function attributes.
Barry Warsaw [Mon, 15 Jan 2001 20:43:18 +0000 (20:43 +0000)]
Add a NEWS item about function attributes.

24 years agoCommitting PEP 232, function attribute feature, approved by Guido.
Barry Warsaw [Mon, 15 Jan 2001 20:40:19 +0000 (20:40 +0000)]
Committing PEP 232, function attribute feature, approved by Guido.
Closes SF patch #103123.

funcobject.h:

    PyFunctionObject: add the func_dict slot.

funcobject.c:

    PyFunction_New(): Initialize the func_dict slot to NULL.

    func_getattr(): Rename to func_getattro() and change the
    signature.  It's more efficient to use attro methods and dig the C
    string out than it is to re-convert a C string to a PyString.

    Also, add support for getting the __dict__ (a.k.a. func_dict)
    attribute, and for getting an arbitrary function attribute.

    func_setattr(): Rename to func_setattro() and change the signature
    for the same reason.  Also add support for setting __dict__
    (a.k.a. func_dict) and any arbitrary function attribute.

    func_dealloc(): Be sure to DECREF the func_dict slot.

    func_traverse(): Be sure to traverse func_dict too.

    PyFunction_Type: make the necessary func_?etattro() changes.

classobject.c:

    instancemethod_memberlist: Add __dict__

    instancemethod_setattro(): New method to set arbitrary attributes
    on methods (really the underlying im_func).  Raise TypeError when
    the instance is bound or when you're trying to set one of the
    reserved im_* attributes.

    instancemethod_getattr(): Renamed to instancemethod_getattro()
    since that's what it really is.  Also, added support fo getting
    arbitrary attributes through the im_func.

    PyMethod_Type: Do the ?etattr{,o} dance.

24 years agoCommitting PEP 232, function attribute feature, approved by Guido.
Barry Warsaw [Mon, 15 Jan 2001 20:30:15 +0000 (20:30 +0000)]
Committing PEP 232, function attribute feature, approved by Guido.
Closes SF patch #103123.

Regression test for function attributes, with output file.

24 years agoDocument function attributes for both the function type and the method
Barry Warsaw [Mon, 15 Jan 2001 20:28:50 +0000 (20:28 +0000)]
Document function attributes for both the function type and the method
type.  The method documentation also includes a new brief discussion
of `bound' vs. `unbound' and why setting an attr on a bound method is
a TypeError.  Includes Skip's suggested text.

24 years agoAdd note about new and improved xrange().
Guido van Rossum [Mon, 15 Jan 2001 19:11:10 +0000 (19:11 +0000)]
Add note about new and improved xrange().

24 years agoSF patch #103158 by Greg Ball: Don't do unsafe arithmetic in xrange
Guido van Rossum [Mon, 15 Jan 2001 18:58:56 +0000 (18:58 +0000)]
SF patch #103158 by Greg Ball: Don't do unsafe arithmetic in xrange
object.

This fixes potential overflows in xrange()'s internal calculations on
64-bit platforms.  The fix is complicated because the sq_length slot
function can only return an int; we want to support
xrange(sys.maxint), which is a 64-bit quantity on most 64-bit
platforms (except Win64).  The solution is hacky but the best
possible: when the range is that long, we can use it in a for loop but
we can't ask for its length (nor can we actually iterate beyond
2**31-1, because the sq_item slot function has the same restrictions
on its arguments.  Fixing those restrictions is a project for another
day...

24 years ago- Make sure to quote the username and password (SF patch #103236 by
Guido van Rossum [Mon, 15 Jan 2001 18:31:13 +0000 (18:31 +0000)]
- Make sure to quote the username and password (SF patch #103236 by
  dogfort).

- Don't drop the data argument when calling open_https() from the
  authentication error handler.

24 years agofrom the really-stupid-bug department: uppercase literals should match
Fredrik Lundh [Mon, 15 Jan 2001 18:28:14 +0000 (18:28 +0000)]
from the really-stupid-bug department: uppercase literals should match
uppercase strings also when the IGNORECASE flag is set (bug #128899)

(also added test cases for recently fixed bugs to the regression suite
-- or in other words, check in re_tests.py too...)

24 years agomwh@sourceforge found that UnicodeError can be raised by compiling.
Guido van Rossum [Mon, 15 Jan 2001 18:13:35 +0000 (18:13 +0000)]
mwh@sourceforge found that UnicodeError can be raised by compiling.
Its base class ValueError can be raised too, so catch that.

24 years agoFixing author's email address.
Barry Warsaw [Mon, 15 Jan 2001 17:08:45 +0000 (17:08 +0000)]
Fixing author's email address.

24 years agoCommitting patch #103216, autodetect of dbmmodule support and building
Barry Warsaw [Mon, 15 Jan 2001 17:07:21 +0000 (17:07 +0000)]
Committing patch #103216, autodetect of dbmmodule support and building
of dbmmodule dynamically by default (otherwise it can pull in
dependencies with libdb that croak pybsddb3).  This change moves the
Setup line for dbmmodule to Setup.config.in.

24 years agoGeoffrey Gerrietts discovered that a KeyError was caught that probably
Guido van Rossum [Mon, 15 Jan 2001 16:53:58 +0000 (16:53 +0000)]
Geoffrey Gerrietts discovered that a KeyError was caught that probably
should have been a NameError.  I'm checking in a change that catches
both, just to be sure -- I can't be bothered trying to understand this
code any more. :-)

24 years agoDocument changed default to passive mode.
Guido van Rossum [Mon, 15 Jan 2001 16:37:05 +0000 (16:37 +0000)]
Document changed default to passive mode.

24 years agoAdd note about ftplib defaulting to passive mode.
Guido van Rossum [Mon, 15 Jan 2001 16:36:08 +0000 (16:36 +0000)]
Add note about ftplib defaulting to passive mode.

24 years agoDefault to passive mode. See SF bug #126851.
Guido van Rossum [Mon, 15 Jan 2001 16:32:49 +0000 (16:32 +0000)]
Default to passive mode.  See SF bug #126851.

This is slightly controversial, but after reading the argumentation in
the bug tracker for and against, I believe this is the right solution.
Let me know if it breaks for you, and how.

24 years agoFix from Jack Jansen for the Mac and the Metrowerks compiler, posted
Andrew M. Kuchling [Mon, 15 Jan 2001 16:09:35 +0000 (16:09 +0000)]
Fix from Jack Jansen for the Mac and the Metrowerks compiler, posted
to the Distutils-SIG and archived at
http://mail.python.org/pipermail/distutils-sig/2000-November/001755.html

24 years agoAdded a separate extension (.carbon.slb) for Carbon dynamic modules.
Jack Jansen [Mon, 15 Jan 2001 16:00:40 +0000 (16:00 +0000)]
Added a separate extension (.carbon.slb) for Carbon dynamic modules.

24 years agoRevert a change I accidentally checked in together with Ping's
Guido van Rossum [Mon, 15 Jan 2001 14:34:20 +0000 (14:34 +0000)]
Revert a change I accidentally checked in together with Ping's
"smallest patch ever".

24 years agoadded "magic" number to the _sre module, to avoid weird errors caused
Fredrik Lundh [Mon, 15 Jan 2001 12:46:09 +0000 (12:46 +0000)]
added "magic" number to the _sre module, to avoid weird errors caused
by compiler/engine mismatches

24 years agoSpeed getline_via_fgets(), by supplying two "fast paths", although one is
Tim Peters [Mon, 15 Jan 2001 10:36:56 +0000 (10:36 +0000)]
Speed getline_via_fgets(), by supplying two "fast paths", although one is
faster than the other.  Should be faster for Mark Favas's 254-character
mail log lines, and *is* 3-4% quicker for my test case with much shorter
lines (but they're typical of *my* text files, and I'm tired of optimizing
for everyone else at my expense <wink> -- in fact, the only one who loses
here is Guido ...).

24 years agoUse the "MS" getline hack (fgets()) by default on non-get_unlocked
Tim Peters [Mon, 15 Jan 2001 06:33:19 +0000 (06:33 +0000)]
Use the "MS" getline hack (fgets()) by default on non-get_unlocked
platforms.  See NEWS for details.

24 years agoWhitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters [Mon, 15 Jan 2001 03:34:38 +0000 (03:34 +0000)]
Whitespace normalization.  Top level of Lib now fixed-point for reindent.py!

24 years agoWhitespace normalization.
Tim Peters [Mon, 15 Jan 2001 03:26:36 +0000 (03:26 +0000)]
Whitespace normalization.

24 years agoWhitespace normalization.
Tim Peters [Mon, 15 Jan 2001 01:36:40 +0000 (01:36 +0000)]
Whitespace normalization.

24 years agoWhitespace normalization.
Tim Peters [Mon, 15 Jan 2001 01:18:21 +0000 (01:18 +0000)]
Whitespace normalization.

24 years agoWhitespace normalization.
Tim Peters [Mon, 15 Jan 2001 00:50:52 +0000 (00:50 +0000)]
Whitespace normalization.

24 years ago-- don't use recursion for unbounded non-greedy repeat
Fredrik Lundh [Sun, 14 Jan 2001 23:55:55 +0000 (23:55 +0000)]
-- don't use recursion for unbounded non-greedy repeat
(bugs #115903, #115696)

This is based on a patch by Darrel Gallion.  I'm not 100%
sure about this fix, but I haven't managed to come up with
any test case it cannot handle...

24 years agoWhitespace normalization.
Tim Peters [Sun, 14 Jan 2001 23:47:14 +0000 (23:47 +0000)]
Whitespace normalization.

24 years agoWhitespace normalization.
Tim Peters [Sun, 14 Jan 2001 23:36:06 +0000 (23:36 +0000)]
Whitespace normalization.

24 years ago- Don't hardcode Unix filename syntax when opening ~/.pdbrc.
Guido van Rossum [Sun, 14 Jan 2001 23:29:48 +0000 (23:29 +0000)]
- Don't hardcode Unix filename syntax when opening ~/.pdbrc.
- Conform to standard coding style in a few more places.

24 years ago- Use mimetypes.types_map to initialize extensions_map.
Guido van Rossum [Sun, 14 Jan 2001 23:21:25 +0000 (23:21 +0000)]
- Use mimetypes.types_map to initialize extensions_map.
- Change the default file type to application/octet-stream.
- Add support to recognize .py, .c, .h files as text/plain (this is
  what I use most :-).

24 years agoSF Patch #103211.
Guido van Rossum [Sun, 14 Jan 2001 23:09:35 +0000 (23:09 +0000)]
SF Patch #103211.

Ping apparently doesn't check in Accepted patches, so I'm doing this
for him.

According to Ping: The name of the controller class should be
"Konqueror", not "Konquerer". (See the website
http://www.konqueror.org/.)

24 years agoMake the copyright message the same as for the "real" interpreter.
Guido van Rossum [Sun, 14 Jan 2001 23:04:22 +0000 (23:04 +0000)]
Make the copyright message the same as for the "real" interpreter.

24 years agoConfig file for shared carbon Python.
Jack Jansen [Sun, 14 Jan 2001 23:02:32 +0000 (23:02 +0000)]
Config file for shared carbon Python.

24 years agoAdded PythonCoreCarbon and PythonInterpreterCarbon targets.
Jack Jansen [Sun, 14 Jan 2001 23:01:36 +0000 (23:01 +0000)]
Added PythonCoreCarbon and PythonInterpreterCarbon targets.

24 years agoWhitespace normalization.
Tim Peters [Sun, 14 Jan 2001 21:54:20 +0000 (21:54 +0000)]
Whitespace normalization.

24 years agoSF Patch #103232 by dougfort: Preserve Nonstandard Port Number in Host
Guido van Rossum [Sun, 14 Jan 2001 21:03:01 +0000 (21:03 +0000)]
SF Patch #103232 by dougfort: Preserve Nonstandard Port Number in Host
Header

Dougfort's comments: httplib does not include ':port ' in the HTTP 1.1
'Host:' header.  This causes problems if the server is not listening
on Port 80.  The test case I use is the login to /manage under Zope,
with Zope listening on port 8080. Zope returns a <frameset> with the
<frame> source URLs lacking the :8080.

24 years agoSRE: stricter pattern syntax checking (covers parts of bug #115900)
Fredrik Lundh [Sun, 14 Jan 2001 21:00:44 +0000 (21:00 +0000)]
SRE: stricter pattern syntax checking (covers parts of bug #115900)

24 years agoWhitespace standardization.
Tim Peters [Sun, 14 Jan 2001 18:09:23 +0000 (18:09 +0000)]
Whitespace standardization.

24 years agoreapplied Fred's "recommended style" patch...
Fredrik Lundh [Sun, 14 Jan 2001 15:15:37 +0000 (15:15 +0000)]
reapplied Fred's "recommended style" patch...

24 years agoSRE fixes for 2.1 alpha:
Fredrik Lundh [Sun, 14 Jan 2001 15:06:11 +0000 (15:06 +0000)]
SRE fixes for 2.1 alpha:

-- added some more docstrings
-- fixed typo in scanner class (#125531)
-- the multiline flag (?m) should't affect the \Z operator (#127259)
-- fixed non-greedy backtracking bug (#123769, #127259)
-- added sre.DEBUG flag (currently dumps the parsed pattern structure)
-- fixed a couple of glitches in groupdict (the #126587 memory leak
   had already been fixed by AMK)

24 years agoDocument filterwarnings(..., append=<bool>).
Guido van Rossum [Sun, 14 Jan 2001 14:10:18 +0000 (14:10 +0000)]
Document filterwarnings(..., append=<bool>).

24 years ago- Added keyword argument 'append' to filterwarnings(); if true, this
Guido van Rossum [Sun, 14 Jan 2001 14:08:40 +0000 (14:08 +0000)]
- Added keyword argument 'append' to filterwarnings(); if true, this
  appends to list of filters instead of inserting at the front.  This
  is useful to add a filter with a lower priority than -W options.

- Cosmetic improvements to a docstring and an error message.

24 years agoReverting a dumb experimental version I checked in by mistake.
Tim Peters [Sun, 14 Jan 2001 05:12:40 +0000 (05:12 +0000)]
Reverting a dumb experimental version I checked in by mistake.

24 years agoSF bug 128713: type(mmap_object) blew up on Linux.
Tim Peters [Sun, 14 Jan 2001 05:05:51 +0000 (05:05 +0000)]
SF bug 128713:  type(mmap_object) blew up on Linux.

24 years agoAdded Jeffery Collins.
Tim Peters [Sun, 14 Jan 2001 05:04:40 +0000 (05:04 +0000)]
Added Jeffery Collins.

24 years agoFix a variety of minor nits and typos caught by Chris Ryland
Fred Drake [Sun, 14 Jan 2001 02:57:14 +0000 (02:57 +0000)]
Fix a variety of minor nits and typos caught by Chris Ryland
<cpr@emsoftware.com>.

24 years agomwh: [ Patch #103228 ] traceback.py nit.
Guido van Rossum [Sat, 13 Jan 2001 22:14:31 +0000 (22:14 +0000)]
mwh: [ Patch #103228 ] traceback.py nit.

When the exception has no message, don't insert a colon after the
exception name.

24 years agoSF Patch #103227 by mwh: make code.py appreciate softspace
Guido van Rossum [Sat, 13 Jan 2001 22:10:41 +0000 (22:10 +0000)]
SF Patch #103227 by mwh: make code.py appreciate softspace

24 years agoNeil discovered a bad DECREF on warnoptions, that caused repeated
Guido van Rossum [Sat, 13 Jan 2001 22:06:05 +0000 (22:06 +0000)]
Neil discovered a bad DECREF on warnoptions, that caused repeated
re-initializing Python (Py_Finalize() followed by Py_Initialize()) to
blow up quickly.  With the DECREF removed I can't get it to fail any
more.  (Except it still leaks, but that's probably a separate issue.)

24 years agoNo text file relying on significant trailing whitespace is robust under
Tim Peters [Sat, 13 Jan 2001 19:16:21 +0000 (19:16 +0000)]
No text file relying on significant trailing whitespace is robust under
modification.  Removed the need for that.

24 years agoSF Patch #103225 by Ping: httplib: smallest Python patch ever
Guido van Rossum [Sat, 13 Jan 2001 16:55:33 +0000 (16:55 +0000)]
SF Patch #103225 by Ping: httplib: smallest Python patch ever

  The ASCII-art diagram at the top of httplib contains a backslash at
  the end of a line, which causes Python to remove the newline. This
  one-character patch adds a space after the backslash so it will
  appear at the end of the line in the docstring as intended.

24 years agoMention new curses.panel module
Andrew M. Kuchling [Sat, 13 Jan 2001 14:53:34 +0000 (14:53 +0000)]
Mention new curses.panel module

24 years agoDocument extensions to .pth files.
Martin v. Löwis [Sat, 13 Jan 2001 09:54:41 +0000 (09:54 +0000)]
Document extensions to .pth files.

24 years agoFix stupidity.
Tim Peters [Sat, 13 Jan 2001 03:45:59 +0000 (03:45 +0000)]
Fix stupidity.

24 years agoGuido found a brand new race in tempfile on Linux, due to Linux changing
Tim Peters [Sat, 13 Jan 2001 03:04:02 +0000 (03:04 +0000)]
Guido found a brand new race in tempfile on Linux, due to Linux changing
pid across threads (but in that case, it's still the same process, and so
still sharing the "template" cache in tempfile.py).  Repaired that, and
added a new std test.
On Linux, someone please run that standalone with more files and/or more
threads; e.g.,

    python lib/test/test_threadedtempfile.py -f 1000 -t 10

to run with 10 threads each creating (and deleting) 1000 temp files.

24 years agoFixed Carbon command-dot handling. There is still a problem, though, and you may...
Jack Jansen [Fri, 12 Jan 2001 23:42:28 +0000 (23:42 +0000)]
Fixed Carbon command-dot handling. There is still a problem, though, and you may have to hit it repeatedly.