Andi Gutmans [Tue, 5 Nov 2002 18:25:09 +0000 (18:25 +0000)]
- ATTENTION: Finally nuke old_function and cfunction. I think it is time
- to get rid of these BC notations. This is from the days of the move from
- PHP/FI 2 -> PHP 3
Andi Gutmans [Tue, 5 Nov 2002 18:16:11 +0000 (18:16 +0000)]
- Change the automatically created variable $clone in __clone() to
- $that as discussed at the PHP Conference. If there are any objections
- alternative names please let me know. The reason for changing it from
- $clone is because $clone sounds as if it's the newly cloned object and
- not the old one.
Fixed a memory leak in array_fill().
The refcount hack is now done for ZE1 only and is slightly faster then the
original.
After this patch array_fill() can consistently create arrays with >65k elements.
Marcus Boerger [Tue, 5 Nov 2002 15:15:51 +0000 (15:15 +0000)]
call php_module_startup after zts has been initialised. This way we could
have called zend_ini_deactivate(TSRMLS_C). But it seems unnecessary.
#tested with ZTS and non ZTS build and many manual fail points.
#now it even compiles without any warning, at least for me
Marcus Boerger [Tue, 5 Nov 2002 15:01:20 +0000 (15:01 +0000)]
-use of emalloc functions
-info was not initialised fulle due to memset parameters
now that the whole structure is initialised we can stop putting 0 in it.
Added 4th argument to virtual_file_ex() that specifies whether or not
realpath() should be used during path resolving. In a number of functions
we do not want to use realpath(), since realpath() will resolve symlinks.
Marcus Boerger [Tue, 5 Nov 2002 11:17:45 +0000 (11:17 +0000)]
silence
#when i get this code right tmp_line is not initialised, too.
#Therefore php_stream_wrapper_log_error() will be called with
#undefined value, will it?
Timm Friebe [Tue, 5 Nov 2002 08:29:57 +0000 (08:29 +0000)]
@- Implemented features/changes requested in Bug #16960 (Timm):
@ . Added a new function sybase_unbuffered_query()
@ . Added a new function sybase_fetch_assoc()
@ . Added sybase_set_message_handler() which enables users to handle
@ server messages in a callback function
@ . Added an ini entry for deadlock retries - retrying deadlocks
@ can cause transaction state to break (sybct.deadlock_retry_count,
@ defaults to -1 "forever").
@ . Fixed sybase_fetch_object() not to return objects with numeric
@ members
@ . Fixed issues with identical fieldnames
@ . Made sybase_fetch_*() functions return correct datatypes
@ . Made phpinfo() section more verbose
@ . Made sybase_query() error messages more verbose
Stig Bakken [Tue, 5 Nov 2002 06:05:48 +0000 (06:05 +0000)]
Added separate functions for setting include_path, for environments
where ini_set has been disabled. New functions: get_include_path(),
set_include_path(), restore_include_path()
Andi Gutmans [Mon, 4 Nov 2002 20:45:28 +0000 (20:45 +0000)]
- Clean up mess made by netware guys. There's no reason not to abstract
- the macro a tiny bit more allowing the code to look cleaner. The more
- #ifdef's code has the more unreadable it becomes.
Marcus Boerger [Mon, 4 Nov 2002 10:43:22 +0000 (10:43 +0000)]
-testing the tests
-moved tests/dirname.phpt to tests/run-test/test007.phpt because many
tests use this to find their additional files.
-no more tests in tests directly only in subdirectories
Fixed ini argument parsing. We now support ini options with = in the argument,
such as the one being used in session test #21, which is once again working
properly.
Fixed bug #20217. When php_flag & php_value are used, the PHP_INI_PERDIR
flag should be used when setting the ini value. Otherwise, the user cannot
modify per-directory ini settings specified via .htaccess.