Commit Graph

21785 Commits

Author SHA1 Message Date
Tim Peters ddea208be9 Give Python a debug-mode pymalloc, much as sketched on Python-Dev.
When WITH_PYMALLOC is defined, define PYMALLOC_DEBUG to enable the debug
allocator.  This can be done independent of build type (release or debug).
A debug build automatically defines PYMALLOC_DEBUG when pymalloc is
enabled.  It's a detected error to define PYMALLOC_DEBUG when pymalloc
isn't enabled.

Two debugging entry points defined only under PYMALLOC_DEBUG:

+ _PyMalloc_DebugCheckAddress(const void *p) can be used (e.g., from gdb)
  to sanity-check a memory block obtained from pymalloc.  It sprays
  info to stderr (see next) and dies via Py_FatalError if the block is
  detectably damaged.

+ _PyMalloc_DebugDumpAddress(const void *p) can be used to spray info
  about a debug memory block to stderr.

A tiny start at implementing "API family" checks isn't good for
anything yet.

_PyMalloc_DebugRealloc() has been optimized to do little when the new
size is <= old size.  However, if the new size is larger, it really
can't call the underlying realloc() routine without either violating its
contract, or knowing something non-trivial about how the underlying
realloc() works.  A memcpy is always done in this case.

This was a disaster for (and only) one of the std tests:  test_bufio
creates single text file lines up to a million characters long.  On
Windows, fileobject.c's get_line() uses the horridly funky
getline_via_fgets(), which keeps growing and growing a string object
hoping to find a newline.  It grew the string object 1000 bytes each
time, so for a million-character string it took approximately forever
(I gave up after a few minutes).

So, also:

fileobject.c, getline_via_fgets():  When a single line is outrageously
long, grow the string object at a mildly exponential rate, instead of
just 1000 bytes at a time.

That's enough so that a debug-build test_bufio finishes in about 5 seconds
on my Win98SE box.  I'm curious to try this on Win2K, because it has very
different memory behavior than Win9X, and test_bufio always took a factor
of 10 longer to complete on Win2K.  It *could* be that the endless
reallocs were simply killing it on Win2K even in the release build.
2002-03-23 10:03:50 +00:00
Skip Montanaro 91cc17d20e Only AttributeError can be raised in this situation - on systems without
getuid or getpid.  posix_getuid & posix_getpid never raise exceptions when
called with no args.
2002-03-23 05:58:52 +00:00
Skip Montanaro 3c4a629bdc import statements only raise ImportError, right? 2002-03-23 05:55:18 +00:00
Skip Montanaro db5d1444a1 tighten up except - only ValueError can be raised in this situation 2002-03-23 05:50:17 +00:00
Skip Montanaro 3c643d8db3 tighten up except - int() only raises ValueError 2002-03-23 05:47:31 +00:00
Skip Montanaro 6ec967d066 added RFC 2396 tests from Aaron Swartz included in bug # 450225.
converted to use unittest
2002-03-23 05:32:10 +00:00
Skip Montanaro b1ba6b0044 no longer needed - converted test_urlparse.py to use unittest 2002-03-23 05:29:59 +00:00
Tim Peters 0c2c8e77fb SF bug 533234: tm_isdst > 1 Passed to strftime.
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.
2002-03-23 03:26:53 +00:00
Neil Schemenauer ed19b88f0b Check in (hopefully) corrected version of last change. 2002-03-23 02:06:50 +00:00
Tim Peters ce7fb9b515 Just whitespace fiddling. 2002-03-23 00:28:57 +00:00
Tim Peters 1221c0a435 Build obmalloc.c directly instead of #include'ing from object.c.
Also move all _PyMalloc_XXX entry points into obmalloc.c.

The Windows build works fine.
The Unix build is changed here (Makefile.pre.in), but not tested.
No other platform's build process has been fiddled.
2002-03-23 00:20:15 +00:00
Neil Schemenauer c24ea08644 Disable the parser hacks that enabled the "yield" keyword using a future
statement.
2002-03-22 23:53:36 +00:00
Neil Schemenauer 12a6d942d8 Undo last commit. It's causing the tests to file. 2002-03-22 23:50:30 +00:00
Neil Schemenauer c155dd4ca9 Disable the parser hacks that allowed the "yield" keyword to be enabled
by a future statement.
2002-03-22 23:38:11 +00:00
Neil Schemenauer 558ba52f10 Remove malloc hooks. 2002-03-22 23:20:15 +00:00
Neil Schemenauer fa79c65235 Match behavior of the pickle.py module more closely. 2002-03-22 23:02:53 +00:00
Neil Schemenauer 79f181395b Add more example exceptions that unpickling can raise. 2002-03-22 22:16:03 +00:00
Neil Schemenauer 94b866a030 Handle os.listdir("") case correctly on Windows. Closes bug 500705. 2002-03-22 20:51:58 +00:00
Neil Schemenauer 398b9f6d6d Disallow open()ing of directories. Closes SF bug 487277. 2002-03-22 20:38:57 +00:00
Skip Montanaro 07c57d4e60 better solution for bug #533234 courtesy of Tim.
Michael: use this version as the bugfix candidate...
2002-03-22 18:35:51 +00:00
Skip Montanaro e8c6a3eef6 guarantee that the dst flag of synthetic "time" tuples passed to strftime
is always 0.  This closes bug #533234.
2002-03-22 18:07:49 +00:00
Walter Dörwald 92b48b739f use stat attributes instead of tuple entries
and remove the unneccessary "import stat" statement.
2002-03-22 17:30:38 +00:00
Fred Drake cf43004b83 Fix broken HTML in the head; this was reported by a user as causing Opera 6.01
to crash.  The user has reported the problem to Opera, but we still should
generate something that passes for HTML.
2002-03-22 17:22:38 +00:00
Neil Schemenauer 9d50d138e7 Add news about pymalloc being enabled. 2002-03-22 17:06:59 +00:00
Neil Schemenauer 1b0e4fcc29 Use pymalloc for realloc() as well. 2002-03-22 15:41:03 +00:00
Andrew M. Kuchling b26ca9db2e Revert part of previous patch: several install_* subcommands expect
.compile to be None, and set it to true if it is.
    Caught by Pearu Peterson.

Bugfix candidate, if the previous change is accepted for
release22-maint.
2002-03-22 15:35:17 +00:00
Neil Schemenauer 16c22976c3 Enable pymalloc by default. 2002-03-22 15:34:49 +00:00
Neil Schemenauer dcc819a5c9 Use pymalloc if it's enabled. 2002-03-22 15:33:15 +00:00
Neil Schemenauer a1a9c51a3e Add pymalloc object memory management functions. These must be
available even if pymalloc is disabled since extension modules might use
them.
2002-03-22 15:28:30 +00:00
Neil Schemenauer ffd5399728 Make PyObject_{NEW,New,Del,DEL} always use the standard malloc (PyMem_*)
and not pymalloc.  Add the functions PyMalloc_New, PyMalloc_NewVar, and
PyMalloc_Del that will use pymalloc if it's enabled.   If pymalloc is
not enabled then they use the standard malloc (PyMem_*).
2002-03-22 15:25:18 +00:00
Jack Jansen 150ed6113c Generate with weaklink stubs, so missing routines (on MacOS 8.6 and earlier)
don't cause import failure.

Fixes 531398, 2.2.1 candidate.
2002-03-22 14:16:39 +00:00
Jack Jansen 9051e0e577 - Weaklink InterfaceLib in _Res module
- forgot to pass libraryflags and stdlibraryflags on to ppc/carbon
  project generation.

First half of fix to 531398.
2002-03-22 14:15:07 +00:00
Tim Peters b26f3639ed Enable pymalloc by default in the Windows build. 2002-03-22 06:32:32 +00:00
Andrew M. Kuchling 76fffd81e9 Add a simple test suite for netrc.py, and remove it from test_sundry 2002-03-22 02:48:57 +00:00
Tim Peters bab22beda8 SF bug 533198: Complex power underflow raises exception.
Konrad was too kind.  Not only did it raise an exception, the specific
exception it raised made no sense.  These are old bugs in complex_pow()
and friends:

1. Raising 0 to a negative power isn't a range error, it's a domain
   error, so changed c_pow() to set errno to EDOM in that case instead
   of ERANGE.

2. Changed complex_pow() to:

A. Used the Py_ADJUST_ERANGE2 macro to try to clear errno of a spurious
   ERANGE error due to underflow in the libm pow() called by c_pow().

B. Produced different exceptions depending on the errno value:
   i) For errno==EDOM, raise ZeroDivisionError instead of ValueError.
      This is for consistency with the non-complex cases 0.0**-2 and
      0**-2 and 0L**-2.
   ii) For errno==ERANGE, raise OverflowError.

Bugfix candidate.
2002-03-22 02:48:46 +00:00
Andrew M. Kuchling 366a1df7f1 [Bug #532115] netrc module was broken
* 'macdef' (macro definition) wasn't parsed correctly
   * account value not reset for a subsequent 'default' line
   * typo: 'whitepace' -> 'whitespace'

Bugfix candidate.
2002-03-22 02:46:41 +00:00
Andrew M. Kuchling 83d042d3a7 [Bug #532136] Change path in README, as suggested by Ralph Corderoy
Bugfix candidate.
2002-03-21 23:52:20 +00:00
Andrew M. Kuchling 3f1822b468 Add missing Boolean options
Remove unused no_compile flag
Initialize the Boolean attribute .compile to 0 instead of None

Bugfix candidate.
2002-03-21 23:46:54 +00:00
Andrew M. Kuchling da9f0bf4dd Add unlisted Boolean options. Thomas H., can you please check that I
got this right?

Bugfix candidate, unless Thomas notes a problem.
2002-03-21 23:44:01 +00:00
Andrew M. Kuchling 5de82a855e [Bug #517451] bdist_rpm didn't list all of its Boolean options.
(Someone should check the other commands for this same error.)

Bugfix candidate.
2002-03-21 23:27:54 +00:00
Jack Jansen 562baabb61 Added an open_pathname() method which opens a resource file by pathname,
possibly converting from AppleSingle.
2002-03-21 22:38:32 +00:00
Jack Jansen b9e6661206 For reasons I don't fully understand we sometimes get unexpected events
in MachoPython. As we don't have MacOS.HandleEvent() we drop these on
the floor (with a print).
2002-03-21 22:36:57 +00:00
Jack Jansen 0e5a733887 Solve sync() signature conflict differently: manually return 0 if
we have GUSI's void sync().
2002-03-21 21:09:36 +00:00
Jack Jansen 99286f96ac New checkin to convert from BinHex to AppleSingle encoding. 2002-03-21 20:18:32 +00:00
Martin v. Löwis 41933dd1b7 Patch #532729: check for sem_init in -lrt. 2002-03-21 15:10:58 +00:00
Neal Norwitz e22d3dfcc1 Fix grammar 2002-03-21 12:58:54 +00:00
Walter Dörwald b25c2b0a4a [Apply SF patch #504943]
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).
2002-03-21 10:38:40 +00:00
Martin v. Löwis 047c05ebc4 Do not insert characters for unicode-escape decoders if the error mode
is "ignore". Fixes #529104.
2002-03-21 08:55:28 +00:00
Andrew M. Kuchling bdf1f19fee [Apply patch #500457 from the PyXML tracker]
Add iterator support to pulldom.DOMEventStream

New feature, so not a bugfix candidate (though it should be safe for inclusion)
2002-03-20 23:56:34 +00:00
Neal Norwitz 3afb2d2bba Remove compiler warnings on Solaris 8.
Can go into 2.2.x, but not necessary.
2002-03-20 21:32:07 +00:00