[ 400998 ] experimental support for extended slicing on lists
somewhat spruced up and better tested than it was when I wrote it.
Includes docs & tests. The whatsnew section needs expanding, and arrays
should support extended slices -- later.
transformed them into the initial_indent and subsequent_indent instance
attributes. Now they actually work as advertised, ie. they are
accounted for in the width of each output line. Plus you can use them
with wrap() as well as fill(), and fill() went from simple-and-broken to
trivial-and-working.
metaclasses.
This is essentially the same problem as that reported in bug 494904
for pickle: deepcopy should treat instances of custom metaclasses the
same way it treats instances of type 'type'.
Bugfix candidate.
and fill() methods. Keep interface of existing wrap() and fill()
functions by going back to having them construct a new TextWrapper
instance on each call, with the preferred width passed to the
constructor.
sentences are separated by two spaces.
Improve _fix_sentence_endings() a bit -- look for ".!?" instead of just
".", and factor out the list of sentence-ending punctuation characters
to a class attribute.
I've made considerable changes to Michael's code, specifically to use
the select() system call directly and to store the timeout as a C
double instead of a Python object; internally, -1.0 (or anything
negative) represents the None from the API.
I'm not 100% sure that all corner cases are covered correctly, so
please keep an eye on this. Next I'm going to try it Windows before
Tim complains.
No way is this a bugfix candidate. :-)
Straightforward fix. Will backport to 2.2. If there's ever a new 2.1
release, this could be backported there too (since it's an issue with
anything that's got both a __reduce__ and a __setstate__).
Straightforward fix. Will backport to 2.2. If there's ever a new 2.1
release, this could be backported there too (since it's an issue with
anything that's got both a __reduce__ and a __setstate__).
This is a conservative version of SF patch 504889. It uses the log
module instead of calling print in various places, and it ignores the
verbose argument passed to many functions and set as an attribute on
some objects. Instead, it uses the verbosity set on the logger via
the command line.
The log module is now preferred over announce() and warn() methods
that exist only for backwards compatibility.
XXX This checkin changes a lot of modules that have no test suite and
aren't exercised by the Python build process. It will need
substantial testing.
While I was at it, I added a tp_clear handler and changed the
tp_dealloc handler to use the clear_slots helper for the tp_clear
handler.
Also tightened the rules for slot names: they must now be proper
identifiers (ignoring the dirty little fact that <ctype.h> is locale
sensitive).
Also set mp->flags = READONLY for the __weakref__ pseudo-slot.
Most of this is a 2.2 bugfix candidate; I'll apply it there myself.
# XXX this isn't used anywhere, and worse, it has the same name as a method
# in Command with subtly different semantics. (This one just has one
# source -> one dest; that one has many sources -> one dest.) Nuke it?
Yes. Nuke it.
modules, distutils does not understand that the build version of the
source tree is needed.
This patch fixes distutils.sysconfig to understand that the running
Python is part of the build tree and needs to use the appropriate
"shape" of the tree. This does not assume anything about the current
directory, so can be used to build 3rd-party modules using Python's
build tree as well.
This is useful since it allows us to use a non-installed debug-mode
Python with 3rd-party modules for testing. It as the side-effect that
set_python_build() is no longer needed (the hack which was added to
allow distutils to be used to build the "standard" extension modules).
This closes SF patch #547734.
BOM_UTF32, BOM_UTF32_LE and BOM_UTF32_BE that represent the Byte
Order Mark in UTF-8, UTF-16 and UTF-32 encodings for little and
big endian systems.
The old names BOM32_* and BOM64_* were off by a factor of 2.
This closes SF bug http://www.python.org/sf/555360
Change the module constructor (module_init) to have the signature
__init__(name:str, doc=None); this prevents the call from type_new()
to succeed. While we're at it, prevent repeated calling of
module_init for the same module from leaking the dict, changing the
semantics so that __dict__ is only initialized if NULL.
Also adding a unittest, test_module.py.
This is an incompatibility with 2.2, if anybody was instantiating the
module class before, their argument list was probably empty; so this
can't be backported to 2.2.x.
The HTTPError class tries to act as a regular response objects for
HTTP protocol errors that include full responses. If the failure is
more basic, like no valid response, the __init__ choked when it tried
to initialize its superclasses in addinfourl hierarchy that requires a
valid response.
The solution isn't elegant but seems to be effective. Do not
initialize the base classes if there isn't a file object containing
the response. In this case, user code expecting to use the addinfourl
methods will fail; but it was going to fail anyway.
It might be cleaner to factor out HTTPError into two classes, only one
of which inherits from addinfourl. Not sure that the extra complexity
would lead to any improved functionality, though.
Partial fix for SF bug # 563665.
Bug fix candidate for 2.1 and 2.2.
-f/--fromfile <filename>
option. This runs all and only the tests named in the file, in the
order given (although -x may weed that list, and -r may shuffle it).
Lines starting with '#' are ignored.
This goes a long way toward helping to automate the binary-search-like
procedure I keep reinventing by hand when a test fails due to interaction
among tests (no failure in isolation, and some unknown number of
predecessor tests need to run first -- now you can stick all the test
names in a file, and comment/uncomment blocks of lines until finding a
minimal set of predecessors).
__call__() can be 2-3x slower than the equivalent normal method.
_handle_message(): The structure of message/rfc822 message has
changed. Now parent's payload is a list of length 1, and the zeroth
element is the Message sub-object. Adjust the printing of such
message trees to reflect this change.
There's some wierdness here, but the test ran before and not after,
so I'm just hacking the change out. Someone more motivated than
me can work out what's really happening.
Raymond: *PLEASE* run the test suite before checking things like
this in!
subclasses.
MIMENonMultipart: Base class for non-multipart/* content type subclass
specializations, e.g. image/gif. This class overrides attach() which
raises an exception, since it makes no sense to attach a subpart to
e.g. an image/gif message.
MIMEMultipart: Base class for multipart/* content type subclass
specializations, e.g. multipart/mixed. Does little more than provide
a useful constructor.
If a rexec instance allows writing in the current directory (a common
thing to do), there's a way to execute bogus bytecode. Fix this by
not allowing imports from .pyc files (in a way that allows a site to
configure things so that .pyc files *are* allowed, if writing is not
allowed).
I'll apply this to 2.2 and 2.1 too.
- Add comment explaining the structure of the stack.
- Minor optimization: make stack tuple directly usable as part of return
value for enter/exit events.
(or how do I "mark" something to be a candidate?)
fixed an old buglet that caused bdb to be unable to
continue in the botframe, after a breakpoint was set.
the key idea is not to set botframe to the bottom level frame,
but its f_back, which actually might be None.
Additional changes: migrated old exception trick to use
sys._getframe(), which exists both in 2.1 and 2.2 .
Note: I believe Mark Hammond needs to look over his code now.
F5 correctly starts up in the debugger, but later on doesn't stop at a given
breakpoint any longer.
kind regards - chris
[ 559250 ] more POSIX signal stuff
Adds support (and docs and tests and autoconfery) for posix signal
mask handling -- sigpending, sigprocmask and sigsuspend.
A MemoryError is now raised when the list cannot be created.
There is a test, but as the comment says, it really only
works for 32 bit systems. I don't know how to improve
the test for other systems (ie, 64 bit or systems
where the data size != addressable size,
e.g. 64 bit data, but 48 bit addressable memory)
instead of calling the getaddrlist() method, since the latter doesn't
work with multiple calls (it will return the empty list for the second
and subsequent calls).
Closes SF bug #555035. Include a unittest.
of the PyUNIT version of the same file. This helps people understand that
this version is the same as the version from the independent PyUNIT
release (confusion was indicated on the PyUNIT mailing list).
email package's Parser to handle the three common line endings.
Certain protocols such as IMAP define CRLF line endings and it doesn't
make sense for the client app to have to normalize the line endings
before handing it message off to the Parser.
_parsebody(): Be more flexible in the matching of line endings for
finding the MIME separators. Accept any of \r, \n and \r\n. Note
that we do /not/ change the line endings in the payloads, we just
accept any of those three around MIME boundaries.
single byte character sets. Also fixed a semantic problem with the
constructor's default arguments. Specifically,
__init__(): Change the maxlinelen argument default to None instead of
MAXLINELEN. The semantics should have been (and now are) that if
maxlinelen is given it is always honored. If it isn't given, but
header_name is given, then the maximum line length is calculated. If
neither are given then the default 76 characters is used.
_split(): If the character set is a single byte character set then we
can split the line at the maxlinelen because we know that encoding the
header won't increase its length. If the charset isn't a single byte
charset then we use the quicker divide-and-conquer line splitting
algorithm as before.
for the email package. The former is now just a shell project that
has some extra files for packaging for independent use (e.g. setup.py
and README).
Added a compatibility layer so that the same API can be used in Python
2.1 and 2.2/2.3 with the major differences shuffled off into helper
modules (_compat21.py and _compat22.py).
Also bumped the package version number to 2.0.3 for some fixes to be
checked in momentarily.
Scot Stevenson. Could be a bug fix candidate, but probably doesn't
matter much unless a certain blue-nosed cat suddenly becomes corporeal
and starts emailing some stmp.py (sic) fronted mailer.
returned a proxy for __class__ whose __bases__ was also a proxy. The
merge_class_dict() helper for dir() assumed incorrectly that __bases__
would always be a tuple and used the in-line tuple API on the proxy.
I will backport this to 2.2 as well.
test if 'callable' has not been supplied is to test for None instead of
False. The previous correction to 'if callable()' was wrong because an unusable
callback would be ignored rather than raising an exception.
and the .seed() and .whseed() methods failed to reset it. In other
words, setting the seed didn't completely determine the sequence of
results produced by random.gauss(). It does now. Programs repeatedly
mixing calls to a seed method with calls to gauss() may see different
results now.
Bugfix candidate (random.gauss() has always been broken in this way),
despite that it may change results.
This now does a dynamic analysis of which elements are so frequently
repeated as to constitute noise. The primary benefit is an enormous
speedup in find_longest_match, as the innermost loop can have factors
of 100s less potential matches to worry about, in cases where the
sequences have many duplicate elements. In effect, this zooms in on
sequences of non-ubiquitous elements now.
While I like what I've seen of the effects so far, I still consider
this experimental. Please give it a try!
On Win2K it thought 'foo' started at byte offset 0 instead of at the
pagesize, and on Win98 it thought 'foo' didn't exist at all. Somehow
or other this is related to the new "in memory file" gimmicks in
bsddb, but the old bsddb we use on Windows sucks so bad anyway I don't
want to bother digging deeper. Flushing the file in test_mmap after
writing to it makes the problem go away, so good enough.
build's "undetected error" problems were originally detected with
extension types, but we can whitebox test the same situations with
new-style classes.
Also add a test that Python doesn't die with SIGXFSZ if it exceeds the
file rlimit. (Assuming this will also test the behavior when the 2GB
limit is exceed on a platform that doesn't have large file support.)
closes SF #514433
can now pass 'None' as the filename for the bsddb.*open functions,
and you'll get an in-memory temporary store.
docs are ripped out of the bsddb dbopen man page. Fred may want to
clean them up.
Considering this for 2.2, but not 2.1.
http://www.python.org/sf/444708
This adds the optional argument for str.strip
to unicode.strip too and makes it possible
to call str.strip with a unicode argument
and unicode.strip with a str argument.
- islink() now returns true for alias files
- walk() no longer follows aliases while traversing
- realpath() implemented, returning an alias-free pathname.
As this could conceivably break existing code I think it isn't a bugfix candidate.
test data: this test fails on WIndows now if universal newlines are
enabled (which they aren't yet, by default). I don't know whether the
test will also fail on Linux now.
SF bug #522264 reported by Evelyn Mitchell.
The code included a comment about "STAR STAR" which was translated
into the code as the bogus attribute token.STARSTAR. This name never
caused an attribute error because it was never retrieved. The code
was based on an old version of the grammar that specified kwargs as
two tokens ('*' '*'). I checked as far back as 2.1 and didn't find
this production.
The fix is simple, because token.DOUBLESTAR is the only token
allowed. Also update the grammar fragment in com_arglist().
XXX I'll bet lots of other grammar fragments in comments are out of
date, probably in this module and in compile.c.
Close a file before trying to unlink it, and apparently Cygwin needs
writes to an mmap'ed file to get flushed before they're visible.
Bugfix candidate, but I think only for the 2.2 line (it's testing
features that I think were new in 2.2).
Change type_get_doc (the get function for __doc__) to look in tp_dict
more often, and if it finds a descriptor in tp_dict, to call it (with
a NULL instance). This means you can add a __doc__ descriptor to a
new-style class that returns instance docs when called on an instance,
and class docs when called on a class -- or the same docs in either
case, but lazily computed.
I'll also check this into the 2.2 maintenance branch.
If a str or unicode method returns the original object,
make sure that for str and unicode subclasses the original
will not be returned.
This should prevent SF bug http://www.python.org/sf/460020
from reappearing.
PyNumber_InPlaceMultiply insisted on calling sq_inplace_repeat if it
existed, even if nb_inplace_multiply also existed and the arguments
weren't right for sq_inplace_repeat. Change this to only use
sq_inplace_repeat if nb_inplace_multiply isn't defined.
Bugfix candidate.
double call to AddressList.getaddrlist(), and /that/ always returns an
empty list for the second and subsequent calls.
Instead, instantiate an AddressList directly, and get the parsed
addresses out of the addresslist attribute.
which requires that if there are ehlo parameters returned with an ehlo
keyword (in the response to EHLO), the keyword and parameters must be
delimited by an ASCII space. Thus responses like
250-AUTH=LOGIN
should be ignored as non-conformant to the RFC (the `=' isn't allowed
in the ehlo keyword).
This is a bug fix candidate.
Add a method zfill to str, unicode and UserString and change
Lib/string.py accordingly.
This activates the zfill version in unicodeobject.c that was
commented out and implements the same in stringobject.c. It also
adds the test for unicode support in Lib/string.py back in and
uses repr() instead() of str() (as it was before Lib/string.py 1.62)
In DatagramRequestHandler.setup(), the wfile initialization should be
StringIO.StringIO(), not StringIO.StringIO(slf.packet).
Bugfix candidate (all the way back to Python 1.5.2 :-).
Highlights: import and friends will understand any of \r, \n and \r\n
as end of line. Python file input will do the same if you use mode 'U'.
Everything can be disabled by configuring with --without-universal-newlines.
See PEP278 for details.
Add optional arg to string methods strip(), lstrip(), rstrip().
The optional arg specifies characters to delete.
Also for UserString.
Still to do:
- Misc/NEWS
- LaTeX docs (I did the docstrings though)
- Unicode methods, and Unicode support in the string methods.
The test function's signature should be
test(methodname, input, output, *args)
but the output argument was omitted. This caused all tests to fail,
because the expected output was passed as the initial argument to the
method call. But because of the way the test works (it compares the
results for a regular string to the results for a UserString instance
with the same value, and it's OK if both raise the same exception) the
test never failed!
I've fixed this, and also cleaned up a few warts in the verbose
output. Finally, I've made it possible to run the test stand-alone in
verbose mode by passing -v as a command line argument.
Now, the test will report failure related to zfill. That's not my
fault, that's a legitimate problem: the string_tests.py file contains
a test for the zfill() method (just added) but this method is not
implemented. The responsible party will surely fix this soon now.
non-us-ascii character sets in headers and bodies. Some API changes
(with DeprecationWarnings for the old APIs). Better RFC-compliant
implementations of base64 and quoted-printable.
Updated test cases. Documentation updates to follow (after I finish
writing them ;).
Change pickling format for bools to use a backwards compatible
encoding. This means you can pickle True or False on Python 2.3
and Python 2.2 or before will read it back as 1 or 0. The code
used for pickling bools before would create pickles that could
not be read in previous Python versions.
ContentHandler. While GC will eventually clean up, it can take longer than
normal for applications that create a lot of strings (or other immutables)
rather without creating many containers.
This closes SF bug #535474.
PEP 285. Everything described in the PEP is here, and there is even
some documentation. I had to fix 12 unit tests; all but one of these
were printing Boolean outcomes that changed from 0/1 to False/True.
(The exception is test_unicode.py, which did a type(x) == type(y)
style comparison. I could've fixed that with a single line using
issubtype(x, type(y)), but instead chose to be explicit about those
places where a bool is expected.
Still to do: perhaps more documentation; change standard library
modules to return False/True from predicates.
457466: popenx() argument mangling hangs python
226766: popen('python -c"...."') tends to hang
Fixes argument quoting in w9xpopen.exe for Windows 9x. w9xpopen.exe
also never attempts to display a MessageBox when not executed
interactively.
Added test_popen() test. This test currently just executes
"python -c ..." as a child process, and checks that the expected
arguments were all recieved correctly by the child process. This
test succeeds for me on Win9x, win2k and Linux, and I hope it does
for other popen supported platforms too :)
I really can't test this, but from reading the discussion in that bug
report, it's likely that this works. It may also close a whole bunch
of other bug reports related to urllib and proxies on Windows, but who
knows.
Walter Doerwald provided a patch, which I've modified in two ways:
1) (Uncontroversial) Removed code to make module work in earlier versions of
Python without the unicode() built-in
2) (Poss. controversial) Instead of making string.zfill take the repr()
of non-string objects, take the str().
Should a warning be added to this branch of the code so that the automatic
str() can be deprecated?
2.2.2 bugfix candidate, assuming the repr()->str() change is deemed OK.
It appears that getcomments() can get called for classes defined in
C. Since these don't have source code, it can't do anything useful.
A function buried many levels deep was raising a TypeError that was
not caught.
Who knows why this broke...
better local_hostname default. According to RFC 2821, it is
recommended that the fqdn hostname be provided in the EHLO/HELO verb
and if that can't be calculated, to use a domain literal.
The rationale for this change is documented in SF patch #497736 which
also had privacy concerns about leaking the fqdn in the EHLO/HELO. We
decided this wasn't a big concern because no user data is leaked, and
the IP will always be leaked. The local_hostname argument is provided
for those clients that are super paranoid.
Using localhost.localdomain may break some strict smtp servers so we
decided against using it as the default.
This fixes the symptom, but PRINT_ITEM has no way to know what (if
anything) PyFile_WriteObject() writes unless the object being printed
is a string. When the object isn't a string, this fix retains the
guess that softspace should be set after PyFile_WriteObject().
We might want to say that it's the job of filelike-object write methods
to leave the file's softspace in the correct state. That would probably
be better -- but everyone relies on PRINT_ITEM to guess for them now.
One more time on this turkey, but duller instead of cleverer.
Curious: The docs say __getslice__ has been deprecated since 2.0, but
list.__getitem__ still doesn't work if you pass it a slice. This makes
it a lot clearer to emulate a list by *being* a list <wink>.
Bugfix candidate. Michael, just pile this patch on top of the others
that went by -- no need to try to pick these apart.
* 'macdef' (macro definition) wasn't parsed correctly
* account value not reset for a subsequent 'default' line
* typo: 'whitepace' -> 'whitespace'
Bugfix candidate.
This patch makes it possible to pass Warning instances as the first
argument to warnings.warn. In this case the category argument
will be ignored. The message text used will be str(warninginstance).
FakeSocket class. Without it, the sendall() call got the method on
the underlying socket object, and that messed up SSL.
Does httplib use other methods of sockets that FakeSocket doesn't support?
Someone should take a look... (I'll try to give it a once-over.)
2.2.1 bugfix candidate.
The proper fix is not quite what was submitted; it's really better to
take the class of the object passed rather than calling PyMethod_New
with NULL pointer args, because that can then cause other core dumps
later.
I also added a testcase for the fix to classmethods() in test_descr.py.
I've already applied this to the 2.2 branch.
(as it is often not on Windows). The code was always designed so that it
would raise an IOError if there was no .netrc. But if there was no $HOME
it would return a KeyError which would be somewhat unexpected for code
that didn't know the algorithm it used to find .netrc. The particular
code that triggered this problem for me was ftpmirror.py which handled
the IOError gracefully, but not the KeyError.
particular, negative indexes work and they are limited by the actual length
of the names they represent (weekday and month names). This closes bug
#503202.
As promised in my response to the bug report, I'm not really fixing
it; in fact, one could argule over what the proper fix should do.
Instead, I'm adding a little magic that raises TypeError if you try to
pickle an instance of a class that has __slots__ but doesn't define or
override __getstate__. This is done by adding a bozo __getstate__
that always raises TypeError.
Bugfix candidate (also the checkin to typeobject.c, of course).
and (b) stop trying to prevent file growth.
Beef up the file.truncate() docs.
Change test_largefile.py to stop assuming that f.truncate() moves the
file pointer to the truncation point, and to verify instead that it leaves
the file position alone. Remove the test for what happens when a
specified size exceeds the original file size (it's ill-defined, according
to the Single Unix Spec).
dropping MS's inadequate _chsize() function. This was inspired by
SF patch 498109 ("fileobject truncate support for win32"), which I
rejected.
libstdtypes.tex: Someone who knows should update the availability
blurb. For example, if it's available on Linux, it would be good to
say so.
test_largefile: Uncommented the file.truncate() tests, and reworked to
do more. The old comment about "permission errors" in the truncation
tests under Windows was almost certainly due to that the file wasn't open
for *write* access at this point, so of course MS wouldn't let you
truncate it. I'd be appalled if a Unixish system did.
CAUTION: Someone should run this test on Linux (etc) too. The
truncation part was commented out before. Note that test_largefile isn't
run by default.
Adapter from SF patch 528038; fixes SF bug 527816.
The wrapper for __nonzero__ should be wrap_inquiry rather than
wrap_unaryfunc, since the slot returns an int, not a PyObject *.
In August, Greg said this looked good, so I'm going ahead with it.
The fix is different from the one in the bug report. Instead of using
a regular expression to extract the host from the url, I use
urlparse.urlsplit.
Martin commented that the patch doesn't address URLs that have basic
authentication username and password in the header. I don't see any
code anywhere in httplib that supports this feature, so I'm not going
to address it for this fix.
Bug fix candidate.
asyncore.poll, the select fails with EINTR, which the
code catches. However, the code fails to clear the
r/w/e arrays (like poll3 does), which means it acts as
if every descriptor had received all possible events.
Bug report and patch by Cesar Eduardo Barros
mmap_find_method(): this obtained the string to find via s#, but it
ignored its length, acting as if it were \0-terminated instead.
Someone please run on Linux too (the extended test_mmap works on Windows).
Bugfix candidate.
Due to the bizarre definition of _PyLong_Copy(), creating an instance
of a subclass of long with a negative value could cause core dumps
later on. Unfortunately it looks like the behavior of _PyLong_Copy()
is quite intentional, so the fix is more work than feels comfortable.
This fix is almost, but not quite, the code that Naofumi Honda added;
in addition, I added a test case.
rexec.
When using a restricted environment, imports of copy will fail with an
AttributeError when trying to access types.CodeType.
Bugfix candidate (all the way back to 1.5.3, but at least 2.1.3 and
2.2.1).
- Use substring search, not re search for user-agent and paths.
- Consider * entry last. Unquote, then requote URLs.
- Treat empty Disallow as "allow everything".
Add test cases. Fixes#523041
There were never tests for the fact that list() always returns a *new*
list object, even when the argument is a list, while tuple() may
return a reference to the argument when it is a tuple. Now there are.
Windows: apply normcase() as well as abspath(). (Note: this isn't
needed to make IDLE work, but it's a good idea anyway.)
Bugfix candidate -- both 2.2.1 and 2.1.3.
--install-script ... command line option to bdist_wininst) at the end
of the installation and at the start of deinstallation. Output
(stdout, stderr) of the script (if any) is displayed in the last
screen at installation, or in a simple message box at deinstallation.
sys.argv[1] for the script will contain '-install' at installation
time or '-remove' at deinstallation time.
The installation script runs in an environment (embedded by the
bdist_wininst runtime) where an additional function is available as
builtin:
create_shortcut(path, description, filename,
[arguments[, workdir[, iconpath, iconindex]]])
Recreated this file after source changes.
SSL support. test_socket.py passes again on Windows.
Added an XXX about adding _ssl exports to the __all__ list (it doesn't
appear to be doing anything about that now, but since I don't have SSL
on this box I can't really tell).
helper module _ssl.
The support for the RAND_* APIs in _ssl is now only enabled
for OpenSSL 0.9.5 and up since they were added in that
release.
Note that socketmodule.* should really be renamed to _socket.* --
unfortunately, this seems to lose the CVS history of the file.
Please review and test... I was only able to test the header file
chaos in socketmodule.c/h on Linux. The test run through fine
and compiles don't give errors or warnings.
WARNING: This patch does *not* include changes to the various
non-Unix build process files.
Fix exit races in test_thread.py and test_threaded_import.py.
I suspect the bug is provokable only under Linux (where child threads
seem to get lots of cycles before they get killed after the main thread
exits), or on multi-processor machines running other OSes.
Bugfix candidate.
Fix for the UTF-8 decoder: it will now accept isolated surrogates
(previously it raised an exception which causes round-trips to
fail).
Added new tests for UTF-8 round-trip safety (we rely on UTF-8 for
marshalling Unicode objects, so we better make sure it works for
all Unicode code points, including isolated surrogates).
Bumped the PYC magic in a non-standard way -- please review. This
was needed because the old PYC format used illegal UTF-8 sequences
for isolated high surrogates which now raise an exception.
installations are present, by always unlinking the destination file
before copying to it. Without the unlink(), the copied file remains
owned by its previous UID, causing the subsequent chmod() to fail.
Bugfix candidate, though it may cause changes on platforms where
file ownership behaves differently.
contain the type of the file (regular file, socket, link, &c.).
This means that install_scripts will now print
"changing mode of <file> to 775" instead of "... to 100775".
2.2 bugfix candidate, I suppose, though this isn't actually fixing a bug.
TemproraryFileWrapper wrapper anymore, and should be immune from the
problem that a temp file inherited by a spawned process caused an
attempt to close the temp file in the spawning process to blow
up (the unlink in TemporaryFileWrapper.close() blew up with a
"Permission denied" error because, despite that the temp file got
closed in the spawning process, the spawned process still had it open
by virtue of C-level file descriptor inheritance). In context,
that bug took days to figure out <wink/sigh>.
This is an ancient race when multiple threads call gettempdir() (or
anything relying on it) for the first time.
Fixed x-platform via the Big Hammer of rearranging the code to serialize
the first calls. Subsequent calls are as fast as before.
Note that the Python test suite can't provoke this bug: it requires
setting up multiple threads making the very first calls into tempfile,
but the test suite uses tempfile several times before getting to
test_threadedtempfile.
Bugfix candidate.
present - at least the swigged file should be named <name>_wrap.c as
this is also SWIG's default. (Even better would be to generate the
wrapped sources in a different location, but I'll leave this for
later).
Newer versions of SWIG don't accept the -dnone flag any more.
Since virtually nobody uses SWIG with distutils, this should do no
harm.
Suggested be Martin Bless on c.l.p.
the finally clause. An exception here could happen when a daemon
thread exits after the threading module has already been trashed by
the import finalization, and there's not much of a point in trying to
insist doing the cleanup in that stage.
This should fix SF bug ##497111: active_limbo_lock error at program
exit.
2.1.2 and 2.2.1 Bugfix candidate!
deepcopy(), _reconstruct(): pass the memo to the other function, so
that recursive data structures built out of new-style objects may be
deeply copied correctly.
2.2.1 bugfix!
deepcopy(), _reconstruct(): pass the memo to the other function, so
that recursive data structures built out of new-style objects may be
deeply copied correctly.
2.2.1 bugfix!
Instead of sending the real user and host, use "anonymous@" (i.e. no
host name at all!) as the default anonymous FTP password. This avoids
privacy violations.
doesn't have the _HEAPTYPE flag set, e.g. for time.struct_time and
posix.stat_result.
This fixes the immediate symptoms of SF bug #496873 (cPickle /
time.struct_time loop), replacing the infinite loop with an exception.
Make dumbdbm merely "dumb", rather than "terminally broken". Without this
patch, it's almost impossible to use dumbdbm _without_ causing horrible
datalossage. With this patch, dumbdbm passes my own horrible torture test,
as well as the roundup test suite.
dumbdbm really could do with a smidgin of a rewrite or two, but that's not
suitable for the release21-maint branch.
rfc822.AddressList incorrectly handles empty address.
"<>" is converted to None and should be "".
AddressList.__str__() fails on None.
I got an email with such an address and my program
failed processing it.
Example:
>>> import rfc822
>>> rfc822.AddressList("<>").addresslist
[('', None)]
>>> str(rfc822.AddressList("<>"))
Traceback (most recent call last):
File "<stdin>", line 1, in ?
File "/usr/lib/python2.1/rfc822.py", line 753, in __str__
return ", ".join(map(dump_address_pair,
self.addresslist))
TypeError: sequence item 0: expected string, None found
[His solution: in the internal routine AddrlistClass.getrouteaddr(),
initialize adlist to "".]
metaclass, reported by Dan Parisien.
Objects that are instances of custom metaclasses, i.e. whose class is
a subclass of 'type', should be pickled the same as new-style classes
(objects whose class is 'type'). This can't be done through a
dispatch table entry, and the __reduce__ trick doesn't work for these,
since it finds the unbound __reduce__ for instances of the class
(inherited from 'object'). So check explicitly using issubclass().
This way, when a socket object is deleted after the socket module has
already been zapped by module shutdown, we don't get annoying warnings
about exceptions in __del__ methods.
crashes.
If no external zip-utility is found, the archive is created by the
zipfile module, which behaves different now than in 2.1: if the
zip-file is created in the root directory if the distribution, it will
contain an (empty) version of itself.
This triggered the above bug - so it's better to create the zip-file
far away in the TMP directory.
paren. This was there to worm around a stupid XEmacs bug, but since I
can't tickle the bug in newer XEmacsen (just tried w/21.4.5) it's
possible the problem has been fixed. We shouldn't have to be working
around editor bugs anyway.
If it crops up again, I'll report it (again) to the XEmacs crowd.
Dietmar Schwertberger.
Bugfix candidate.
"""
RISCOS/Modules/getpath_riscos.c:
Include trailing '\0' when using strncpy [copy
strlen(...)+1 characters].
Lib/plat-riscos/riscospath.py:
Use riscosmodule.expand for os.path.abspath.
[fixes problems with site.py where
abspath("<Python$Dir>") returned
join(os.getcwd(), "<Python$Dir>") as e.g.
"SCSI::SCSI4.$.<Python$Dir>" because "<Python$Dir>"
wasn't recognised as an absolute path.]
"""
Rev 1.20 introduced a call to getpeername() in the dispatcher
constructor. This only works for a connected socket. Apparently
earlier versions of the code worked with un-connected sockets, e.g. a
listening socket.
It's not clear that the code is supposed to accept these sockets,
because it sets self.connected = 1 when passed a socket. But it's
also not clear that it should be a fatal error to pass a listening
socket.
The solution, for now, is to put a try/except around the getpeername()
call and continue if it fails. The self.addr attribute is used
primarily (only?) to produce a nice repr for the object, so it hardly
matters. If there is a real error on a connected socket, it's likely
that subsequent calls will fail too.
Fix for SF bug #492345. (I could've sworn I checked this in, but
apparently I didn't!)
This code:
class Classic:
pass
class New(Classic):
__metaclass__ = type
attempts to create a new-style class with only classic bases -- but it
doesn't work right. Attempts to fix it so it works caused problems
elsewhere, so I'm now raising a TypeError in this case.
>
> When using 'distutils' (shipped with Python 2.1) I've found that my
> Python scripts installed with a first line of:
>
> #!/usr/bin/python2.1None
>
> This is caused by distutils trying to patch the first line of the python
> script to use the current interpreter.
- the repr of unicode. Jython only add the u'' if the string contains char
values > 255.
- A unicode arg to unicode() is perfectly valid in jython.
- A test buffer() test. No buffer() on Jython
This closes patch "[ #490920 ] Jython and test_unicode".
Using grid methods on ScrolledText widgets does not
work as expected. It either fails to pack a widget, or
can even cause Tk to lock up.
The problem is that the .grid method is being called on
the text widget, not the frame widget. This can lead
to the well-known lockup in Tk when a frame's children
are managed by both the pack and grid managers. Even
if it doesn't lock up, the frame is never placed within
the intended widget.
Program fragment:
>>> import ScrolledText
>>> s = ScrolledText.ScrolledText()
>>> s.grid(row=0, column=0, rowspan=2)
The following patch uses the same hack to copy the
'grid' and 'place' geometry manager methods to the
ScrolledText instance as is already used for the 'pack'
manager.
backed out of broken minimal repeat patch from July
also fixed a couple of minor potential resource leaks in pattern_subx
(Guido had already fixed the big one)
type.__module__ behavior.
This adds the module name and a dot in front of the type name in every
type object initializer, except for built-in types (and those that
already had this). Note that it touches lots of Mac modules -- I have
no way to test these but the changes look right. Apologies if they're
not. This also touches the weakref docs, which contains a sample type
object initializer. It also touches the mmap test output, because the
mmap type's repr is included in that output. It touches object.h to
put the correct description in a comment.
open_http():
In urllib.py library module, URLopener.open_https()
returns a class instance of addinfourl() with its
self.url property missing the protocol.
Instead of "https://www.someurl.com", it becomes
"://www.someurl.com".
1. Acknowledge the welknown difference that jython
allows continue in the finally clause.
2. Avoid using _testcapi when running with jython.
This closes patch "[ #490417 ] Jython and test_exceptions"
twice! Fixed this by avoiding the import of test_email, which loads
the module a second time in that situation, and fiddled the __main__
section to resemble other test suites using unittest.
annoying that often you have to hit ^C numerous times before it
works. The solution: before the "except:" clause, insert "except
KeyboardInterrupt: raise". This propagates KeyboardInterrupt out,
stopping the test in its tracks.
distutils for the library modules built as shared objects. A better solution
appears possible, but with the threat that the distutils becomes more
magical ("complex").
This closes SF bug #458343.
initialized, this will be None, but the functions will still work (there will
simply be a bogus parent on the screen). Allowing the parent to be None
is useful when testing the functions from an interactive interpreter.
Add an optional keyword paramter "show" to the _QueryString class; when given
it is used to set the -show option to the entry widget. This allows passing
show="*" or the like to askstring(), making it useful for requesting
passwords/passphrases from the user.
This closes SF bug #438517.
Changed a docstring to be less font-lock-hostile.
Big Hammer to implement -Qnew as PEP 238 says it should work (a global
option affecting all instances of "/").
pydebug.h, main.c, pythonrun.c: define a private _Py_QnewFlag flag, true
iff -Qnew is passed on the command line. This should go away (as the
comments say) when true division becomes The Rule. This is
deliberately not exposed to runtime inspection or modification: it's
a one-way one-shot switch to pretend you're using Python 3.
ceval.c: when _Py_QnewFlag is set, treat BINARY_DIVIDE as
BINARY_TRUE_DIVIDE.
test_{descr, generators, zipfile}.py: fiddle so these pass under
-Qnew too. This was just a matter of s!/!//! in test_generators and
test_zipfile. test_descr was trickier, as testbinop() is passed
assumptions that "/" is the same as calling a "__div__" method; put
a temporary hack there to call "__truediv__" instead when the method
name is "__div__" and 1/2 evaluates to 0.5.
Three standard tests still fail under -Qnew (on Windows; somebody
please try the Linux tests with -Qnew too! Linux runs a whole bunch
of tests Windows doesn't):
test_augassign
test_class
test_coercion
I can't stay awake longer to stare at this (be my guest). Offhand
cures weren't obvious, nor was it even obvious that cures are possible
without major hackery.
Question: when -Qnew is in effect, should calls to __div__ magically
change into calls to __truediv__? See "major hackery" at tail end of
last paragraph <wink>.
It was easier than I thought, assuming that no other things contribute
to the instance size besides slots -- a pretty good bet. With a test
suite, no less!
happy if one could delete the __dict__ attribute of an instance. I
love to make Jim happy, so here goes...
- New-style objects now support deleting their __dict__. This is for
all intents and purposes equivalent to assigning a brand new empty
dictionary, but saves space if the object is not used further.
int_mul(): new and vastly simpler overflow checking. Whether it's
faster or slower will likely vary across platforms, favoring boxes
with fast floating point. OTOH, we no longer have to worry about
people shipping broken LONG_BIT definitions <0.9 wink>.
There's now a new structmember code, T_OBJECT_EX, which is used for
all __slot__ variables (except __weakref__, which has special behavior
anyway). This new code raises AttributeError when the variable is
NULL rather than converting NULL to None.
In goahead(), use a bound version of rawdata.startswith() since we use the
same method all the time and never change the value of rawdata. This can
save a lot of bound method creation.
Rather than tweaking the inheritance of type object slots (which turns
out to be too messy to try), this fix adds a __hash__ to the list and
dict types (the only mutable types I'm aware of) that explicitly
raises an error. This has the advantage that list.__hash__([]) also
raises an error (previously, this would invoke object.__hash__([]),
returning the argument's address); ditto for dict.__hash__.
The disadvantage for this fix is that 3rd party mutable types aren't
automatically fixed. This should be added to the rules for creating
subclassable extension types: if you don't want your object to be
hashable, add a tp_hash function that raises an exception.
Also, it's possible that I've forgotten about other mutable types for
which this should be done.
Kurt B. Kaiser
Guido van Rossum
Neal Norwitz
Michael W. Hudson
Martin v. Löwis
Steven M. Gava
Greg Ward
Skip Montanaro
Raymond Hettinger
Jeremy Hylton
Walter Dörwald
Piers Lauder
Fred Drake
Tim Peters
Barry Warsaw
Andrew M. Kuchling
Christian Tismer
Chui Tey
Andrew MacIntyre
Thomas Heller
Neil Schemenauer
Jack Jansen
Anthony Baxter
Mark Hammond
Marc-André Lemburg
Paul Prescod
Just van Rossum
Fredrik Lundh
Steve Purcell
Finn Bock
Ka-Ping Yee