]> granicus.if.org Git - python/log
python
23 years ago- Don't return mac-style pathnames in unix-Python.
Jack Jansen [Wed, 8 Aug 2001 13:46:49 +0000 (13:46 +0000)]
- Don't return mac-style pathnames in unix-Python.
- Fixed up a lot more prototypes (gcc also wants them on static routines)
- Fixed various other gcc warnings.

23 years agoRemoved special rule for Mac/Python/macglue.o, it is not needed anymore.
Jack Jansen [Wed, 8 Aug 2001 13:18:36 +0000 (13:18 +0000)]
Removed special rule for Mac/Python/macglue.o, it is not needed anymore.

23 years agoSplit macglue.c into two: a new mactoolboxglue.c (in ./Python)
Jack Jansen [Wed, 8 Aug 2001 13:17:31 +0000 (13:17 +0000)]
Split macglue.c into two: a new mactoolboxglue.c (in ./Python)
with functionality needed for both unix-Python and MacPython and a
new smaller ./Mac/Python/macglue.c which contains MacPython stuff only.

pymactoolbox.h has moved to ./Include from ./Mac/Include and now also
contains the relevant stuff from macglue.h.

The net effect of this is that the ./Mac subdirectory is not needed
anymore for building the unix-Python core on MacOSX (it is needed
for building the extension modules).

23 years agoPut conditional S_IFMT definition into pyport.h.
Martin v. Löwis [Wed, 8 Aug 2001 12:54:28 +0000 (12:54 +0000)]
Put conditional S_IFMT definition into pyport.h.

23 years agoRemove redundant check for 'getaddrinfo' (it's already checked the block
Thomas Wouters [Wed, 8 Aug 2001 10:39:03 +0000 (10:39 +0000)]
Remove redundant check for 'getaddrinfo' (it's already checked the block
before.)

23 years agoPut conditional S_ISDIR definition(s) into pyport.h.
Martin v. Löwis [Wed, 8 Aug 2001 10:28:06 +0000 (10:28 +0000)]
Put conditional S_ISDIR definition(s) into pyport.h.

23 years agoMerged in bugfix from PyUnit CVS for problem reported by Gary Todd.
Steve Purcell [Wed, 8 Aug 2001 07:57:26 +0000 (07:57 +0000)]
Merged in bugfix from PyUnit CVS for problem reported by Gary Todd.
If 'unittest.py' was run from the command line with the name of a test
case class as a parameter, it failed with an ugly error. (Which was a
shame, because the documentation says you can do that.)

The problem was the old 'is the class X that you imported from me the same
as my class X?' gotcha.

23 years agoRegenerated token.py to account for new DOUBLESLASH and DOUBLESLASHEQUAL.
Tim Peters [Wed, 8 Aug 2001 06:35:56 +0000 (06:35 +0000)]
Regenerated token.py to account for new DOUBLESLASH and DOUBLESLASHEQUAL.

23 years agoRepair the Windows build (S_ISDIR() macro doesn't exist).
Tim Peters [Wed, 8 Aug 2001 06:24:48 +0000 (06:24 +0000)]
Repair the Windows build (S_ISDIR() macro doesn't exist).
Somebody else should feel free to repair this a different way; see Python-
Dev for discussion.

23 years agoAdd option to push the development docs to SF without announcing, for
Fred Drake [Wed, 8 Aug 2001 05:41:01 +0000 (05:41 +0000)]
Add option to push the development docs to SF without announcing, for
when the changes are just too small.

23 years agoMore names....
Fred Drake [Wed, 8 Aug 2001 05:39:52 +0000 (05:39 +0000)]
More names....

23 years agoClean up some of the markup here -- be more consistent in the use of
Fred Drake [Wed, 8 Aug 2001 05:39:29 +0000 (05:39 +0000)]
Clean up some of the markup here -- be more consistent in the use of
semantic labels instead of presentational markup.

23 years ago"Thin" version of floor division docs: add // to the list of operators
Fred Drake [Wed, 8 Aug 2001 05:37:21 +0000 (05:37 +0000)]
"Thin" version of floor division docs:  add // to the list of operators
and //= to the list of other delimiter tokens.

I'll work on it again when it's not so late...

23 years agoPatch #448227: Raise an exception when a directory is passed to execfile.
Martin v. Löwis [Wed, 8 Aug 2001 05:30:36 +0000 (05:30 +0000)]
Patch #448227: Raise an exception when a directory is passed to execfile.

23 years agoAdd new tokens // and //=, in support of PEP 238.
Guido van Rossum [Wed, 8 Aug 2001 05:04:07 +0000 (05:04 +0000)]
Add new tokens // and //=, in support of PEP 238.

23 years agoImplement PEP 238 in its (almost) full glory.
Guido van Rossum [Wed, 8 Aug 2001 05:00:18 +0000 (05:00 +0000)]
Implement PEP 238 in its (almost) full glory.

This introduces:

- A new operator // that means floor division (the kind of division
  where 1/2 is 0).

- The "future division" statement ("from __future__ import division)
  which changes the meaning of the / operator to implement "true
  division" (where 1/2 is 0.5).

- New overloadable operators __truediv__ and __floordiv__.

- New slots in the PyNumberMethods struct for true and floor division,
  new abstract APIs for them, new opcodes, and so on.

I emphasize that without the future division statement, the semantics
of / will remain unchanged until Python 3.0.

Not yet implemented are warnings (default off) when / is used with int
or long arguments.

This has been on display since 7/31 as SF patch #443474.

Flames to /dev/null.

23 years agobeginning of work on the conf. handling smarts
Steven M. Gava [Wed, 8 Aug 2001 01:30:38 +0000 (01:30 +0000)]
beginning of work on the conf. handling smarts

23 years agoCommented out most of the MacPython modules. There is a discussion on the
Jack Jansen [Wed, 8 Aug 2001 00:36:53 +0000 (00:36 +0000)]
Commented out most of the MacPython modules. There is a discussion on the
pythonmac-sig about turning this all into a package, so in the mean time
there is no reason to scribble all over people's disks. Interested parties
can uncomment them.

23 years agoPart of SF patch [#431848] mathmodule.c: doc strings & conversion, from
Tim Peters [Tue, 7 Aug 2001 22:10:00 +0000 (22:10 +0000)]
Part of SF patch [#431848] mathmodule.c: doc strings & conversion, from
Peter Schneider-Kamp.
Clarified some docstrings in the spirit of the patch; left out the
degrees() and radians() functions (see the patch comments on SF).

23 years agoFix SF bug [ #447370 ] typo in urllib2.py
Jeremy Hylton [Tue, 7 Aug 2001 21:12:25 +0000 (21:12 +0000)]
Fix SF bug [ #447370 ] typo in urllib2.py

Also fix another bug caught by pychecker-- HTTPError() raised when
redirect limit exceed did not pass an fp object.  Had to change method
to keep fp object around until it's certain that the error won't be
raised.

Remove useless line in do_proxy().

23 years agoApply two small changes to the Windows code, according to SF bug
Guido van Rossum [Tue, 7 Aug 2001 19:55:10 +0000 (19:55 +0000)]
Apply two small changes to the Windows code, according to SF bug
#427345.  These are supposed to support binary data and avoid
buffering problems on Windows.

23 years agoRemove make_re() function; this is no longer needed since _sre and pcre
Fred Drake [Tue, 7 Aug 2001 19:49:15 +0000 (19:49 +0000)]
Remove make_re() function; this is no longer needed since _sre and pcre
are now allowed by ok_builtin_modules.  This effectively backs out
revision 1.26.

This closes SF bug #448546.

23 years agoAdd a test that xml.sax.saxutils.XMLGenerator does the right thing
Fred Drake [Tue, 7 Aug 2001 19:17:06 +0000 (19:17 +0000)]
Add a test that xml.sax.saxutils.XMLGenerator does the right thing
when quoting attribute values that contain single & double quotes.

This provides the rest of the regression test for SF bug #440351.

23 years agoMake sure XMLGenerator uses quoteattr() instead of escape() to quote
Fred Drake [Tue, 7 Aug 2001 19:14:46 +0000 (19:14 +0000)]
Make sure XMLGenerator uses quoteattr() instead of escape() to quote
attribute values.  Just using escape() can (and always has) led to broken
XML being generated.  This makes sure it always produces the right thing.

This actually closes SF bug #440351.

23 years agoFix the test so it uses IterableUserDict for the "for x in dict" test.
Guido van Rossum [Tue, 7 Aug 2001 17:50:06 +0000 (17:50 +0000)]
Fix the test so it uses IterableUserDict for the "for x in dict" test.

23 years agoRemove the __iter__ method from the UserDict class -- it can silently
Guido van Rossum [Tue, 7 Aug 2001 17:40:42 +0000 (17:40 +0000)]
Remove the __iter__ method from the UserDict class -- it can silently
break old code (in extreme cases).  See SF bug #448153.

Add a new subclass IterableUserDict that has the __iter__ method.

Note that for new projects, unless backwards compatibility with
pre-2.2 Python is required, subclassing 'dictionary' is recommended;
UserDict might become deprecated.

23 years ago- Rename PyType_InitDict() to PyType_Ready().
Guido van Rossum [Tue, 7 Aug 2001 17:24:28 +0000 (17:24 +0000)]
- Rename PyType_InitDict() to PyType_Ready().

- Add an explicit call to PyType_Ready(&PyList_Type) to pythonrun.c
  (just for the heck of it, really -- we should either explicitly
  ready all types, or none).

23 years agoAdd the NannyNag exception class and the process_tokens() function to
Guido van Rossum [Tue, 7 Aug 2001 17:19:25 +0000 (17:19 +0000)]
Add the NannyNag exception class and the process_tokens() function to
__all__, to indicate these are implied as part of the public API.

IDLE's "Check Module" command uses this, and it broke once already
because the reset_globals() and tokeneater() functions were deleted
when Neil converted this to using the generator API of tokenizer.
(See SF bug #448835.)

23 years agoFix for SF bug #448835.
Guido van Rossum [Tue, 7 Aug 2001 17:15:36 +0000 (17:15 +0000)]
Fix for SF bug #448835.

Fix this to work with the new (still undocumented) tabnanny API.

I'm afraid Stephen will have to add this fix to the IDLE fork code
base as well.

23 years agoRemove the test for the 'object' type -- it can no longer be
Guido van Rossum [Tue, 7 Aug 2001 16:53:42 +0000 (16:53 +0000)]
Remove the test for the 'object' type -- it can no longer be
instantiated.  (Its use as a base class is adequately tested by other
tests.)

23 years agoCosmetics:
Guido van Rossum [Tue, 7 Aug 2001 16:40:56 +0000 (16:40 +0000)]
Cosmetics:

- Add comment blocks explaining add_operators() and override_slots().
  (This file could use some more explaining, but this is all I had
  breath for today. :)

- Renamed the argument 'base' of add_wrappers() to 'wraps' because
  it's not a base class (which is what the 'base' identifier is used
  for elsewhere).

Small nits:

- Fix add_tp_new_wrapper() to avoid overwriting an existing __new__
  descriptor in tp_defined.

- In add_operators(), check the return value of add_tp_new_wrapper().

Functional change:

- Remove the tp_new functionality from PyBaseObject_Type; this means
  you can no longer instantiate the 'object' type.  It's only useful
  as a base class.

- To make up for the above loss, add tp_new to dynamic types.  This
  has to be done in a hackish way (after override_slots() has been
  called, with an explicit call to add_tp_new_wrapper() at the very
  end) because otherwise I ran into recursive calls of slot_tp_new().
  Sigh.

23 years agosilence warnings about import *
Jeremy Hylton [Tue, 7 Aug 2001 16:38:19 +0000 (16:38 +0000)]
silence warnings about import *

23 years agoBrought up-to-date.
Jack Jansen [Tue, 7 Aug 2001 15:34:47 +0000 (15:34 +0000)]
Brought up-to-date.

23 years agoGot rid of (hopefully) the last 68k-mac related ifdefs.
Jack Jansen [Tue, 7 Aug 2001 15:32:37 +0000 (15:32 +0000)]
Got rid of (hopefully) the last 68k-mac related ifdefs.

23 years agoGot rid of 68k-Mac and other outdated ifdefs.
Jack Jansen [Tue, 7 Aug 2001 15:30:23 +0000 (15:30 +0000)]
Got rid of 68k-Mac and other outdated ifdefs.

23 years agoGot rid of mactcp cruft, added Mlte and CF modules.
Jack Jansen [Tue, 7 Aug 2001 15:29:17 +0000 (15:29 +0000)]
Got rid of mactcp cruft, added Mlte and CF modules.

23 years agoRemoved unused variable.
Jack Jansen [Tue, 7 Aug 2001 15:28:52 +0000 (15:28 +0000)]
Removed unused variable.

23 years agoYet another project.
Jack Jansen [Tue, 7 Aug 2001 15:26:26 +0000 (15:26 +0000)]
Yet another project.

23 years agoGot rid of (hopefully) the last 68k Mac-related ifdefs and various other cruft.
Jack Jansen [Tue, 7 Aug 2001 15:14:53 +0000 (15:14 +0000)]
Got rid of (hopefully) the last 68k Mac-related ifdefs and various other cruft.

23 years agoRemoved outdated stuff.
Jack Jansen [Tue, 7 Aug 2001 15:12:04 +0000 (15:12 +0000)]
Removed outdated stuff.

23 years agoDocument two new items
Andrew M. Kuchling [Tue, 7 Aug 2001 14:28:58 +0000 (14:28 +0000)]
Document two new items
Correct error noticed by Keith Briggs
Re-indent a paragraph

23 years agoProject files for the new plugin modules, plus a few that hadn't been added yet but...
Jack Jansen [Tue, 7 Aug 2001 14:21:56 +0000 (14:21 +0000)]
Project files for the new plugin modules, plus a few that hadn't been added yet but should have been.

23 years agoBit another bullet: all toolbox modules are now in dynamically loaded modules. Everyt...
Jack Jansen [Tue, 7 Aug 2001 13:53:25 +0000 (13:53 +0000)]
Bit another bullet: all toolbox modules are now in dynamically loaded modules. Everything still seems to work, but the ConfigurePythonXXX on initial install may still need work.

23 years agoBit the bullet and enabled garbage collection (finally).
Jack Jansen [Tue, 7 Aug 2001 12:33:32 +0000 (12:33 +0000)]
Bit the bullet and enabled garbage collection (finally).
Also updated pyconfig.h to the current state of pyconfig.h.in.

23 years agomore fleshing out, this time primarily the 'general' page
Steven M. Gava [Tue, 7 Aug 2001 03:28:25 +0000 (03:28 +0000)]
more fleshing out, this time primarily the 'general' page

23 years agoAnother bug fix for recent import * warning (caught by Thomas Wouters)
Jeremy Hylton [Mon, 6 Aug 2001 20:34:25 +0000 (20:34 +0000)]
Another bug fix for recent import * warning (caught by Thomas Wouters)

Only return if symtable_warn() returns -1, indicating that the warning
was turned into an error.

23 years ago+ Turn off patch creation (this isn't a patch release! got turned on by
Tim Peters [Mon, 6 Aug 2001 20:01:19 +0000 (20:01 +0000)]
+ Turn off patch creation (this isn't a patch release!  got turned on by
  mistake).
+ Arrange for Win2K Add/Remove to show a Python icon.

I think this "does it" -- a full install/uninstall can now be done on a
Win2K box from an ordinary (not Admin, not Power User) user acct, incl.
file extension registration, Start Menu entries, and full Add/Remove.

23 years agoFix error message for import * in function/class scope
Jeremy Hylton [Mon, 6 Aug 2001 19:55:17 +0000 (19:55 +0000)]
Fix error message for import * in function/class scope

23 years agoFix SF bug [ #445474 ] warn about import * inside functions
Jeremy Hylton [Mon, 6 Aug 2001 19:45:40 +0000 (19:45 +0000)]
Fix SF bug [ #445474 ] warn about import * inside functions

Reported by the Man himself.

23 years agoTabify the file (pointed out by Thomas Wouters)
Andrew M. Kuchling [Mon, 6 Aug 2001 19:33:52 +0000 (19:33 +0000)]
Tabify the file (pointed out by Thomas Wouters)

23 years agoUpdate listing of subdirectories
Andrew M. Kuchling [Mon, 6 Aug 2001 18:55:22 +0000 (18:55 +0000)]
Update listing of subdirectories
Update link to a running instance of the FAQwizard

23 years agoRemove spurious "closed" attribute definition from the memberlist
Guido van Rossum [Mon, 6 Aug 2001 18:51:38 +0000 (18:51 +0000)]
Remove spurious "closed" attribute definition from the memberlist
table.  (reported as an aside in SF #446049).

23 years agoRemove various outdated files. (Leaving find_recursionlimit.py alone,
Andrew M. Kuchling [Mon, 6 Aug 2001 18:44:56 +0000 (18:44 +0000)]
Remove various outdated files.  (Leaving find_recursionlimit.py alone,
   as Neil pointed out it isn't the same as sys.getrecursionlimit)

23 years agoUpdate cheatsheet to 2.0 from Brunning/Gruet's quick reference
Andrew M. Kuchling [Mon, 6 Aug 2001 17:43:49 +0000 (17:43 +0000)]
Update cheatsheet to 2.0 from Brunning/Gruet's quick reference

23 years agoAdd 'yield' as a keyword
Andrew M. Kuchling [Mon, 6 Aug 2001 17:42:53 +0000 (17:42 +0000)]
Add 'yield' as a keyword
Fix typo in comment

23 years agoA totally new way to do the __new__ wrapper. This should address the
Guido van Rossum [Mon, 6 Aug 2001 16:50:37 +0000 (16:50 +0000)]
A totally new way to do the __new__ wrapper.  This should address the
problem brought up in SF bug #444229.

23 years agoOops, arg type for the glue routine was wrong.
Jack Jansen [Mon, 6 Aug 2001 15:32:30 +0000 (15:32 +0000)]
Oops, arg type for the glue routine was wrong.

23 years agoShow the MacPython IDE version in the interactive banner.
Just van Rossum [Mon, 6 Aug 2001 11:12:18 +0000 (11:12 +0000)]
Show the MacPython IDE version in the interactive banner.

23 years agoOops, this file was apparently never added to the repository. Now it is.
Jack Jansen [Mon, 6 Aug 2001 10:08:56 +0000 (10:08 +0000)]
Oops, this file was apparently never added to the repository. Now it is.

23 years agoReplace moddir and incdir by
Jack Jansen [Sun, 5 Aug 2001 22:31:19 +0000 (22:31 +0000)]
Replace moddir and incdir by
moddirlist and incdirlist, lists of source and include
directories that are searched for modules.
This is needed because the Mac modules and include files
live in the Mac subtree.

In addition (and that's actually what the mod is all about) on OSX we build all the Mac extension modules.

23 years agoPatch #448194: Debuging negative reference counts.
Martin v. Löwis [Sun, 5 Aug 2001 21:23:03 +0000 (21:23 +0000)]
Patch #448194: Debuging negative reference counts.

23 years agoAdd a comment on time.time() returning non-decreasing values except when
Fred Drake [Sun, 5 Aug 2001 15:43:04 +0000 (15:43 +0000)]
Add a comment on time.time() returning non-decreasing values except when
the clock is set back.

This closes SF bug #447945.

23 years ago+ Made installer more reluctant to overwrite MS C runtime DLLs -- it's
Tim Peters [Sun, 5 Aug 2001 13:58:48 +0000 (13:58 +0000)]
+ Made installer more reluctant to overwrite MS C runtime DLLs -- it's
  been overwriting them even if they have the same version, not just if
  they're an older version (and our installers have always done this).
+ Added an "Advanced Options" subdialog to "Select Components".  Allows
  to do a non-admin install even if you have Administrator rights, and
  to skip registering file extensions and/or creating Start Menu
  shortcuts.  Since so far these installers have been tested only by me,
  and Win2K has been full of surprises, I want those options available
  out in the field.
Lots of web searching turned up what should have been obvious:  Because
Windows Installer is a native Win2K service, it can run at a higher
privilege level than the user invoking it.  So MSI installs don't bash
into these permission gotchas on Win2K, but Wise 8.1 does (it's just
another app to Win2K, and we're not alone in wrestling with this; but,
like changing int division in Python, Win2K is doing a right thing <wink>).

23 years agofix for Bug ID 448100 - "test code using NL instead of CRNL"
Piers Lauder [Sun, 5 Aug 2001 10:43:03 +0000 (10:43 +0000)]
fix for Bug ID 448100 - "test code using NL instead of CRNL"

23 years agobegan font / indentation page
Steven M. Gava [Sun, 5 Aug 2001 08:00:28 +0000 (08:00 +0000)]
began font / indentation page

23 years agoRename pydoc.pyw to pydocgui.pyw: Since we changed Python to allow .pyw
Tim Peters [Sun, 5 Aug 2001 05:43:20 +0000 (05:43 +0000)]
Rename pydoc.pyw to pydocgui.pyw:  Since we changed Python to allow .pyw
files to satisfy imports, pydoc.pyw was just importing itself!

23 years ago+ Redid file associations on Win2K in what I believe is "the right way".
Tim Peters [Sun, 5 Aug 2001 05:19:15 +0000 (05:19 +0000)]
+ Redid file associations on Win2K in what I believe is "the right way".
  Can't test it until getting to a Win2K box, because the non-Admin way
  of setting file associations on Win2K doesn't work on any other flavor
  of Windows (and other flavors of Windows never need Admin privs to
  do it the old way).
+ Consequently got rid of the "Register file associations" Component and
  associated GUI.
+ Added a line to the summary saying whether or not this is an Admin-level
  install (I fear that will be an important clue someday).
+ Minor fiddling to the summary to reduce the # of lines.  Added a
  horizontal scrollbar in case the install path is very long.
+ Reworked the way the Main and Tools components share pydoc.pyw; cleaner
  and simpler.

23 years agoNew runtime vrbl DOADMIN. When false, we never try to write under HKLM,
Tim Peters [Sun, 5 Aug 2001 04:12:30 +0000 (04:12 +0000)]
New runtime vrbl DOADMIN.  When false, we never try to write under HKLM,
and install the Python and MS runtime DLLs into the Python dir instead of
a system dir.
Initial value is taken from new compiler vrbl _DOADMIN_ (default true),
and forced to false if the user doesn't have admin privs.
This makes it possible to *test* non-admin installs on machines where the
distinction doesn't exist (like my home box), via just changing _DOADMIN_.
It may also be useful for users who don't *want* an installer to
scribble into their system dir (for example, me(! most days)), but that
would require adding more GUI to let them get at it.

23 years ago+ Add version info to Win2K add/remove.
Tim Peters [Sat, 4 Aug 2001 23:23:27 +0000 (23:23 +0000)]
+ Add version info to Win2K add/remove.
+ Fiddle vrbls so Win2K add/remove can display version w/o future manual
  script fiddling.
+ Break apart the mysterious wizard-generated Win2K "Edit 3 Registry Keys"
  script items by hand into 3 separate items, so you can see what the heck
  they're doing in the script view.
+ pydoc.pyw was a problem:  it's installed by both the Main and Tools
  components.  So when both were selected, the second time it got
  installed Wise figured it was overwriting a pre-existing version, and
  made a backup copy in BACKUP.  A rollback-uninstall then restored that,
  leaving the Tools/Scripts/ directory non-empty, and so Wise couldn't
  remove that directory (or any above it).  Fixed by installing pydoc.pyw
  at most once.
+ Rearranged and commented the "register file extensions" section, because
  it was confusing and needs more work:  turns out it's not true that
  Win2K requires Admin privs to register file extensions, BUT, if you
  don't have Admin privs, Win2K requires a new way to register file
  extensions, and a way that doesn't blow up but doesn't do any good either
  on earlier Windows flavors.  I think I know how to get this done, but am
  too depressed to do it right now <0.7 wink>.

23 years agoAdd more constants. Contributed by itojun.
Martin v. Löwis [Sat, 4 Aug 2001 22:37:23 +0000 (22:37 +0000)]
Add more constants. Contributed by itojun.

23 years agoAuto-detect hstrerror. Raise socket.herror in PyH_Error. Register the three
Martin v. Löwis [Sat, 4 Aug 2001 22:32:03 +0000 (22:32 +0000)]
Auto-detect hstrerror. Raise socket.herror in PyH_Error. Register the three
exception classes in the module dictionary.

23 years agoDocument IPv6 changes. Contributed by itojun.
Martin v. Löwis [Sat, 4 Aug 2001 22:22:45 +0000 (22:22 +0000)]
Document IPv6 changes. Contributed by itojun.

23 years agoHack to workaround bug #445984: attempt to import _socket before
Greg Ward [Sat, 4 Aug 2001 22:22:19 +0000 (22:22 +0000)]
Hack to workaround bug #445984: attempt to import _socket before
starting the test suite proper.  If _socket fails to build, that will
make this test fail with an ImportError -- handled by the test harness
as "no such module _socket" -- instead of an AttributeError deep in
CGIHTTPServer.

23 years agofurther work on keybinding page
Steven M. Gava [Sat, 4 Aug 2001 13:58:22 +0000 (13:58 +0000)]
further work on keybinding page

23 years agoDerived from SF patch #446899 Permit import of .pyw under Windows, from
Tim Peters [Sat, 4 Aug 2001 08:12:36 +0000 (08:12 +0000)]
Derived from SF patch #446899 Permit import of .pyw under Windows, from
David Bolen.

23 years agoPlug a memory leak in Py_InitModule4(): when PyDict_SetItemString() failed,
Fred Drake [Sat, 4 Aug 2001 03:11:25 +0000 (03:11 +0000)]
Plug a memory leak in Py_InitModule4():  when PyDict_SetItemString() failed,
the object being inserted was not being DECREFed.

This closes SF bug #444486.

23 years agoLots of new text and example code on embedding Python in C, contributed
Fred Drake [Sat, 4 Aug 2001 01:58:36 +0000 (01:58 +0000)]
Lots of new text and example code on embedding Python in C, contributed
by Albert Hofkamp.  Some editing has been done for style and markup
consistency.

This also supplies an example of importing modules and calling a function
defined in the module, so this closes SF bug #440037 as well.

(The long example code was moved to a separate file so that it would
format properly.)

23 years agoNote that the mimetypes.MimeTypes class was added in Python 2.2.
Fred Drake [Sat, 4 Aug 2001 00:48:49 +0000 (00:48 +0000)]
Note that the mimetypes.MimeTypes class was added in Python 2.2.

23 years agoAdd documentation for the new aspects of the mimetypes module.
Fred Drake [Fri, 3 Aug 2001 21:03:14 +0000 (21:03 +0000)]
Add documentation for the new aspects of the mimetypes module.

This closes the SF bug (feature request) #439710.

23 years agoRefactor so that it is easier to work with alternate MIME types databases,
Fred Drake [Fri, 3 Aug 2001 21:01:44 +0000 (21:01 +0000)]
Refactor so that it is easier to work with alternate MIME types databases,
and programmatically extend the database in different ways.

This closes the SF bug (feature request) #439710.

23 years agoDon't use any characters C doesn't guarantee are safe for text-mode files.
Tim Peters [Fri, 3 Aug 2001 20:40:18 +0000 (20:40 +0000)]
Don't use any characters C doesn't guarantee are safe for text-mode files.
This should stop the bizarre translations Jack was getting from Mac CVS.

23 years agoFix stupid bug: when migrating these tests from the Zope repository, the
Fred Drake [Fri, 3 Aug 2001 19:53:01 +0000 (19:53 +0000)]
Fix stupid bug: when migrating these tests from the Zope repository, the
names of the test methods were not changed from the Zope-standard "check"
names to the Python-standard "test_" names, so the tests were not actually
being run.

Added test of hexadecimal character references as a regression check for
SF bug #445196.

23 years agoChange some comments into docstrings.
Fred Drake [Fri, 3 Aug 2001 19:50:59 +0000 (19:50 +0000)]
Change some comments into docstrings.

Fix handling of hexadecimal character references (legal in XHTML) so that
they are properly interpreted as character references.
This fixes SF bug #445196.

23 years ago+ Moved IDLE from the "tools" component to the main component.
Tim Peters [Fri, 3 Aug 2001 19:04:13 +0000 (19:04 +0000)]
+ Moved IDLE from the "tools" component to the main component.
+ Ditto pydoc.
  (IMO, both should have been done long ago -- simply didn't occur to
   me before)
+ Build the summary text into a vrbl instead of a temp file.  Doh!  Less
  fiddling, and should avoid another class of Win2K permission problems.
Bug:  the "auto vertical scrollbar" control on the summary page doesn't
work (never creates a scrollbar, no matter how much text).  So forced a
vertical scrollbar there.

23 years agoApply the new \mailheader macro where appropriate, and fix a few small
Fred Drake [Fri, 3 Aug 2001 18:39:36 +0000 (18:39 +0000)]
Apply the new \mailheader macro where appropriate, and fix a few small
markup inconsistencies.

23 years agoDefine the \mailheader macro and document conventions for using it.
Fred Drake [Fri, 3 Aug 2001 18:36:17 +0000 (18:36 +0000)]
Define the \mailheader macro and document conventions for using it.

23 years agoMinor markup nits and slight style-guide conformance changes.
Fred Drake [Fri, 3 Aug 2001 17:11:33 +0000 (17:11 +0000)]
Minor markup nits and slight style-guide conformance changes.

23 years agoPyMac_GetFSSpec must be called via the dylib glue code on OSX.
Jack Jansen [Fri, 3 Aug 2001 15:39:27 +0000 (15:39 +0000)]
PyMac_GetFSSpec must be called via the dylib glue code on OSX.

23 years agoIf we use frameworks we should include CoreServices/CoreServices.h, not the
Jack Jansen [Fri, 3 Aug 2001 15:36:23 +0000 (15:36 +0000)]
If we use frameworks we should include CoreServices/CoreServices.h, not the
old-style CoreFoundation.h.

23 years agoTXNTSMCheck isn't available under Mach-O Carbon.
Jack Jansen [Fri, 3 Aug 2001 15:34:29 +0000 (15:34 +0000)]
TXNTSMCheck isn't available under Mach-O Carbon.

23 years ago- On OSX add -no-cpp-precomp to OPT. The user still has to manually specify
Jack Jansen [Fri, 3 Aug 2001 15:32:23 +0000 (15:32 +0000)]
- On OSX add -no-cpp-precomp to OPT. The user still has to manually specify
  the --with-suffix=.exe, but it seems that that is also true for cygwin
  (or not? should I automatically set it?)
- Got --with-next-framework to build on OSX. This is only the build bit,
  the install still has to be done manually. Moreover, the Python build order
  isn't really suited to frameworks (where you want to do 'build lib',
  'install lib and framework', 'link executable against installed framework'
  in that order).

23 years agoICFindConfigFile isn't needed, and it's gone under carbon.
Jack Jansen [Fri, 3 Aug 2001 14:12:36 +0000 (14:12 +0000)]
ICFindConfigFile isn't needed, and it's gone under carbon.

23 years agoEasyDialogs has a working version of this.
Jack Jansen [Fri, 3 Aug 2001 14:10:54 +0000 (14:10 +0000)]
EasyDialogs has a working version of this.

23 years agoThese stopped working a long time ago, and they're not worth fixing.
Jack Jansen [Fri, 3 Aug 2001 14:10:27 +0000 (14:10 +0000)]
These stopped working a long time ago, and they're not worth fixing.

23 years agoObsolete, EasyDialogs has the same code (but working:-).
Jack Jansen [Fri, 3 Aug 2001 14:09:33 +0000 (14:09 +0000)]
Obsolete, EasyDialogs has the same code (but working:-).

23 years agoObsolete: uses old applescript support.
Jack Jansen [Fri, 3 Aug 2001 14:07:52 +0000 (14:07 +0000)]
Obsolete: uses old applescript support.

23 years agoObsolete, uses the old non-package-based applescript suite support.
Jack Jansen [Fri, 3 Aug 2001 14:06:17 +0000 (14:06 +0000)]
Obsolete, uses the old non-package-based applescript suite support.

23 years agoThis "test" showed an obsolete way to get scrolled lists in dialogs. Removed.
Jack Jansen [Fri, 3 Aug 2001 14:05:29 +0000 (14:05 +0000)]
This "test" showed an obsolete way to get scrolled lists in dialogs. Removed.

23 years agoThis "test" showed an obsolete way to get scrolled lists in dialogs. Removed.
Jack Jansen [Fri, 3 Aug 2001 14:02:07 +0000 (14:02 +0000)]
This "test" showed an obsolete way to get scrolled lists in dialogs. Removed.

23 years agoRemoved: it doesn't work anymore, and it wasn't all that useful anyway. The FrameWork...
Jack Jansen [Fri, 3 Aug 2001 13:58:16 +0000 (13:58 +0000)]
Removed: it doesn't work anymore, and it wasn't all that useful anyway. The FrameWork testcode shows similar functionality.