Pierre Joye [Thu, 4 Nov 2010 01:16:20 +0000 (01:16 +0000)]
- expose php_sys_(l)stat, doing will prevent me to have to provide pecl's builds for <5.3.3 and another set for 5.3.4+. Trunk remains the same, A(B/P)I can be broken there
- Fixed bug #53198 (changing INI setting "from" with ini_set did not have any
effect)
#Made "from" a proper INI setting and bound it to a global variable.
#Previously, it was simply read from the hash table with the parsed ini file
#by using cfg_get_string (I wonder why this mechanism still exists...)
Andrey Hristov [Fri, 29 Oct 2010 15:02:39 +0000 (15:02 +0000)]
- More features for the profiling, create aggregates and dump them
on file close.
- Also add a trace modifier to switch on and off the profiling.
- With additional compiler switch the profiling can be completely omitted,
of course it makes sense only when --enable-debug. Because otherwise
there is no tracing, thus no profiling.
- Added a fix for Windows for handling trace files on different devices
(special handing of ':' )
- Fixed bug #49687 (utf8_decode vulnerabilities and deficiencies in the number
of reported malformed sequences). (Gustavo)
#Made a public interface for get_next_char/utf-8 in trunk to use in utf8_decode.
#In PHP 5.3, trunk's get_next_char was copied to xml.c because 5.3's
#get_next_char is different and is not prepared to recover appropriately from
#errors.
- Implemented request #44164, zlib.output_compression is now implicitly
disabled when the header "Content-length" is set.
#One could argue that any output handler could change the size of the
#response, so this exception for zlib.output_compression is an
#inconsistency. However, zlib.output_compression is presented as a
#performance setting, whose value should have no effect on the
#correctness of the scripts. This was not the case. Setting the
#header "content-length" and enabling zlib.output_compression was
#a recipe for infringing section 4.4 of RFC 2616.
- Small optimization of the fix of bug #53071. It's not necessary to destroy
the debug info as it's not holding references anymore (the fix removed
the refcount increments).
Adam Harvey [Mon, 18 Oct 2010 02:10:29 +0000 (02:10 +0000)]
Fix bug #53089 (php.ini should use portable example of find) by using POSIX
compliant syntax in the suggested find command for cleaning up session files in
the shipped php.ini files.
- [DOC] Reverted rev #304382 and rev #304380, as I figured out a way to
fix the erratic behavior without breaking backwards compatibility. Namely,
$offset retains SEEK_SET behavior but actually SEEK_CUR is passed to
_php_stream_seek, if possible, by moving the offset stream->position bytes.
- Addresses bug #53006.
- Fixed forward stream seeking emulation in streams that don't support seeking
in situations where the read operation gives back less data than requested
and when there was data in the buffer before the emulation started. Also made
more consistent its behavior -- should return failure every time less data
than was requested was skipped.
- Small performance improvement by correcting off-by-one error that generate an
invalid call to the seek handler or read handler. in _php_stream_seek.
- Added a 3rd parameter to get_html_translation_table. It now takes a charset
hint, like htmlentities et al.
- Fixed bug #49407 (get_html_translation_table doesn't handle UTF-8).
- Fixed bug #25927 (get_html_translation_table calls the ' ' instead of
').
- Fixed tests for get_html_translation_table and unified the Windows and
non-Windows versions of the tests.
- Removed redundant local variable in dns_get_record.
- (5.3) Fixed bug in the Windows implementation of
dns_get_record, where the two last parameters wouldn't be
filled unless the type were DNS_ANY (Gustavo).
- Fixed a typo in rev #304208 (24 instead of 34/'"').
- Improved the test bug53021.phpt to reflect other fixes in rev #304208.
- Updated NEWS to reflect other fixes in rev #304208.
- mysqli.c: Fix unused variables, they are only used in non-mysqlnd mode
- mysqli_api.c (PHP_5_3 only): Fix constness, add_property_string expects a char *, not a const char *
- mysqli_prop.c: Cast to long, as its below the LONG_MAX and therefore safe
- mysqli_result_iterator.c: Cast to ulong as the iterator member expects that rather than a my_longlong
# In trunk only warnings regarding the zend_property_info
# structure is present and PHP_5_3 is warning free now
- Fixed bug #52981 (Unicode casing table was out-of-date).
Updated with UnicodeData-6.0.0d7.txt and included the
source of the generator program with the distribution.
#The replaced tables, generated circa 2002, seem to reflect
#Unicode 3.2. I was unable to generate the same property
#offsets with Unicode 3.2 data, but all the tests I made
#indicate php_unicode_is_prop() is returning the correct
#values. The replaced file merely says it used a "modified
#version" of ucgendat, which is not very helpful. The results
#I got were not significantly different, only slightly higher
#offsets at two properties, which were carried over to the
#subsequent properties.
#I was, however, able to replicate precisely the casing table.
#The extent of the "modifications" besides omitting most of
#the tables, a slightly different layout and the casing table
#offsets having been multiplied by 3 is unclear.
#The test suite showed no regressions; however, it's very poor
#in testing the modified portion of the extension.
Ulf Wendel [Mon, 4 Oct 2010 10:02:58 +0000 (10:02 +0000)]
Sorry for the commit but compiler warnings are annoying and so are white spaces at the end of a line, if one uses an editor which removes them on save and diff's get blown
up by white space changes...
Changes:
- remove end of line white space
- fix compiler warning "ext/mysqli/mysqli_priv.h:156:27: warning: no newline at end of file"
- fix compiler warning "ext/mysqli/mysqli_fe.h:135:25: warning: no newline at end of file"
- fix compiler warning "ext/mysqli/mysqli.c:896:11: warning: extra tokens at end of #ifdef directive"
- Implemented FR #50692, not uploaded files don't count towards
max_file_uploads limit.
- As a side improvement, temporary files are not opened for
empty uploads and, in debug mode, 0-length uploads.
Felipe Pena [Sun, 3 Oct 2010 16:01:38 +0000 (16:01 +0000)]
- Fixed bug #52971 (PCRE-Meta-Characters not working with utf-8)
# In PCRE, by default, \d, \D, \s, \S, \w, and \W recognize only ASCII
# characters, even in UTF-8 mode. However, this can be changed by setting
# the PCRE_UCP option.