Commit Graph

157 Commits

Author SHA1 Message Date
Raymond Hettinger b1b915c796 Issue 18719: Remove a false optimization
Remove an unused early-out test from the critical path for
dict and set lookups.

When the strings already have matching lengths, kinds, and hashes,
there is no additional information gained by checking the first
characters (the probability of a mismatch is already known to
be less than 1 in 2**64).
2013-08-13 18:16:34 -07:00
Benjamin Peterson d2b58a9880 only recursively expand in the format spec (closes #17644) 2013-05-17 17:34:30 -05:00
Ezio Melotti 6b02772c13 Remove trailing whitespace. 2013-04-21 04:07:51 +03:00
Serhiy Storchaka 18809fa94e Remove unused defines. 2013-02-23 14:48:16 +02:00
Serhiy Storchaka 18ba40b945 Check for NULL before the pointer aligning in fastsearch_memchr_1char.
There is no guarantee that NULL is aligned.
2013-01-15 13:27:28 +02:00
Victor Stinner ab60de478d Issue #8271: Fix compilation on Windows 2012-11-04 23:59:15 +01:00
Ezio Melotti f7ed5d111b #8271: the utf-8 decoder now outputs the correct number of U+FFFD characters when used with the "replace" error handler on invalid utf-8 sequences. Patch by Serhiy Storchaka, tests by Ezio Melotti. 2012-11-04 23:21:38 +02:00
Antoine Pitrou ca8aa4acf6 Issue #15144: Fix possible integer overflow when handling pointers as integer values, by using Py_uintptr_t instead of size_t.
Patch by Serhiy Storchaka.
2012-09-20 20:56:47 +02:00
Victor Stinner b3f5501250 Close #15534: Fix a typo in the fast search function of the string library (_s => s)
Replace _s with ptr to avoid future confusion. Add also non regression tests.
2012-08-02 23:05:01 +02:00
Mark Dickinson 01ac8b6ab1 Use correct types for ASCII_CHAR_MASK integer constants. 2012-07-07 14:08:48 +02:00
Mark Dickinson 106c4145ff Issue #14923: Optimize continuation-byte check in UTF-8 decoding. Patch by Serhiy Storchaka. 2012-06-23 21:45:14 +01:00
Antoine Pitrou a759d4e9f4 Make private function static (from `make smelly`) 2012-06-21 17:26:28 +02:00
Antoine Pitrou 27f6a3b0bf Issue #15026: utf-16 encoding is now significantly faster (up to 10x).
Patch by Serhiy Storchaka.
2012-06-15 22:15:23 +02:00
Victor Stinner d7b7c7472b Issue #14993: Use standard "unsigned char" instead of a unsigned char bitfield 2012-06-04 22:52:12 +02:00
Victor Stinner d3f0882dfb Issue #14744: Use the new _PyUnicodeWriter internal API to speed up str%args and str.format(args)
* Formatting string, int, float and complex use the _PyUnicodeWriter API. It
   avoids a temporary buffer in most cases.
 * Add _PyUnicodeWriter_WriteStr() to restore the PyAccu optimization: just
   keep a reference to the string if the output is only composed of one string
 * Disable overallocation when formatting the last argument of str%args and
   str.format(args)
 * Overallocation allocates at least 100 characters: add min_length attribute
   to the _PyUnicodeWriter structure
 * Add new private functions: _PyUnicode_FastCopyCharacters(),
   _PyUnicode_FastFill() and _PyUnicode_FromASCII()

The speed up is around 20% in average.
2012-05-29 12:57:52 +02:00
Antoine Pitrou 63065d761e Issue #14624: UTF-16 decoding is now 3x to 4x faster on various inputs.
Patch by Serhiy Storchaka.
2012-05-15 23:48:04 +02:00
Antoine Pitrou ca5f91b888 Issue #14738: Speed-up UTF-8 decoding on non-ASCII data. Patch by Serhiy Storchaka. 2012-05-10 16:36:02 +02:00
Victor Stinner 3b1a74a9c3 Rename unicode_write_t structure and its methods to "_PyUnicodeWriter" 2012-05-09 22:25:00 +02:00
Victor Stinner ee4544c920 Issue #14744: Inline unicode_writer_write_char() and unicode_write_str()
Optimize also PyUnicode_Format(): call unicode_writer_prepare() only once
per argument.
2012-05-09 22:24:08 +02:00
Victor Stinner 202fdca133 Close #14716: str.format() now uses the new "unicode writer" API instead of the
PyAccu API. For example, it makes str.format() from 25% to 30% faster on Linux.
2012-05-07 12:47:02 +02:00
Antoine Pitrou d0acb411ef Issue #14387: Do not include accu.h from Python.h. 2012-03-22 14:42:18 +01:00
Victor Stinner 41a863cb81 Issue #13706: Fix format(int, "n") for locale with non-ASCII thousands separator
* Decode thousands separator and decimal point using PyUnicode_DecodeLocale()
   (from the locale encoding), instead of decoding them implicitly from latin1
 * Remove _PyUnicode_InsertThousandsGroupingLocale(), it was not used
 * Change _PyUnicode_InsertThousandsGrouping() API to return the maximum
   character if unicode is NULL
 * Replace MIN/MAX macros by Py_MIN/Py_MAX
 * stringlib/undef.h undefines STRINGLIB_IS_UNICODE
 * stringlib/localeutil.h only supports Unicode
2012-02-24 00:37:51 +01:00
Benjamin Peterson 21e0da228d remove some usage of Py_UNICODE_TOUPPER/LOWER 2012-01-11 21:00:42 -05:00
Victor Stinner 6099a03202 Issue #13624: Write a specialized UTF-8 encoder to allow more optimization
The main bottleneck was the PyUnicode_READ() macro.
2011-12-18 14:22:26 +01:00
Victor Stinner f8eac00779 Issue #13623: Fix a performance regression introduced by issue #12170 in
bytes.find() and handle correctly OverflowError (raise the same ValueError than
the error for -1).
2011-12-18 01:17:41 +01:00
Victor Stinner b37b17423b Replace PyUnicode_FromUnicode(NULL, 0) by PyUnicode_New(0, 0)
Create an empty string with the new Unicode API.
2011-12-01 03:18:59 +01:00
Antoine Pitrou 0a3229de6b Issue #13417: speed up utf-8 decoding by around 2x for the non-fully-ASCII case.
This almost catches up with pre-PEP 393 performance, when decoding needed
only one pass.
2011-11-21 20:39:13 +01:00
Victor Stinner 0fc35196bb stringlib: remove unused STRINGLIB_FILL 2011-11-20 19:30:15 +01:00
Victor Stinner 7931d9a951 Replace PyUnicodeObject type by PyObject
* _PyUnicode_CheckConsistency() now takes a PyObject* instead of void*
 * Remove now useless casts to PyObject*
2011-11-04 00:22:48 +01:00
Victor Stinner 9db1a8b69f Replace PyUnicodeObject* by PyObject* where it was irrevelant
A Unicode string can now be a PyASCIIObject, PyCompactUnicodeObject or
PyUnicodeObject. Aliasing a PyASCIIObject* or PyCompactUnicodeObject* to
PyUnicodeObject* is wrong
2011-10-23 20:04:37 +02:00
Antoine Pitrou ac65d96777 Issue #12170: The count(), find(), rfind(), index() and rindex() methods
of bytes and bytearray objects now accept an integer between 0 and 255
as their first argument.  Patch by Petri Lehtinen.
2011-10-20 23:54:17 +02:00
Antoine Pitrou 5b9f4c1539 Fix typo 2011-10-17 19:21:04 +02:00
Antoine Pitrou c198d0599b Add a comment explaining this heuristic. 2011-10-13 18:07:37 +02:00
Antoine Pitrou dda339e6d2 Simplify heuristic for when to use memchr 2011-10-13 17:58:11 +02:00
Antoine Pitrou dd4e2f0153 Issue #13155: Optimize finding the optimal character width of an unicode string 2011-10-13 00:02:27 +02:00
Victor Stinner d218bf14cc stringlib: Fix STRINGLIB_STR for UCS2/UCS4 2011-10-12 00:14:32 +02:00
Victor Stinner 8cc70dcf70 Fix fastsearch for UCS2 and UCS4
* If needle is 0, try (p[0] >> 16) & 0xff for UCS4
 * Disable fastsearch_memchr_1char() if needle is zero for UCS2 and UCS4
2011-10-11 23:22:22 +02:00
Antoine Pitrou 2c3b2302ad Issue #13134: optimize finding single-character strings using memchr 2011-10-11 20:29:21 +02:00
Martin v. Löwis c47adb04b3 Change PyUnicode_KIND to 1,2,4. Drop _KIND_SIZE and _CHARACTER_SIZE. 2011-10-07 20:55:35 +02:00
Antoine Pitrou 4574e62c6e Fix massive slowdown in string formatting with str.format.
Example:
./python -m timeit -s "f='{}' + '-' * 1024 + '{}'; s='abcd' * 16384" "f.format(s, s)"

-> before: 547 usec per loop
-> after: 13 usec per loop
-> 3.2: 22.5 usec per loop
-> 2.7: 12.6 usec per loop
2011-10-07 02:26:47 +02:00
Antoine Pitrou dbf697ae5c Fix compilation warnings under 64-bit Windows 2011-10-06 15:34:41 +02:00
Victor Stinner c3cec7868b Add asciilib: similar to ucs1, ucs2 and ucs4 library, but specialized to ASCII
ucs1, ucs2 and ucs4 libraries have to scan created substring to find the
maximum character, whereas it is not need to ASCII strings. Because ASCII
strings are common, it is useful to optimize ASCII.
2011-10-05 21:24:08 +02:00
Victor Stinner e57b1c0da1 Mark PyUnicode_FromUCS[124] as private 2011-09-28 22:20:48 +02:00
Martin v. Löwis d63a3b8beb Implement PEP 393. 2011-09-28 07:41:54 +02:00
Mark Dickinson c7d93b7614 Issue #1621: Fix undefined behaviour from signed overflow in datetime module hashes, array and list iterations, and get_integer (stringlib/string_format.h) 2011-09-25 15:34:32 +01:00
Mark Dickinson 36f27c995a Issue #1621: Fix undefined behaviour from signed overflow in get_integer (stringlib/formatter.h) 2011-09-24 19:11:53 +01:00
Eric V. Smith 12ebefc9d3 Closes #12579. Positional fields with str.format_map() now raise a ValueError instead of SystemError. 2011-07-18 14:03:41 -04:00
Jesus Cea 6159ee3cf5 MERGE: startswith and endswith don't accept None as slice index. Patch by Torsten Becker. (closes #11828) 2011-04-20 17:42:50 +02:00
Jesus Cea ac4515063c startswith and endswith don't accept None as slice index. Patch by Torsten Becker. (closes #11828) 2011-04-20 17:09:23 +02:00
Ezio Melotti 4969f709cc #11515: Merge with 3.1. 2011-03-15 05:59:46 +02:00