]> granicus.if.org Git - python/log
python
24 years agoDonn Cave <donn@oz.net>:
Fred Drake [Sat, 7 Oct 2000 12:25:17 +0000 (12:25 +0000)]
Donn Cave <donn@oz.net>:
Generated files for BeOS R5.

24 years agoRecord bugs found when comparing the module with DOM Core Level 2.
Martin v. Löwis [Sat, 7 Oct 2000 12:10:28 +0000 (12:10 +0000)]
Record bugs found when comparing the module with DOM Core Level 2.

24 years agoPrevent possible buffer overflow exploits under Windows. As per (the very quick...
Mark Hammond [Sat, 7 Oct 2000 11:10:50 +0000 (11:10 +0000)]
Prevent possible buffer overflow exploits under Windows.  As per (the very quick) patch Patch #101801.

24 years agoSRE didn't handle character category followed by hyphen inside a
Fredrik Lundh [Sat, 7 Oct 2000 10:16:19 +0000 (10:16 +0000)]
SRE didn't handle character category followed by hyphen inside a
character class.  Fix provided by Andrew Kuchling.  Closes bug
#116251.

24 years ago[ Bug #116174 ] using %% in cstrings sometimes fails with unicode paramsFix for the...
Marc-André Lemburg [Sat, 7 Oct 2000 08:54:09 +0000 (08:54 +0000)]
[ Bug #116174 ] using %% in cstrings sometimes fails with unicode paramsFix for the bug reported in Bug #116174: "%% %s" % u"abc" failed due
to the way string formatting delegated work to the Unicode formatting
function.

24 years agoUpdated test with a case which checks for the bug reported in
Marc-André Lemburg [Sat, 7 Oct 2000 08:52:45 +0000 (08:52 +0000)]
Updated test with a case which checks for the bug reported in

24 years agoPossible fix for Skip's bug 116136 (sre recursion limit hit in tokenize.py).
Tim Peters [Sat, 7 Oct 2000 05:09:39 +0000 (05:09 +0000)]
Possible fix for Skip's bug 116136 (sre recursion limit hit in tokenize.py).
tokenize.py has always used naive regexps for matching string literals,
and that appears to trigger the sre recursion limit on Skip's platform (he
has very long single-line string literals).  Replaced all of tokenize.py's
string regexps with the "unrolled" forms used in IDLE, where they're known to
handle even absurd (multi-megabyte!) string literals without trouble.  See
Friedl's book for explanation (at heart, the naive regexps create a backtracking
choice point for each character in the literal, while the unrolled forms create
none).

24 years agoPrep Windows installer for 2.0c1: title and build number.
Tim Peters [Sat, 7 Oct 2000 04:04:07 +0000 (04:04 +0000)]
Prep Windows installer for 2.0c1:  title and build number.

24 years agoFix for next iteration of SF bug 115690 (Unicode headaches in IDLE). The
Tim Peters [Fri, 6 Oct 2000 23:09:00 +0000 (23:09 +0000)]
Fix for next iteration of SF bug 115690 (Unicode headaches in IDLE).  The
parsing functions in support of auto-indent weren't expecting Unicode
strings, but text.get() can now return them (although it remains muddy as
to exactly when or why that can happen).  Fixed that with a Big Hammer.

24 years agoAdd .toxml test case, as proposed by Alex Martelli in bug report #116244.
Martin v. Löwis [Fri, 6 Oct 2000 22:42:55 +0000 (22:42 +0000)]
Add .toxml test case, as proposed by Alex Martelli in bug report #116244.

24 years agominidom: access attribute value before printing it
Martin v. Löwis [Fri, 6 Oct 2000 22:36:03 +0000 (22:36 +0000)]
minidom: access attribute value before printing it
         correct order of constructor args in createAttributeNS
pulldom: use symbolic names for uri and localnames
         correct usage of createAttribute and setAttributeNode signatures.

24 years agoAdd a test case for reporting the file name, and for reporting an error
Martin v. Löwis [Fri, 6 Oct 2000 21:13:23 +0000 (21:13 +0000)]
Add a test case for reporting the file name, and for reporting an error
for incomplete input.

24 years agoIn an incremental parser, close the parser after feeding everything.
Martin v. Löwis [Fri, 6 Oct 2000 21:12:12 +0000 (21:12 +0000)]
In an incremental parser, close the parser after feeding everything.

24 years agoDon't use a file object as system id; try to propagate the file name to
Martin v. Löwis [Fri, 6 Oct 2000 21:11:20 +0000 (21:11 +0000)]
Don't use a file object as system id; try to propagate the file name to
the InputSource.

24 years agoDon't report a final chunk for an external entity parser.
Martin v. Löwis [Fri, 6 Oct 2000 21:10:34 +0000 (21:10 +0000)]
Don't report a final chunk for an external entity parser.

24 years agoMove translation from expat.error to SAXParseException into feed, so that
Martin v. Löwis [Fri, 6 Oct 2000 21:08:59 +0000 (21:08 +0000)]
Move translation from expat.error to SAXParseException into feed, so that
callers of feed will get a SAXException.
In close, feed the last chunk first before calling endDocument, so that
the parser may report errors before the end of the document. Don't do
anything in a nested parser.
Don't call endDocument in parse; that will be called in close.
Use self._source for finding the SystemID; XML_GetBase will be cleared in
case of an error.

24 years agoInclude more information from the docstrings.
Fred Drake [Fri, 6 Oct 2000 21:07:14 +0000 (21:07 +0000)]
Include more information from the docstrings.

24 years agoAnother name.
Fred Drake [Fri, 6 Oct 2000 21:00:40 +0000 (21:00 +0000)]
Another name.

24 years agoDonn Cave <donn@oz.net>:
Fred Drake [Fri, 6 Oct 2000 20:42:33 +0000 (20:42 +0000)]
Donn Cave <donn@oz.net>:
Fix large file support for BeOS.

This closes SourceForge patch #101773.  Refer to the patch discussion for
information on possible alternate fixes.

24 years agoFix a couple of typos in docstrings.
Fred Drake [Fri, 6 Oct 2000 20:28:46 +0000 (20:28 +0000)]
Fix a couple of typos in docstrings.

24 years agoFix really bad typo, noted by Neil Schemenauer <nas@arctrix.com>.
Fred Drake [Fri, 6 Oct 2000 20:16:50 +0000 (20:16 +0000)]
Fix really bad typo, noted by Neil Schemenauer <nas@arctrix.com>.

24 years agoAdd notes on the requirements for subclasses.
Fred Drake [Fri, 6 Oct 2000 20:04:48 +0000 (20:04 +0000)]
Add notes on the requirements for subclasses.

This closes SourceForge bug #115928.

24 years agoRevise the versioning information to say that this was revised in 1.6,
Fred Drake [Fri, 6 Oct 2000 20:01:23 +0000 (20:01 +0000)]
Revise the versioning information to say that this was revised in 1.6,
not added then, and note what the change was (ncurses, change to a package).

24 years agoMade a number of revisions suggested by Fredrik Lundh.
Fred Drake [Fri, 6 Oct 2000 19:59:22 +0000 (19:59 +0000)]
Made a number of revisions suggested by Fredrik Lundh.
Revised the first paragraph so it doesn't sound like it was written
when 7-bit strings were assumed; note that Unicode strings can be used.

24 years agotest_linuxaudio:
Jeremy Hylton [Fri, 6 Oct 2000 19:39:55 +0000 (19:39 +0000)]
test_linuxaudio:
    read the header from the .au file and do a sanity check
    pass only the data to the audio device
    call flush() so that program does not exit until playback is complete
    call all the other methods to verify that they work minimally
    call setparameters with a bunch of bugs arguments

linuxaudiodev.c:
    use explicit O_WRONLY and O_RDONLY instead of 1 and 0
    add a string name to each of the entries in audio_types[]
    add AFMT_A_LAW to the list of known formats
    add x_mode attribute to lad object, stores imode from open call
    test ioctl return value as == -1, not < 0
    in read() method, resize string before return
    add getptr() method, that calls does ioctl on GETIPTR or GETOPTR
        depending on x_mode
    in setparameters() method, do better error checking and raise
        ValueErrors; also use ioctl calls recommended by Open Sound
        System Programmer's Guido (www.opensound.com)
    use PyModule_AddXXX to define names in module

24 years agoIt turns out that Guido does not like or encourage the use of the term
Fred Drake [Fri, 6 Oct 2000 19:39:47 +0000 (19:39 +0000)]
It turns out that Guido does not like or encourage the use of the term
"disciplines" for the __*__() methods, so they should be referred to as
"methods" or "special methods", as appropriate in context.

24 years ago__getslice__(): Make this use the constructor form that gets a sequence
Fred Drake [Fri, 6 Oct 2000 19:26:01 +0000 (19:26 +0000)]
__getslice__():  Make this use the constructor form that gets a sequence
                 as a parameter; this was the only use of the base
                 constructor or surgical alteration of another object's
                 data attribute.

This change simplifies the constructor requirements for subclasses.

This relates to SourceForge bug #115928.

24 years agoAdded a new "base" type, IOobject for which most of the
Jim Fulton [Fri, 6 Oct 2000 19:24:23 +0000 (19:24 +0000)]
Added a new "base" type, IOobject for which most of the
operations are defined.  This will, hopefully clarify
some of the logic.

Added close test to raise proper error when operations
are performed on closed StringIOs.

Added a position argument to the truncate method.

Added a size argument to readline.

Added PyArg_Parse calls for methods that don't take arguments to
make sure they don't take arguments.

24 years agoUncommented tests that failed for cStringIO,
Jim Fulton [Fri, 6 Oct 2000 19:21:32 +0000 (19:21 +0000)]
Uncommented tests that failed for cStringIO,

Added missing clode to make the clode test test a close. ;)

24 years agoAdding Jeremy's new test_import (SF patch 101709).
Tim Peters [Fri, 6 Oct 2000 18:46:22 +0000 (18:46 +0000)]
Adding Jeremy's new test_import (SF patch 101709).

24 years agoAdd SAXReaderNotAvailable, and use it to distinguish between an
Martin v. Löwis [Fri, 6 Oct 2000 17:41:52 +0000 (17:41 +0000)]
Add SAXReaderNotAvailable, and use it to distinguish between an
ImportError, and a missing driver.

24 years agoRemoving these scripts. redemo.py lives on in Tools/scripts/.
Guido van Rossum [Fri, 6 Oct 2000 17:38:42 +0000 (17:38 +0000)]
Removing these scripts.  redemo.py lives on in Tools/scripts/.
regexdemo.py is obsolete with the regex module.

24 years agoI'm moving redemo.py here from Demo/tkinter/guido, since it is
Guido van Rossum [Fri, 6 Oct 2000 17:37:12 +0000 (17:37 +0000)]
I'm moving redemo.py here from Demo/tkinter/guido, since it is
somewhat useful to learn regular expressions, and this way it'll be
installed on Windows.

This closes bug report 115609.

24 years ago[ Bug #113803 ] [2.0b1 NT4.0] printing non asci char causes idle to abort
Guido van Rossum [Fri, 6 Oct 2000 16:58:26 +0000 (16:58 +0000)]
[ Bug #113803 ] [2.0b1 NT4.0] printing non asci char causes idle to abort
http://sourceforge.net/bugs/?func=detailbug&bug_id=113803&group_id=5470

Add Unicode support and error handling to AsString().  Both AsString()
and Merge() now return NULL and set a proper Python exception
condition when an error happens; Merge() and other callers of
AsString() check for errors from AsString().  Also fixed cleanup in
Merge() and Tkapp_Call() return cleanup code; the fv array was not
necessarily completely initialized, causing calls to ckfree() with
garbage arguments!

(Also reindented some lines that were longer than 80 chars and
reformatted some code that used an alien coding standard.)

24 years agoUse the cvsinfo module instead of a module stuff off in my personal
Fred Drake [Fri, 6 Oct 2000 16:37:47 +0000 (16:37 +0000)]
Use the cvsinfo module instead of a module stuff off in my personal
collection.

24 years agoSupport module to help work with checked-out CVS trees.
Fred Drake [Fri, 6 Oct 2000 16:36:48 +0000 (16:36 +0000)]
Support module to help work with checked-out CVS trees.

24 years agoDonn Cave <donn@oz.net>:
Fred Drake [Fri, 6 Oct 2000 16:17:21 +0000 (16:17 +0000)]
Donn Cave <donn@oz.net>:
New README for BeOS R5.

This closes SourceForge patch #101779.

24 years agoDonn Cave <donn@u.washington.edu>:
Fred Drake [Fri, 6 Oct 2000 16:11:21 +0000 (16:11 +0000)]
Donn Cave <donn@u.washington.edu>:
Script to regenerate platform-specific modules of constants.

[I moved common paths to variables for easier reading by humans. -- FLD]

This closes SourceForge patch #101781.

24 years agoDonn Cave <donn@u.washington.edu>:
Fred Drake [Fri, 6 Oct 2000 15:57:45 +0000 (15:57 +0000)]
Donn Cave <donn@u.washington.edu>:
Added definition of VERSION so this works as expected.

24 years agoDonn Cave <donn@u.washington.edu>:
Fred Drake [Fri, 6 Oct 2000 15:48:38 +0000 (15:48 +0000)]
Donn Cave <donn@u.washington.edu>:
Do not assume that all platforms using a MetroWorks compiler can use
POSIX threads; the assumption breaks on BeOS.  This fix only helps
for BeOS.

This closes SourceForge patch #101772.

24 years agoNorman Vine <nhv@users.sourceforge.net>:
Fred Drake [Fri, 6 Oct 2000 15:37:06 +0000 (15:37 +0000)]
Norman Vine <nhv@users.sourceforge.net>:
tcp.h is not available on CygWin, so do not try to include it there.

This closes SourceForge patch #101802.

24 years agois_zipfile() description: Use the same name for the parameter as the
Fred Drake [Fri, 6 Oct 2000 15:29:56 +0000 (15:29 +0000)]
is_zipfile() description:  Use the same name for the parameter as the
      code, in case someone wants to use it as a keyword paramter.

ZIP_DEFLATED description:  Do not reveal the specific value of the
      constant, since code should only use the symbolic name.

24 years agoAdded a name.
Fred Drake [Fri, 6 Oct 2000 15:25:01 +0000 (15:25 +0000)]
Added a name.

24 years agoAdded section on threads problems on Reliant UNIX; this relates to
Fred Drake [Fri, 6 Oct 2000 01:58:48 +0000 (01:58 +0000)]
Added section on threads problems on Reliant UNIX; this relates to
bug #113797.  We should be able to resolve this for the next release.

Reflowed the comments on Monterey (64-bit AIX) to match the flow of the
other platform-specific sections.

24 years ago[ Bug #110677 ] PRIVATE: various minor Tkinter things (PR#388)
Guido van Rossum [Fri, 6 Oct 2000 00:38:51 +0000 (00:38 +0000)]
[ Bug #110677 ] PRIVATE: various minor Tkinter things (PR#388)
http://sourceforge.net/bugs/?func=detailbug&group_id=5470&bug_id=110677

Canvas.CanvasItem & Canvas.Group:
- bind lacks an optional "add" param
- unbind lacks an optional "funcid" param
- tkraise/lower should call self.canvas.tag_XXXX

(markus.oberhumer@jk.uni-linz.ac.at)

Note: I'm *not* fixing "bbox() return value is inconsistent with
Canvas.bbox()" -- it might break existing code.

24 years agoSF bug 115831 and Ping's SF patch 101751, 0.0**-2.0 returns inf rather than
Tim Peters [Fri, 6 Oct 2000 00:36:09 +0000 (00:36 +0000)]
SF bug 115831 and Ping's SF patch 101751, 0.0**-2.0 returns inf rather than
raise ValueError.  Checked in the patch as far as it went, but also changed
all of ints, longs and floats to raise ZeroDivisionError instead when raising
0 to a negative number.  This is what 754-inspired stds require, as the "true
result" is an infinity obtained from finite operands, i.e. it's a singularity.
Also changed float pow to not be so timid about using its square-and-multiply
algorithm.  Note that what math.pow does is unrelated to what builtin pow
does, and will still vary by platform.

24 years agoRemove some debugging messages - although this code is a complete hack, we dont need...
Mark Hammond [Thu, 5 Oct 2000 22:10:33 +0000 (22:10 +0000)]
Remove some debugging messages - although this code is a complete hack, we dont need to announce it to the world every time they use freeze!

24 years agoAdd support for "import re" -- it uses pre, but user code does not need
Fred Drake [Thu, 5 Oct 2000 20:42:44 +0000 (20:42 +0000)]
Add support for "import re" -- it uses pre, but user code does not need
to.

24 years agoThe _PyTuple_Resize() last_is_sticky flag must now always be false.
Neil Schemenauer [Thu, 5 Oct 2000 19:38:24 +0000 (19:38 +0000)]
The _PyTuple_Resize() last_is_sticky flag must now always be false.

24 years agoSimplify _PyTuple_Resize by not using the tuple free list and dropping
Neil Schemenauer [Thu, 5 Oct 2000 19:36:49 +0000 (19:36 +0000)]
Simplify _PyTuple_Resize by not using the tuple free list and dropping
support for the last_is_sticky flag.  A few hard to find bugs may be
fixed by this patch since the old code was buggy.

24 years agoSF "bug" 115973: patches from Norman Vine so that shared libraries and
Tim Peters [Thu, 5 Oct 2000 19:24:26 +0000 (19:24 +0000)]
SF "bug" 115973:  patches from Norman Vine so that shared libraries and
Tkinter work under Cygwin.  Accepted on faith & reasonableness.

24 years agotranslation(): Minor optimization patch which avoids instantiating the
Barry Warsaw [Thu, 5 Oct 2000 18:48:12 +0000 (18:48 +0000)]
translation(): Minor optimization patch which avoids instantiating the
default value's instance unless it's absolutely necessary.

24 years agoChange all occurances of
Barry Warsaw [Thu, 5 Oct 2000 18:45:53 +0000 (18:45 +0000)]
Change all occurances of

test -d "$directory"

to

test ! -z "directory" -a -d "directory"

Apparently, on SunOS 4.1.4_JL (and other?) OSes, -d on an empty string
always returns true.  This closes SF bug #115392.

24 years ago[ Patch #101730 ] Add initial static support for Darwin/MacOSX.
Guido van Rossum [Thu, 5 Oct 2000 18:00:06 +0000 (18:00 +0000)]
[ Patch #101730 ] Add initial static support for Darwin/MacOSX.
By D.K. Wolfe.

24 years agoAdded Py_FPROTO macro which was available in Python 1.5.x and below.
Marc-André Lemburg [Thu, 5 Oct 2000 17:25:45 +0000 (17:25 +0000)]
Added Py_FPROTO macro which was available in Python 1.5.x and below.
This should not be used for new code, but will probably make porting
old extensions to 2.0 a lot easier.

Also see Bug #116011.

24 years agoFix Bug #115907: encode '=' as '=3D' and not '=='
Jeremy Hylton [Thu, 5 Oct 2000 17:24:33 +0000 (17:24 +0000)]
Fix Bug #115907: encode '=' as '=3D' and not '=='

24 years agoDocument the lookbehind assertions (closing bug#115119)
Andrew M. Kuchling [Thu, 5 Oct 2000 15:22:28 +0000 (15:22 +0000)]
Document the lookbehind assertions (closing bug#115119)

24 years agoFix for SF bug #115987: PyInstance_HalfBinOp does not initialize the
Thomas Wouters [Thu, 5 Oct 2000 12:43:25 +0000 (12:43 +0000)]
Fix for SF bug #115987: PyInstance_HalfBinOp does not initialize the
result-object-pointer that is passed in, when an exception occurs during
coercion. The pointer has to be explicitly initialized in the caller to avoid
putting trash on the Python stack.

24 years agoDocument expand() method of MatchObjects
Andrew M. Kuchling [Thu, 5 Oct 2000 12:35:29 +0000 (12:35 +0000)]
Document expand() method of MatchObjects

24 years agoDetect conflicting Python DLL on module import under Windows - as per [ Patch #101676 ]
Mark Hammond [Thu, 5 Oct 2000 10:54:45 +0000 (10:54 +0000)]
Detect conflicting Python DLL on module import under Windows - as per [ Patch #101676 ]

24 years agoBetter add at least a short docstring to the module!
Fred Drake [Thu, 5 Oct 2000 05:20:55 +0000 (05:20 +0000)]
Better add at least a short docstring to the module!

24 years agoAdd a link to the new acks.html file.
Fred Drake [Thu, 5 Oct 2000 05:17:29 +0000 (05:17 +0000)]
Add a link to the new acks.html file.

24 years agoAdd rules for generating the acks.html file at the top of the document tree.
Fred Drake [Thu, 5 Oct 2000 05:16:56 +0000 (05:16 +0000)]
Add rules for generating the acks.html file at the top of the document tree.

24 years agoIgnore the acks.html file, since it is generated.
Fred Drake [Thu, 5 Oct 2000 05:16:12 +0000 (05:16 +0000)]
Ignore the acks.html file, since it is generated.

24 years agoNew script to convert the ACKS file to a nicely formatted HTML file.
Fred Drake [Thu, 5 Oct 2000 05:15:29 +0000 (05:15 +0000)]
New script to convert the ACKS file to a nicely formatted HTML file.
Uses the new support module.

24 years agoUse the new support module instead of including all the getopt
Fred Drake [Thu, 5 Oct 2000 05:14:26 +0000 (05:14 +0000)]
Use the new support module instead of including all the getopt
processing and style information directly.

24 years agoFactor out some of the presentation and shared code from mkmodindex and
Fred Drake [Thu, 5 Oct 2000 05:11:57 +0000 (05:11 +0000)]
Factor out some of the presentation and shared code from mkmodindex and
a mkackshtml (not yet checked in).

24 years agoList of acknowledgements culled from CVS checkin logs and memory.
Fred Drake [Thu, 5 Oct 2000 04:05:30 +0000 (04:05 +0000)]
List of acknowledgements culled from CVS checkin logs and memory.
The documentation ACKS list will be maintained separately from the
CPython ACKS list.

24 years agoNew tool for normalizing indentation of .py files.
Tim Peters [Thu, 5 Oct 2000 03:48:38 +0000 (03:48 +0000)]
New tool for normalizing indentation of .py files.

24 years agoUpdate for next planned release.
Fred Drake [Thu, 5 Oct 2000 03:24:00 +0000 (03:24 +0000)]
Update for next planned release.

24 years agoMove LONG_BIT from intobject.c to pyport.h. #error if it's already been
Tim Peters [Thu, 5 Oct 2000 01:42:25 +0000 (01:42 +0000)]
Move LONG_BIT from intobject.c to pyport.h.  #error if it's already been
#define'd to an unreasonable value (several recent gcc systems have
misdefined it, causing bogus overflows in integer multiplication).  Nuke
CHAR_BIT entirely.

24 years agoEnable the binascii module for Win64. It builds and passes the test suite.
Trent Mick [Wed, 4 Oct 2000 20:57:29 +0000 (20:57 +0000)]
Enable the binascii module for Win64. It builds and passes the test suite.
(I had explicitly disabled it a while ago, possibly unecessarily, along with
rgbimg, audioop, and imageop, which are advertised as "not for 64-bit
platforms.)

24 years agoFix for test_class.py on Win64. id(self), which on Win64 returns a
Trent Mick [Wed, 4 Oct 2000 17:50:59 +0000 (17:50 +0000)]
Fix for test_class.py on Win64. id(self), which on Win64 returns a
PyLong, was used for the return value of a class __hash__ method, which
*must* return a PyInt. Solution: hash() the id(self) value.

24 years ago- update Neil's email address
Neil Schemenauer [Wed, 4 Oct 2000 16:34:09 +0000 (16:34 +0000)]
- update Neil's email address

24 years ago- do not start collection during processing of an exception
Neil Schemenauer [Wed, 4 Oct 2000 16:25:07 +0000 (16:25 +0000)]
- do not start collection during processing of an exception

24 years ago- Fix a GC bug caused by PyDict_New() failing.
Neil Schemenauer [Wed, 4 Oct 2000 16:22:26 +0000 (16:22 +0000)]
- Fix a GC bug caused by PyDict_New() failing.

24 years ago- fix a GC bug caused by malloc() failing
Neil Schemenauer [Wed, 4 Oct 2000 16:20:41 +0000 (16:20 +0000)]
- fix a GC bug caused by malloc() failing

24 years agoAdded deprecation notices to atof(), atoi(), and atol(), recommending
Fred Drake [Wed, 4 Oct 2000 13:59:52 +0000 (13:59 +0000)]
Added deprecation notices to atof(), atoi(), and atol(), recommending
use of the float(), int(), and long() built-ins instead.

Fixed minor markup nit elsewhere (use of \optional).

24 years agoAdded \versionchanged comment to os.popen() noting that it was unreliable
Fred Drake [Wed, 4 Oct 2000 13:57:27 +0000 (13:57 +0000)]
Added \versionchanged comment to os.popen() noting that it was unreliable
in earlier versions of Python; this is useful information for people
interested in writing code that is portable across Python versions.

Suggested by Peter Funk <pf@artcom-gmbh.de>.

24 years agoRemove styling information that conflicts with or duplicates the stylesheet;
Fred Drake [Wed, 4 Oct 2000 13:39:24 +0000 (13:39 +0000)]
Remove styling information that conflicts with or duplicates the stylesheet;
different browsers resolve the conflicts differently, and the "proper"
resolution is not what we actually want.

Reported by Peter Funk <pf@artcom-gmbh.de>.

24 years agoRewrites to section on new development process, after Usenet discussion
Andrew M. Kuchling [Wed, 4 Oct 2000 12:40:44 +0000 (12:40 +0000)]
Rewrites to section on new development process, after Usenet discussion
    of the text

24 years agoUse \obindex{...} instead of \indexii{...}{type} in many places; this is
Fred Drake [Wed, 4 Oct 2000 04:21:19 +0000 (04:21 +0000)]
Use \obindex{...} instead of \indexii{...}{type} in many places; this is
more consistent with other index entries in the documentation.

24 years agotest_popen2 broke on Windows shortly after 2.0b2 was released. Fixed it.
Tim Peters [Tue, 3 Oct 2000 23:07:13 +0000 (23:07 +0000)]
test_popen2 broke on Windows shortly after 2.0b2 was released.  Fixed it.

24 years agoSupport non-namespace elements in *ElementNS of XMLGenerator.
Martin v. Löwis [Tue, 3 Oct 2000 22:35:29 +0000 (22:35 +0000)]
Support non-namespace elements in *ElementNS of XMLGenerator.

24 years agoPreliminary tool to troll through the CVS logs and LaTeX sources for
Fred Drake [Tue, 3 Oct 2000 22:10:25 +0000 (22:10 +0000)]
Preliminary tool to troll through the CVS logs and LaTeX sources for
the names of people that should be in the ACKS file.

This relies on some personal code that is not yet available, but should
be by the time we release 2.0c1.

24 years ago_PyUnicode_Fini(): Initialize the local freelist walking variable `u'
Barry Warsaw [Tue, 3 Oct 2000 20:45:26 +0000 (20:45 +0000)]
_PyUnicode_Fini(): Initialize the local freelist walking variable `u'
after unicode_empty has been freed, otherwise it might not point to
the real start of the unicode_freelist.  Final closure for SF bug
#110681, Jitterbug PR#398.

24 years agoAccept keyword arguments for (most) pattern and match object
Fredrik Lundh [Tue, 3 Oct 2000 20:43:34 +0000 (20:43 +0000)]
Accept keyword arguments for (most) pattern and match object
methods.  Closes buglet #115845.

24 years agoRecompile pattern if (?x) flag was found inside the pattern during the
Fredrik Lundh [Tue, 3 Oct 2000 19:22:26 +0000 (19:22 +0000)]
Recompile pattern if (?x) flag was found inside the pattern during the
first scan.  Closes bug #115040.

24 years agoIn _PyUnicode_Fini(), decref unicode_empty before tearng down the free
Guido van Rossum [Tue, 3 Oct 2000 18:09:04 +0000 (18:09 +0000)]
In _PyUnicode_Fini(), decref unicode_empty before tearng down the free
list.  Discovered by Barry, fix approved by MAL.

24 years agoRemove old note that os.popen() on Windows is unreliable; this is no
Fred Drake [Tue, 3 Oct 2000 17:14:27 +0000 (17:14 +0000)]
Remove old note that os.popen() on Windows is unreliable; this is no
longer true.

24 years agoFix a few problems with the _Printer class and the license variable.
Guido van Rossum [Tue, 3 Oct 2000 17:11:37 +0000 (17:11 +0000)]
Fix a few problems with the _Printer class and the license variable.

1. repr(license) will no longer print to stdout and read from stdin;
you have to use license().  `license` is a short message explaining
this.

2. Use lazy initialization so that startup isn't slowed down by the
search for the LICENSE file.

3. repr(license) actually returns the desired string, rather than
printing to stdout and returning ''.  (Why didn't we think of this
before?)

4. Use the pythonlabs license URL as the license fallback instead of
the CNRI license handle.

24 years agoClarify that isatty is only guaranteed to return true for slave ends of
Thomas Wouters [Tue, 3 Oct 2000 16:54:24 +0000 (16:54 +0000)]
Clarify that isatty is only guaranteed to return true for slave ends of
terminals, not the master end (though it does, on most systems.)

24 years agoOn some systems (like Solaris), the master end of a tty/pty pair is
Thomas Wouters [Tue, 3 Oct 2000 16:51:08 +0000 (16:51 +0000)]
On some systems (like Solaris), the master end of a tty/pty pair is
apparently not considered a terminal, and so isatty(3) returns false. So we
skip the test for ttyness of the master side and just check the slave side,
which should really be a terminal.

24 years agoBump the release number for the upcoming candidate.
Fred Drake [Tue, 3 Oct 2000 16:50:48 +0000 (16:50 +0000)]
Bump the release number for the upcoming candidate.

24 years agoFixed negative lookahead/lookbehind. Closes bug #115618.
Fredrik Lundh [Tue, 3 Oct 2000 16:29:23 +0000 (16:29 +0000)]
Fixed negative lookahead/lookbehind.  Closes bug #115618.

24 years ago_PyImport_Fini(): Closed small memory leak when an embedded app calls
Barry Warsaw [Tue, 3 Oct 2000 16:02:05 +0000 (16:02 +0000)]
_PyImport_Fini(): Closed small memory leak when an embedded app calls
Py_Initialize()/Py_Finalize() in a loop.  _PyImport_Filetab needed to
be deallocated.  Partial closure of SF #110681, Jitterbug PR#398.

24 years agoJim Ahlstrom sent a few corrections to my changes. (Thanks!)
Fred Drake [Tue, 3 Oct 2000 15:16:31 +0000 (15:16 +0000)]
Jim Ahlstrom sent a few corrections to my changes.  (Thanks!)

24 years agoUndo Ping's change.
Guido van Rossum [Tue, 3 Oct 2000 13:51:09 +0000 (13:51 +0000)]
Undo Ping's change.

CGI scripts should *not* use /usr/bin/env, since on systems that don't
come standard with Python installed, Python isn't on the default $PATH.

Too bad that this breaks on Linux, where Python is in /usr/bin which
is on the default path -- the point is that you must manually edit
your CGI scripts when you install them.

24 years agoChange first line to #!/usr/bin/env python (really just to test check-in).
Ka-Ping Yee [Tue, 3 Oct 2000 08:32:00 +0000 (08:32 +0000)]
Change first line to #!/usr/bin/env python (really just to test check-in).

24 years ago$OFF_SITE_LINK_ICON: Don't define here; simply defining it overrides a
Fred Drake [Tue, 3 Oct 2000 06:05:25 +0000 (06:05 +0000)]
$OFF_SITE_LINK_ICON:  Don't define here; simply defining it overrides a
        definition provided by previously loaded configuration code, and
        testing whether it's defined isn't needed since the default was
        false anyway.

get_link_icon():  Add support for $OFF_SITE_LINK_ICON_HEIGHT and
        $OFF_SITE_LINK_ICON_WIDTH, giving the dimensions of the icon
        being used.  This can make for faster page display.  Both are
        optional.

make_my_titlegraphic():  Fix insertion of the off-site icon link.

do_env_funcdesc():  Remove debugging print.

handle_rfclike_reference():  Remove trailing colon from first line; it
        doesn't really make sense and looks bad if we add an icon to
        mark off-site links.