]> granicus.if.org Git - python/log
python
22 years agoFix SF #659228, 'realpath' function missing from os.path
Neal Norwitz [Fri, 3 Jan 2003 18:01:57 +0000 (18:01 +0000)]
Fix SF #659228, 'realpath' function missing from os.path

Also added realpath = abspath for os2emx, similar to windows/mac
which also don't really implement realpath.

Backport candidate, I think?

22 years agoWrite PEP 301 section
Andrew M. Kuchling [Fri, 3 Jan 2003 16:52:27 +0000 (16:52 +0000)]
Write PEP 301 section
Mention difference between 2.2.2 and 2.3 True and False

22 years agoMention that imaplib now supports SSL -- this wasn't noted before.
Guido van Rossum [Fri, 3 Jan 2003 16:33:49 +0000 (16:33 +0000)]
Mention that imaplib now supports SSL -- this wasn't noted before.

22 years agoqualify known Solaris versions related to the binutils breakage.
Skip Montanaro [Fri, 3 Jan 2003 16:26:23 +0000 (16:26 +0000)]
qualify known Solaris versions related to the binutils breakage.

22 years agoAdd SSL support for imaplib; add empty PEP301 section
Andrew M. Kuchling [Fri, 3 Jan 2003 16:24:28 +0000 (16:24 +0000)]
Add SSL support for imaplib; add empty PEP301 section

22 years agoupdate info about binutils 2.13 breakage on Solaris.
Skip Montanaro [Fri, 3 Jan 2003 16:17:08 +0000 (16:17 +0000)]
update info about binutils 2.13 breakage on Solaris.

22 years ago[Patch #658093 ] Documentation support for PEP 301
Andrew M. Kuchling [Fri, 3 Jan 2003 15:42:14 +0000 (15:42 +0000)]
[Patch #658093 ] Documentation support for PEP 301
  Add two sections to this manual about package meta-data and about
  registering packages

22 years ago[Patch #658094 ] PEP 301 implementation
Andrew M. Kuchling [Fri, 3 Jan 2003 15:29:28 +0000 (15:29 +0000)]
[Patch #658094 ] PEP 301 implementation
   Add the 'register' distutils command

22 years ago[Patch #658094] PEP 301 implementation
Andrew M. Kuchling [Fri, 3 Jan 2003 15:24:36 +0000 (15:24 +0000)]
[Patch #658094] PEP 301 implementation
  Add 'classifiers' keyword to DistributionMetadata

22 years agoFix for bug #661136
Just van Rossum [Fri, 3 Jan 2003 11:18:56 +0000 (11:18 +0000)]
Fix for bug #661136
Lesson learned: kids should not be allowed to use API's starting
with an underscore :-/
zipimport in 2.3a1 is even more broken than I thought: I attemped
to _PyString_Resize a string created by PyString_FromStringAndSize,
which fails for strings with length 0 or 1 since the latter returns
an interned string in those cases. This would cause a SystemError
with empty source files (and no matching pyc) in the zip archive.
I rewrote the offending code to simply allocate a new buffer and
avoid _PyString_Resize altogether.
Added a test that would've caught the problem.

22 years agoAdd contributor.
Raymond Hettinger [Fri, 3 Jan 2003 10:41:50 +0000 (10:41 +0000)]
Add contributor.

22 years agoRemove debugging prints.
Michael W. Hudson [Fri, 3 Jan 2003 10:25:20 +0000 (10:25 +0000)]
Remove debugging prints.

22 years agoSF patch 660559: Use METH_O and METH_NOARGS where possible
Raymond Hettinger [Fri, 3 Jan 2003 08:24:58 +0000 (08:24 +0000)]
SF patch 660559: Use METH_O and METH_NOARGS where possible

Simplify code and speed access by using PyArg_UnpackTuple, METH_O and
METH_NOARGS in three modules that can benefit from it.

22 years agoFixed markup.
David Goodger [Fri, 3 Jan 2003 03:30:21 +0000 (03:30 +0000)]
Fixed markup.

22 years agoUpdated (2.3 OK now)
David Goodger [Fri, 3 Jan 2003 03:29:58 +0000 (03:29 +0000)]
Updated (2.3 OK now)

22 years agoFix error in previous correction; thanks, Just!
Andrew M. Kuchling [Thu, 2 Jan 2003 23:50:18 +0000 (23:50 +0000)]
Fix error in previous correction; thanks, Just!

22 years agoJack complained that on test_crlf_separation() was failing on MacOS9
Barry Warsaw [Thu, 2 Jan 2003 22:48:36 +0000 (22:48 +0000)]
Jack complained that on test_crlf_separation() was failing on MacOS9
because the test file, msg_26.txt which has \r\n line endings, was
getting munged by cvs, which knows to do line ending conversions for
text files.  But we want \r\n to be preserved on all platforms, so we
cvs admin'd the file to be -kb (binary), which means we have to open
the file in binary mode to preserve these line ends.  Hopefully this
will be the end of the thrashing on this issue (but probably not).

Test passes on *nix now, and Tim confirms it passes on Windows.  We'll
leave it to Jack to test MacOS.

22 years agoFix PEP 302 description; bump version number
Andrew M. Kuchling [Thu, 2 Jan 2003 21:33:15 +0000 (21:33 +0000)]
Fix PEP 302 description; bump version number

22 years agoThe tzinfo methods utcoffset() and dst() must return a timedelta object
Tim Peters [Thu, 2 Jan 2003 21:28:08 +0000 (21:28 +0000)]
The tzinfo methods utcoffset() and dst() must return a timedelta object
(or None) now.  In 2.3a1 they could also return an int or long, but that
was an unhelpfully redundant leftover from an earlier version wherein
they couldn't return a timedelta.  TOOWTDI.

22 years agoAllow list sort's comparison function to explicitly be None. See SF patch
Skip Montanaro [Thu, 2 Jan 2003 20:51:08 +0000 (20:51 +0000)]
Allow list sort's comparison function to explicitly be None.  See SF patch
661092.

22 years ago1. Remove obsolete, incorrect comment on non-package installation
Kurt B. Kaiser [Thu, 2 Jan 2003 20:33:26 +0000 (20:33 +0000)]
1. Remove obsolete, incorrect comment on non-package installation
2. Add more .txt files to installation
3. Fix the reference to Visual Python, s/b VPython

22 years agoastimezone() internals: if utcoffset() returns a duration, complain if
Tim Peters [Thu, 2 Jan 2003 19:35:54 +0000 (19:35 +0000)]
astimezone() internals:  if utcoffset() returns a duration, complain if
dst() returns None (instead of treating that as 0).

22 years agoAdded a section to record datetime changes. There's apparently going to
Tim Peters [Thu, 2 Jan 2003 19:10:19 +0000 (19:10 +0000)]
Added a section to record datetime changes.  There's apparently going to
be an unbounded number of API changes <0.6 wink>.

22 years agoThe astimezone() correctness proof endured much pain to prove what
Tim Peters [Thu, 2 Jan 2003 17:55:03 +0000 (17:55 +0000)]
The astimezone() correctness proof endured much pain to prove what
turned out to be 3 special cases of a single more-general result.
Proving the latter instead is a real simplification.

22 years agoImprove exception handling.
Kurt B. Kaiser [Thu, 2 Jan 2003 17:09:34 +0000 (17:09 +0000)]
Improve exception handling.

22 years agoThere are two more copyright notices in the Windows world:
Tim Peters [Thu, 2 Jan 2003 17:09:19 +0000 (17:09 +0000)]
There are two more copyright notices in the Windows world:

PC/python_nt.rc sets up the DLL version resource (displayed when you
right-click on the DLL and select Properties).

PCbuld/python20.wse sets up the installer version resource (displayed
when you right-click on the installer .exe and select Properties).  Turns
out this one hadn't been updated since 2001 <frown>!

22 years agoSF bug 661086: datetime.today() truncates microseconds.
Tim Peters [Thu, 2 Jan 2003 16:32:54 +0000 (16:32 +0000)]
SF bug 661086: datetime.today() truncates microseconds.
On Windows, it was very common to get microsecond values (out of
.today() and .now()) of the form 480999, i.e. with three trailing
nines.  The platform precision is .001 seconds, and fp rounding
errors account for the rest.  Under the covers, that 480999 started
life as the fractional part of a timestamp, like .4809999978.
Rounding that times 1e6 cures the irritation.

Confession:  the platform precision isn't really .001 seconds.  It's
usually worse.  What actually happens is that MS rounds a cruder value
to a multiple of .001, and that suffers its own rounding errors.

A tiny bit of refactoring added a new internal utility to round
doubles.

22 years agoUpdate the copyright year.
Guido van Rossum [Thu, 2 Jan 2003 16:31:35 +0000 (16:31 +0000)]
Update the copyright year.

22 years agoAnother copyright update. (JvR: can you backport this to the 2.3a1
Guido van Rossum [Thu, 2 Jan 2003 16:27:15 +0000 (16:27 +0000)]
Another copyright update.  (JvR: can you backport this to the 2.3a1
release branch?)

22 years agoSF bug 660795: logging missing from Python 2.3a1 for Windows.
Tim Peters [Thu, 2 Jan 2003 16:02:27 +0000 (16:02 +0000)]
SF bug 660795:  logging missing from Python 2.3a1 for Windows.
Added the logging package.  In the meantime, Neal Norwitz added a
test_logging.py to the std test suite, which would have caught this
oversight in the Windows installer.

22 years agoAdd some version info for new methods and class
Neal Norwitz [Thu, 2 Jan 2003 15:32:00 +0000 (15:32 +0000)]
Add some version info for new methods and class

22 years agoSF #660795
Neal Norwitz [Thu, 2 Jan 2003 14:56:39 +0000 (14:56 +0000)]
SF #660795

Add a test for logging from Vinay Sajip (module author)

22 years agoFix an example
Andrew M. Kuchling [Thu, 2 Jan 2003 13:42:32 +0000 (13:42 +0000)]
Fix an example

22 years agoReplaced imp.set_frozenmodules() cruft with proper zipimport support.
Just van Rossum [Thu, 2 Jan 2003 13:13:01 +0000 (13:13 +0000)]
Replaced imp.set_frozenmodules() cruft with proper zipimport support.
This work uncovered the zipimport bug in 2.3a1 -- wish I'd had time to
do this before the release :-(.

22 years agoUgh, zipimport is virtually broken in 2.3a1 :-( It worked by accident in
Just van Rossum [Thu, 2 Jan 2003 12:55:48 +0000 (12:55 +0000)]
Ugh, zipimport is virtually broken in 2.3a1 :-( It worked by accident in
the test set as it only tested with a zip archive in the current directory,
but it doesn't work at all for packages when the zip archive was specified
as an absolute path. It's a real embarrassing bug: a strchr call should
have been strrchr; fever apparently implies dyslexia.

Second stupid bug: the zipimport test failed with a name error
__importer__ (which I had renamed to __loader__ everywhere but here).
I would've sworn I ran the test after that change but that can't be true.
What I don't understand that noone reported a failing test_zipimport.py
before the release of 2.3a1.

22 years agoEMX fork() emulation not good enough to cope with test_socketserver
Andrew MacIntyre [Thu, 2 Jan 2003 12:49:00 +0000 (12:49 +0000)]
EMX fork() emulation not good enough to cope with test_socketserver

22 years agoOS/2 sockets do not support AF_UNIX, even though EMX headers define it
Andrew MacIntyre [Thu, 2 Jan 2003 12:45:34 +0000 (12:45 +0000)]
OS/2 sockets do not support AF_UNIX, even though EMX headers define it

22 years agocatch up with zipimport changes to std getpathp.c
Andrew MacIntyre [Thu, 2 Jan 2003 12:41:58 +0000 (12:41 +0000)]
catch up with zipimport changes to std getpathp.c

22 years agobring structure closer to std config.c, whitespace normalisation
Andrew MacIntyre [Thu, 2 Jan 2003 12:40:41 +0000 (12:40 +0000)]
bring structure closer to std config.c, whitespace normalisation

22 years agofix a merge mistake - readline not built by default
Andrew MacIntyre [Thu, 2 Jan 2003 12:38:39 +0000 (12:38 +0000)]
fix a merge mistake - readline not built by default

22 years ago- documented Ellipsis, NotImplemented
Fred Drake [Thu, 2 Jan 2003 05:13:51 +0000 (05:13 +0000)]
- documented Ellipsis, NotImplemented
- minor markup changes
- indented for consistency with newer content

22 years agoAdd dependency info for the recently added lib/libconsts.tex.
Fred Drake [Thu, 2 Jan 2003 05:00:12 +0000 (05:00 +0000)]
Add dependency info for the recently added lib/libconsts.tex.

22 years agoDocument that apply() is deprecated. See:
Fred Drake [Thu, 2 Jan 2003 04:54:04 +0000 (04:54 +0000)]
Document that apply() is deprecated.  See:
http://mail.python.org/pipermail/python-dev/2003-January/031556.html

22 years agoCompleted astimezone's correctness proof. That doesn't mean it's
Tim Peters [Thu, 2 Jan 2003 03:14:59 +0000 (03:14 +0000)]
Completed astimezone's correctness proof.  That doesn't mean it's
correct by your lights, it means that-- barring coding errors --it
implements what it intended to implement.

22 years agoClearing out old patch queue. Patch #558547, make SocketServer more
Anthony Baxter [Thu, 2 Jan 2003 03:07:48 +0000 (03:07 +0000)]
Clearing out old patch queue. Patch #558547, make SocketServer more
robust. This makes socketserver's close() method callable repeatedly
without error - similar to other file-like objects.

22 years agoAdd byext.py
Guido van Rossum [Thu, 2 Jan 2003 02:24:22 +0000 (02:24 +0000)]
Add byext.py

22 years agoA quicker astimezone() implementation, rehabilitating an earlier
Tim Peters [Wed, 1 Jan 2003 21:51:37 +0000 (21:51 +0000)]
A quicker astimezone() implementation, rehabilitating an earlier
suggestion from Guido, along with a formal correctness proof of the
trickiest bit.  The intricacy of the proof reveals how delicate this
is, but also how robust the conclusion:  correctness doesn't rely on
dst() returning +- one hour (not all real time zones do!), it only
relies on:

1. That dst() returns a (any) non-zero value if and only if daylight
   time is in effect.

and

2. That the tzinfo subclass implements a consistent notion of time zone.

The meaning of "consistent" was a hidden assumption, which is now an
explicit requirement in the docs.  Alas, it's an unverifiable (by the
datetime implementation) requirement, but so it goes.

22 years agomention built-in constants.
Skip Montanaro [Wed, 1 Jan 2003 20:37:14 +0000 (20:37 +0000)]
mention built-in constants.

22 years agoprocess libconsts.tex
Skip Montanaro [Wed, 1 Jan 2003 20:34:00 +0000 (20:34 +0000)]
process libconsts.tex

22 years agonew section - builtin constants
Skip Montanaro [Wed, 1 Jan 2003 20:33:38 +0000 (20:33 +0000)]
new section - builtin constants

22 years agoadd find-uname.py
Skip Montanaro [Wed, 1 Jan 2003 20:27:13 +0000 (20:27 +0000)]
add find-uname.py

22 years agoSearch for Unicode character names using regular expressions.
Skip Montanaro [Wed, 1 Jan 2003 20:26:47 +0000 (20:26 +0000)]
Search for Unicode character names using regular expressions.

22 years agoSplit OPT make variable into OPT and BASECFLAGS. The latter contains those
Skip Montanaro [Wed, 1 Jan 2003 20:07:49 +0000 (20:07 +0000)]
Split OPT make variable into OPT and BASECFLAGS.  The latter contains those
compiler flags which are necessary to get a clean compile.  The former is
for user-specified optimizer, debug, trace fiddling.  See patch 640843.

Add /sw/lib and /sw/include to setup.py search paths on Darwin to take
advantage of fink goodies.

Add scriptsinstall target to Makefile to install certain scripts from
Tools/scripts directory.

22 years agoMove _PyInt_Init() into pythonrun.h, since all the other _Init()
Neal Norwitz [Wed, 1 Jan 2003 15:18:32 +0000 (15:18 +0000)]
Move _PyInt_Init() into pythonrun.h, since all the other _Init()
functions are here.  Suggested by Skip.

22 years agoRevert last change -- test works on HPUX again after Martin's checkin
Neal Norwitz [Wed, 1 Jan 2003 14:53:27 +0000 (14:53 +0000)]
Revert last change -- test works on HPUX again after Martin's checkin
to 'properly configure the slave terminal'

See SF patch # 656590 for the details.

22 years agoSilly little script to print statistics (files, lines, words) by
Guido van Rossum [Wed, 1 Jan 2003 14:41:25 +0000 (14:41 +0000)]
Silly little script to print statistics (files, lines, words) by
extension.  Could use some work, but already very useful.

22 years agoExpose I_ constants. Auto-detect stropts.h. Properly configure the slave terminal.
Martin v. Löwis [Wed, 1 Jan 2003 09:51:12 +0000 (09:51 +0000)]
Expose I_ constants. Auto-detect stropts.h. Properly configure the slave terminal.

22 years agoRevert merge from 2.3 alpha 1 release branch; the change does not
Fred Drake [Wed, 1 Jan 2003 04:50:32 +0000 (04:50 +0000)]
Revert merge from 2.3 alpha 1 release branch; the change does not
apply to the trunk.

22 years agodatetimetz_astimezone(): Speed optimizations -- although I'd rather
Tim Peters [Wed, 1 Jan 2003 04:48:01 +0000 (04:48 +0000)]
datetimetz_astimezone():  Speed optimizations -- although I'd rather
find a more elegant algorithm (OTOH, the hairy new implementation allows
user-written tzinfo classes to be elegant, so it's a big win even if
astimezone() remains hairy).

Darn!  I've only got 10 minutes left to get falling-down drunk!  I suppose
I'll have to smoke crack instead now.

22 years agoThe failure of the last-second addition to the timezone coversion test is
Tim Peters [Wed, 1 Jan 2003 04:18:51 +0000 (04:18 +0000)]
The failure of the last-second addition to the timezone coversion test is
understood now:  it can't work.  Added comments explaining why (it's "the
usual"-- unrepresentable hours in local time --but in a slightly different
guise).

22 years agoMerging in changes from r23a1-branch. Doc/makefile had conflicts, which
Tim Peters [Wed, 1 Jan 2003 02:14:12 +0000 (02:14 +0000)]
Merging in changes from r23a1-branch.  Doc/makefile had conflicts, which
I leave to Fred to sort out.

22 years agoMore installation info. Bump alpha version.
Kurt B. Kaiser [Wed, 1 Jan 2003 01:14:13 +0000 (01:14 +0000)]
More installation info.  Bump alpha version.

22 years agoDebugger was tracing through rpc.py when IDLEfork was not started
Kurt B. Kaiser [Wed, 1 Jan 2003 00:26:41 +0000 (00:26 +0000)]
Debugger was tracing through rpc.py when IDLEfork was not started
from its source directory.  Generalize the "workaround" (though
the latter seems a reasonable solution?) to handle this.

22 years agoImprove exception handling.
Kurt B. Kaiser [Tue, 31 Dec 2002 23:18:00 +0000 (23:18 +0000)]
Improve exception handling.

22 years agoTry to get compilation working for cygwin
Neal Norwitz [Tue, 31 Dec 2002 21:55:16 +0000 (21:55 +0000)]
Try to get compilation working for cygwin

22 years agoPatch #660485: Cygwin _tkinter Tcl/Tk 8.3 patch
Jason Tishler [Tue, 31 Dec 2002 20:30:46 +0000 (20:30 +0000)]
Patch #660485: Cygwin _tkinter Tcl/Tk 8.3 patch

The attached patch enables Cygwin Python to
build cleanly against the latest Cygwin Tcl/Tk
which is based on Tcl/Tk 8.3. It also prevents
building against the real X headers, if installed.

22 years agoThe trunk is (nominally) post-alpha.
Fred Drake [Tue, 31 Dec 2002 20:26:25 +0000 (20:26 +0000)]
The trunk is (nominally) post-alpha.

22 years agoMerge to trunk from release branch:
Guido van Rossum [Tue, 31 Dec 2002 19:50:03 +0000 (19:50 +0000)]
Merge to trunk from release branch:

Plug the leak that Tim just reported.

22 years agoMake two tests non-locale-dependent
Barry Warsaw [Tue, 31 Dec 2002 19:27:45 +0000 (19:27 +0000)]
Make two tests non-locale-dependent

22 years agoResults of a rewrite pass
Andrew M. Kuchling [Tue, 31 Dec 2002 18:34:54 +0000 (18:34 +0000)]
Results of a rewrite pass

22 years ago- use classdesc where we can (for better indexing)
Fred Drake [Tue, 31 Dec 2002 18:31:48 +0000 (18:31 +0000)]
- use classdesc where we can (for better indexing)
- more style consistency crud

22 years agoReplace all but one explicit emit('SET_LINENO') with call to set_lineno().
Jeremy Hylton [Tue, 31 Dec 2002 18:26:17 +0000 (18:26 +0000)]
Replace all but one explicit emit('SET_LINENO') with call to set_lineno().

Remove broken code in visitDict().  I assume the code was trying to
add set lineno events for each line of a dict constructor, but I think
it was using the wrong object (node instead of k or v).

22 years agoDon't let the docstring end up in __main__.__doc__
Tony Lownds [Tue, 31 Dec 2002 18:22:37 +0000 (18:22 +0000)]
Don't let the docstring end up in __main__.__doc__

22 years agotest01_close_dbenv_before_db(): Added an XXX comment that this test is
Barry Warsaw [Tue, 31 Dec 2002 18:21:43 +0000 (18:21 +0000)]
test01_close_dbenv_before_db(): Added an XXX comment that this test is
BerkeleyDB version dependent.

22 years agoComment out test, since it hangs on HPUX, still investigating
Neal Norwitz [Tue, 31 Dec 2002 18:21:11 +0000 (18:21 +0000)]
Comment out test, since it hangs on HPUX, still investigating

22 years agoSF patch [ 597919 ] compiler package and SET_LINENO
Jeremy Hylton [Tue, 31 Dec 2002 18:17:44 +0000 (18:17 +0000)]
SF patch [ 597919 ] compiler package and SET_LINENO

A variety of changes from Michael Hudson to get the compiler working
with 2.3.  The primary change is the handling of SET_LINENO:

# The set_lineno() function and the explicit emit() calls for
# SET_LINENO below are only used to generate the line number table.
# As of Python 2.3, the interpreter does not have a SET_LINENO
# instruction.  pyassem treats SET_LINENO opcodes as a special case.

A few other small changes:
 - Remove unused code from pycodegen and pyassem.
 - Fix error handling in parsermodule.  When PyParser_SimplerParseString()
   fails, it sets an exception with detailed info.  The parsermodule
   was clobbering that exception and replacing it was a generic
   "could not parse string" exception.  Keep the original exception.

22 years agoGeneral style conformance. Markup some unmarked constructs.
Fred Drake [Tue, 31 Dec 2002 18:13:11 +0000 (18:13 +0000)]
General style conformance.  Markup some unmarked constructs.

22 years agoUpdate (slightly) for Alpha release
Kurt B. Kaiser [Tue, 31 Dec 2002 18:12:04 +0000 (18:12 +0000)]
Update (slightly) for Alpha release

22 years agoRemove bogus test; the master is not a terminal on Solaris and HP-UX.
Martin v. Löwis [Tue, 31 Dec 2002 18:05:15 +0000 (18:05 +0000)]
Remove bogus test; the master is not a terminal on Solaris and HP-UX.

22 years agoMove history to HISTORY.txt
Kurt B. Kaiser [Tue, 31 Dec 2002 17:57:44 +0000 (17:57 +0000)]
Move history to HISTORY.txt

22 years agoUpdate for Alpha 0 Release
Kurt B. Kaiser [Tue, 31 Dec 2002 17:56:18 +0000 (17:56 +0000)]
Update for Alpha 0 Release

22 years agoAdd recipe for creating NEWS.html.
Guido van Rossum [Tue, 31 Dec 2002 17:51:30 +0000 (17:51 +0000)]
Add recipe for creating NEWS.html.

22 years agoA new, and much hairier, implementation of astimezone(), building on
Tim Peters [Tue, 31 Dec 2002 17:36:56 +0000 (17:36 +0000)]
A new, and much hairier, implementation of astimezone(), building on
an idea from Guido.  This restores that the datetime implementation
never passes a datetime d to a tzinfo method unless d.tzinfo is the
tzinfo instance whose method is being called.  That in turn allows
enormous simplifications in user-written tzinfo classes (see the Python
sandbox US.py and EU.py for fully fleshed-out examples).

d.astimezone(tz) also raises ValueError now if d lands in the one hour
of the year that can't be expressed in tz (this can happen iff tz models
both standard and daylight time).  That it used to return a nonsense
result always ate at me, and it turned out that it seemed impossible to
force a consistent nonsense result under the new implementation (which
doesn't know anything about how tzinfo classes implement their methods --
it can only infer properties indirectly).  Guido doesn't like this --
expect it to change.

New tests of conversion between adjacent DST-aware timezones don't pass
yet, and are commented out.

Running the datetime tests in a loop under a debug build leaks 9
references per test run, but I don't believe the datetime code is the
cause (it didn't leak the last time I changed the C code, and the leak
is the same if I disable all the tests that invoke the only function
that changed here).  I'll pursue that next.

22 years agoEliminate C++ comment.
Martin v. Löwis [Tue, 31 Dec 2002 17:34:30 +0000 (17:34 +0000)]
Eliminate C++ comment.

22 years agoUse funcdesc instead of classdesc to be consistent with out sections.
Raymond Hettinger [Tue, 31 Dec 2002 17:24:50 +0000 (17:24 +0000)]
Use funcdesc instead of classdesc to be consistent with out sections.

22 years agoFurther cleanup of exceptions. All interpolation-related exceptions
Fred Drake [Tue, 31 Dec 2002 17:23:27 +0000 (17:23 +0000)]
Further cleanup of exceptions.  All interpolation-related exceptions
now derive from InterpolationError, which is not raised directly (only
subclasses get raised).  This matches what the docs already said.

22 years agoBernhard Herzog's paragraph and string-filling code. I've been using it for
Skip Montanaro [Tue, 31 Dec 2002 16:56:20 +0000 (16:56 +0000)]
Bernhard Herzog's paragraph and string-filling code.  I've been using it for
a month or two with great success.  Barry may want to tweak it some, but I
think it's a worthwhile enough addition to get some more people trying it
out.

22 years agoKeep __main__ namespace clean
Tony Lownds [Tue, 31 Dec 2002 16:52:44 +0000 (16:52 +0000)]
Keep __main__ namespace clean

22 years agomodulefinder.py moved to Lib/
Just van Rossum [Tue, 31 Dec 2002 16:38:01 +0000 (16:38 +0000)]
modulefinder.py moved to Lib/

22 years agoSpelling fix
Raymond Hettinger [Tue, 31 Dec 2002 16:37:03 +0000 (16:37 +0000)]
Spelling fix

22 years agoFix an out-of-bound index in pmerge() discovered by Zooko (SF bug
Guido van Rossum [Tue, 31 Dec 2002 16:33:01 +0000 (16:33 +0000)]
Fix an out-of-bound index in pmerge() discovered by Zooko (SF bug
645404).  I'm not 100% sure this is the right fix, so I'll keep the
bug report open for Samuele, but this fixes the index error and passes
the test suite (and I can't see why it *shouldn't* be the right fix
:-).

22 years agopatch attached to sf item #643711:
Just van Rossum [Tue, 31 Dec 2002 16:33:00 +0000 (16:33 +0000)]
patch attached to sf item #643711:
any_missing() returns less bogus missing modules.

- I've rewritten scan_code() more or less from scratch,
factored bits and pieces out for readability.
- keep track of global assignments and failed imports per
module; use this to determine whether the Y in "from X
import Y" is a submodule or just a global name. This is not
100% doable: you can't tell which symbols are imported when
doing a star import of a non-Python module short of actually
importing it.
- added a new method to ModuleFinder: any_missing_maybe(),
which returns *two* lists, one with certain misses, one with
possible misses. The possible misses are *very* often false
alarms, so it's useful to keep this list separate.
any_misses() now simply returns the union of
any_missing_maybe().

TODO: documentation, test_modulefinder.py

22 years agoAdd markup for time object.
Raymond Hettinger [Tue, 31 Dec 2002 16:30:49 +0000 (16:30 +0000)]
Add markup for time object.
Cleanup whitespace.
Fix unbalanced parenthesis.

22 years agomoving modulefinder.py to the standard library
Just van Rossum [Tue, 31 Dec 2002 16:27:33 +0000 (16:27 +0000)]
moving modulefinder.py to the standard library

22 years agoFix compilation errors on HPUX11
Neal Norwitz [Tue, 31 Dec 2002 16:16:07 +0000 (16:16 +0000)]
Fix compilation errors on HPUX11

22 years agoWhitespace Normalization
Kurt B. Kaiser [Tue, 31 Dec 2002 16:03:23 +0000 (16:03 +0000)]
Whitespace Normalization

22 years agoRemoved the now-untrue (or soon-to-be untrue) part of the astimezone()
Tim Peters [Tue, 31 Dec 2002 16:01:47 +0000 (16:01 +0000)]
Removed the now-untrue (or soon-to-be untrue) part of the astimezone()
docs.  Replaced it with an XXX block, because the hoped-for treatment
of DST endcases remains unclear (Guido doesn't really like raising an
exception when it's impossible to deliver a correct result, but so
far I have no way in hand to consistently deliver a defined incorrect
result either).

22 years agoremoved unused get_short() function
Just van Rossum [Tue, 31 Dec 2002 15:47:42 +0000 (15:47 +0000)]
removed unused get_short() function

22 years agoSet the release date.
Guido van Rossum [Tue, 31 Dec 2002 15:47:36 +0000 (15:47 +0000)]
Set the release date.

22 years agoDocument the new ,netrc awareness in nntplib.
Eric S. Raymond [Tue, 31 Dec 2002 15:28:44 +0000 (15:28 +0000)]
Document the new ,netrc awareness in nntplib.