cpython-lru_cache-weakref: Modules/structmodule.c history

patch #1493701: performance enhancements for struct module

Tue, 23 May 2006 19:11:34 +0000

patch #1493701: performance enhancements for struct module

revert #1493701

Tue, 23 May 2006 19:09:51 +0000

revert #1493701

struct_pack(): Repair new assert-fail crash in

Thu, 16 Feb 2006 20:19:46 +0000

struct_pack(): Repair new assert-fail crash in
debug-build test_struct on a box where plain "char"
is signed.

Use Py_ssize_t for counts and sizes.

Thu, 16 Feb 2006 14:30:23 +0000

Use Py_ssize_t for counts and sizes.

Check return result from Py_InitModule*(). This API can fail.

Thu, 19 Jan 2006 06:09:39 +0000

Check return result from Py_InitModule*(). This API can fail.

Probably should be backported.

SF bug #1072182: bad arg type to isspace in struct module

Fri, 26 Aug 2005 08:42:13 +0000

SF bug #1072182: bad arg type to isspace in struct module

Patch #1011240: SystemError generated by struct.pack('P', 'foo').

Mon, 27 Sep 2004 19:27:51 +0000

Patch #1011240: SystemError generated by struct.pack('P', 'foo').

Getting rid of code dependent on GUSI or the MetroWerks compiler.

Wed, 19 Nov 2003 22:52:23 +0000

Getting rid of code dependent on GUSI or the MetroWerks compiler.

Rename LONG_LONG to PY_LONG_LONG. Fixes #710285.

Sat, 29 Mar 2003 10:06:18 +0000

Rename LONG_LONG to PY_LONG_LONG. Fixes #710285.

New private API functions _PyFloat_{Pack,Unpack}(4,8}. This is a

Thu, 20 Mar 2003 20:53:32 +0000

New private API functions _PyFloat_{Pack,Unpack}(4,8}. This is a
refactoring to get all the duplicates of this delicate code out of the
cPickle and struct modules.

SF bug 705836: struct.pack of floats in non-native endian order

Thu, 20 Mar 2003 18:32:13 +0000

SF bug 705836: struct.pack of floats in non-native endian order

pack_float, pack_double, save_float: All the routines for creating
IEEE-format packed representations of floats and doubles simply ignored
that rounding can (in rare cases) propagate out of a long string of
1 bits. At worst, the end-off carry can (by mistake) interfere with
the exponent value, and then unpacking yields a result wrong by a factor
of 2. In less severe cases, it can end up losing more low-order bits
than intended, or fail to catch overflow *caused* by rounding.

Bugfix candidate, but I already backported this to 2.2.

In 2.3, this code remains in severe need of refactoring.

(Most of) SF patch 601369 (Christos Georgiou): obmalloc,structmodule:

Tue, 03 Sep 2002 18:42:21 +0000

(Most of) SF patch 601369 (Christos Georgiou): obmalloc,structmodule:
64bit, big endian (issue 2 only).

This adds a bunch of memcpy calls via a temporary variable to avoid
alignment errors. That's needed for some platforms.

Excise DL_EXPORT/DL_IMPORT from Modules/*. Required adding a prototype

Fri, 02 Aug 2002 02:27:13 +0000

Excise DL_EXPORT/DL_IMPORT from Modules/*. Required adding a prototype
for Py_Main().

Thanks to Kalle Svensson and Skip Montanaro for the patches.

Patch #568124: Add doc string macros.

Thu, 13 Jun 2002 20:33:02 +0000

Patch #568124: Add doc string macros.

Get rid of more PyArg_Parse & METH_OLDARGS.

Tue, 02 Apr 2002 18:17:57 +0000

Get rid of more PyArg_Parse & METH_OLDARGS.
PyArg_Parse( "s" ) -> PyString_AsString
PyArg_Parse( "t#" ) -> PyString_AsStringAndSize

Fix typo.

Thu, 14 Feb 2002 07:16:30 +0000

Fix typo.

Use PyModule_AddObject() instead of accessing the module dict directly.

Thu, 14 Feb 2002 07:11:23 +0000

Use PyModule_AddObject() instead of accessing the module dict directly.

SF bug 705836: struct.pack of floats in non-native endian order

Thu, 20 Mar 2003 18:31:20 +0000

Backport 2.57 from trunk:

Mon, 23 Sep 2002 20:54:04 +0000

Backport 2.57 from trunk:

(Most of) SF patch 601369 (Christos Georgiou): obmalloc,structmodule:
64bit, big endian (issue 2 only).

This adds a bunch of memcpy calls via a temporary variable to avoid
alignment errors. That's needed for some platforms.

SF patch #474590 -- RISC OS support

Wed, 24 Oct 2001 20:42:55 +0000

SF patch #474590 -- RISC OS support

The 'p' (Pascal string) pack code acts unreasonably when the string size

Sat, 15 Sep 2001 02:35:15 +0000

The 'p' (Pascal string) pack code acts unreasonably when the string size
and count exceed 255. Changed to preserve as much of the string as
possible (instead of count%256 characters).

SF bug #442520: test_struct fails on SPARC.

Wed, 18 Jul 2001 20:47:31 +0000

SF bug #442520: test_struct fails on SPARC.
The ob_sval member of a string object isn't necessarily aligned to better
than a native long, so the new "q" and "Q" struct codes can't get away w/
casting tricks on platforms where LONG_LONG requires stricter-than-long
alignment. After I thought of a few elaborate workarounds, Guido bashed
me over the head with the obvious memcpy approach, herewith implemented.

The new {b,l}p_{u,}longlong() didn't check get_pylong()'s return for NULL.

Wed, 13 Jun 2001 01:26:35 +0000

The new {b,l}p_{u,}longlong() didn't check get_pylong()'s return for NULL.
Repaired that, and added appropriate tests for it to test_struct.py.

Added q/Q standard (x-platform 8-byte ints) mode in struct module.

Tue, 12 Jun 2001 01:22:22 +0000

Added q/Q standard (x-platform 8-byte ints) mode in struct module.
This completes the q/Q project.

longobject.c _PyLong_AsByteArray: The original code had a gross bug:
the most-significant Python digit doesn't necessarily have SHIFT
significant bits, and you really need to count how many copies of the sign
bit it has else spurious overflow errors result.

test_struct.py: This now does exhaustive std q/Q testing at, and on both
sides of, all relevant power-of-2 boundaries, both positive and negative.

NEWS: Added brief dict news while I was at it.

Trimmed trailing whitespace.

Mon, 11 Jun 2001 16:57:33 +0000

Trimmed trailing whitespace.

Simplify some convolution by simply not recognizing 'q' and 'Q' at all

Mon, 11 Jun 2001 16:51:56 +0000

Simplify some convolution by simply not recognizing 'q' and 'Q' at all
in native mode on platforms that don't HAVE_LONG_LONG.

Make clear in the docstring that "std" applies to both size and alignment,

Mon, 11 Jun 2001 16:45:33 +0000

Make clear in the docstring that "std" applies to both size and alignment,
not just to alignment. Spotted by Guido.

Bugfix candidate.

Initial support for 'q' and 'Q' struct format codes: for now, only in

Sun, 10 Jun 2001 23:40:19 +0000

Initial support for 'q' and 'Q' struct format codes: for now, only in
native mode, and only when config #defines HAVE_LONG_LONG. Standard mode
will eventually treat them as 8-byte ints across all platforms, but that
likely requires a new set of routines in longobject.c first (while
sizeof(long) >= 4 is guaranteed by C, there's nothing in C we can rely
on x-platform to hold 8 bytes of int, so we'll have to roll our own;
I'm thinking of a simple pair of conversion functions, Python long
to/from sized vector of unsigned bytes; that may be useful for GMP
conversions too; std q/Q would call them with size fixed at 8).

test_struct.py: In addition to adding some native-mode 'q' and 'Q' tests,
got rid of unused code, and repaired a non-portable assumption about
native sizeof(short) (it isn't 2 on some Cray boxes).

libstruct.tex: In addition to adding a bit of 'q'/'Q' docs (more needed
later), removed an erroneous footnote about 'I' behavior.

backport tim's 2.44

Thu, 01 Nov 2001 13:34:10 +0000

backport tim's 2.44
Make clear in the docstring that "std" applies to both size and alignment,
not just to alignment. Spotted by Guido.

not normally bothering with docstring cleanups, but in this case Tim _did_
note it as a bugfix candidate, so I'll be nice :)

Repair portability of sign extension when reading signed ints on boxes

Sun, 08 Apr 2001 23:39:38 +0000

Repair portability of sign extension when reading signed ints on boxes
where sizeof(long)==8. This *was* broken on boxes where signed right
shifts didn't sign-extend, but not elsewhere. Unfortunately, apart
from the Cray T3E I don't know of such a box, and Guido has so far
refused to buy me any Cray machines for home Python testing .

More immediately interesting would be if someone could please test
this on *any* sizeof(long)==8 box, to make sure I didn't break it.

Rationalize use of limits.h, moving the inclusion to Python.h.

Tue, 26 Sep 2000 05:46:01 +0000

Rationalize use of limits.h, moving the inclusion to Python.h.
Add definitions of INT_MAX and LONG_MAX to pyport.h.
Remove includes of limits.h and conditional definitions of INT_MAX
and LONG_MAX elsewhere.

This closes SourceForge patch #101659 and bug #115323.

Use symbolic constants for allowable short ranges.

Fri, 15 Sep 2000 08:10:33 +0000

Use symbolic constants for allowable short ranges.

Check range for bytes and shorts. Closes bug #110845.

Fri, 15 Sep 2000 07:31:57 +0000

Check range for bytes and shorts. Closes bug #110845.

REMOVED all CWI, CNRI and BeOpen copyright markings.

Fri, 01 Sep 2000 23:29:29 +0000

REMOVED all CWI, CNRI and BeOpen copyright markings.
This should match the situation in the 1.6b1 tree.

merge Include/my*.h into Include/pyport.h

Mon, 31 Jul 2000 15:28:04 +0000

merge Include/my*.h into Include/pyport.h

marked my*.h as obsolete

Bunch of minor ANSIfications: 'void initfunc()' -> 'void initfunc(void)',

Fri, 21 Jul 2000 06:00:07 +0000

Bunch of minor ANSIfications: 'void initfunc()' -> 'void initfunc(void)',
and a couple of functions that were missed in the previous batches. Not
terribly tested, but very carefully scrutinized, three times.

All these were found by the little findkrc.py that I posted to python-dev,
which means there might be more lurking. Cases such as this:

long
func(a, b)
long a;
long b; /* flagword */
{

and other cases where the last ; in the argument list isn't followed by a
newline and an opening curly bracket. Regexps to catch all are welcome, of
course ;)

ANSI-fication

Mon, 10 Jul 2000 12:29:26 +0000

ANSI-fication

Nuke all remaining occurrences of Py_PROTO and Py_FPROTO.

Sun, 09 Jul 2000 03:09:57 +0000

Nuke all remaining occurrences of Py_PROTO and Py_FPROTO.

Change copyright notice - 2nd try.

Fri, 30 Jun 2000 23:58:06 +0000

Change copyright notice - 2nd try.

Change copyright notice.

Fri, 30 Jun 2000 23:50:40 +0000

Change copyright notice.

Michael Hudson :

Thu, 01 Jun 2000 02:02:46 +0000

Michael Hudson :
Removed PyErr_BadArgument() calls and replaced them with more useful
error messages.

Massive patch by Skip Montanaro to add ":name" to as many

Tue, 29 Feb 2000 13:59:29 +0000

Massive patch by Skip Montanaro to add ":name" to as many
PyArg_ParseTuple() format string arguments as possible.

Add DL_EXPORT() to all modules that could possibly be used

Fri, 04 Dec 1998 18:50:17 +0000

Add DL_EXPORT() to all modules that could possibly be used
on BeOS or Windows.

Patches from Greg Stein to support 'P' format in struct module's

Fri, 18 Sep 1998 14:14:13 +0000

Patches from Greg Stein to support 'P' format in struct module's
native format, as void* (translated to Python int or long).
Also adds PyLong_FromVoidPtr and PyLong_AsVoidPtr to longobject.c.

Unsigned 1 and 2 byte sized formats shouldn't result in long integer values!

Mon, 29 Jun 1998 04:00:40 +0000

Unsigned 1 and 2 byte sized formats shouldn't result in long integer values!

get_long(): Use PyErr_ExceptionMatches(...) instead of PyErr_Occurred(...).

Thu, 28 May 1998 04:35:49 +0000

get_long(): Use PyErr_ExceptionMatches(...) instead of PyErr_Occurred(...).

Make new gcc -Wall happy

Fri, 10 Apr 1998 22:27:42 +0000

Make new gcc -Wall happy

Oops, left in a non-standard multi-line doc string that GCC finds okay

Sun, 21 Dec 1997 06:46:20 +0000

Oops, left in a non-standard multi-line doc string that GCC finds okay
but other compilers don't like.

Added doc strings, suggested by Charles G. Waldman (but massaged quite a bit).

Fri, 19 Dec 1997 04:24:24 +0000

Added doc strings, suggested by Charles G. Waldman (but massaged quite a bit).

Fix due to Bill Noon for problem discovered by Ken Manheimer: packing

Tue, 04 Nov 1997 17:12:33 +0000

Fix due to Bill Noon for problem discovered by Ken Manheimer: packing
0.0 as float or double would yield the representation for 1.0!

Apply two changes, systematically:

Wed, 01 Oct 1997 04:29:29 +0000

Apply two changes, systematically:

(1) Use PyErr_NewException("module.class", NULL, NULL) to create the
exception object.

(2) Remove all calls to Py_FatalError(); instead, return or
ignore the errors -- the import code now checks PyErr_Occurred()
after calling a module's init function, so it's no longer a
fatal error for the initialization to fail.

Also did some small cleanups, e.g. removed unnecessary test for
"already initialized" from initfpectl(), and unified
initposix()/initnt().

I haven't checked this very thoroughly, so while the changes are
pretty trivial -- beware of untested code!

Contribution by Hannu Krosing (with some changes).

Fri, 05 Sep 1997 07:08:39 +0000

Contribution by Hannu Krosing (with some changes).

Added 'p' format character for Pascal string (i.e. leading length
byte). This uses the count prefix line 's' does, except that the
count includes the length byte; i.e. '10p' takes 10 bytes packed but
has space for a length byte and 9 data bytes.

Ignore whitespace between formats (not internal to a count+format).

Tue, 26 Aug 1997 20:39:54 +0000

Ignore whitespace between formats (not internal to a count+format).

Keep Microsoft VC happy.

Fri, 11 Apr 1997 20:44:04 +0000

Keep Microsoft VC happy.

Changed the ``add/sub_offset'' hacks for dealing with C's unsigned

Fri, 03 Jan 1997 19:08:16 +0000

Changed the ``add/sub_offset'' hacks for dealing with C's unsigned
int/long types, and use the new PyLong_FromUnsignedLong() and
PyLong_AsUnsignedLong() interfaces instead.

Semantic change: the 'I' format will now always return a long int.

Fill pad bytes with zeros (fixing a bug dating from the very first version!).

Fri, 03 Jan 1997 15:40:33 +0000

Fill pad bytes with zeros (fixing a bug dating from the very first version!).

Scratch the ears of gcc -Wall.

Fri, 03 Jan 1997 00:26:28 +0000

Scratch the ears of gcc -Wall.

Ok, ok, I've fixed gradual underflow on packing too.

Thu, 02 Jan 1997 23:23:20 +0000

Ok, ok, I've fixed gradual underflow on packing too.

Still don't know what to do with Inf/NaN, so I raise an exception on
pack(), and something random decided by ldexp() will happen on
unpack().

Oops -- unpack float/double didn't do the right thing if e==0.

Thu, 02 Jan 1997 22:31:07 +0000

Oops -- unpack float/double didn't do the right thing if e==0.

Support float and double in non-native formats.

Thu, 02 Jan 1997 22:21:36 +0000

Support float and double in non-native formats.
These use the ANSI/IEEE standard, which is also used by XDR;
so the _xdr module may become obsolete.