TLS is already used in TSRM, the way exporting the tsrm cache through
a thread local variable is not portable. Additionally, the current
patch suffers from bugs which are hard to find, but prevent it to
be worky with apache. What is done here is mainly uses the idea
from the RFC patch, but
- __thread variable is removed
- offset math and declarations are removed
- extra macros and definitions are removed
What is done merely is
- use an inline function to access the tsrm cache. The function uses
the portable tsrm_tls_get macro which is cheap
- all the TSRM_* macros are set to placebo. Thus this opens the way
remove them later
Except that, the logic is old. TSRMLS_FETCH will have to be done once
per thread, then tsrm_get_ls_cache() can be used. Things seeming to be
worky are cli, cli server and apache. I also tried to enable bz2
shared and it has worked out of the box. The change is yet minimal
diffing to the current master bus is a worky start, IMHO. Though will
have to recheck the other previously done SAPIs - embed and cgi.
The offsets can be added to the tsrm_resource_type struct, then
it'll not be needed to declare them in the userspace. Even the
"done" member type can be changed to int16 or smaller, then adding
the offset as int16 will not change the struct size. As well on the
todo might be removing the hashed storage, thread_id != thread_id and
linked list logic in favour of the explicit TLS operations.
* master: (48 commits)
change locale - looks like not everybody has sl_SI
Fix bug #66921 - Wrong argument type hint for function intltz_from_date_time_zone
fix format
Fix bug #67052 (NumberFormatter::parse() resets LC_NUMERIC setting)
Make sure the generator script also creates a newline at the end of file
Add newline at end of file to prevent compilation warning
Fix handling of session user module custom handlers.
Reference bug report instead of github issue in NEWS file
add more exts for Travis
Update NEWS
Fix phpdbg.1 man page installation when build != src directory
BFN for bug #67551 (php://input temp file will be located in sys_temp_dir instead of upload_tmp_dir)
reorder
restore API compatibility
finish
refactor php_stream_temp_create{,_ex} and use it for the php://input stream
refactor _php_stream_fopen_{temporary_,tmp}file()
fix length overflow of HTTP_RAW_POST_DATA
Update NEWS
Fixed bug #67215 (php-cgi work with opcache, may be segmentation fault happen)
...
Conflicts:
ext/opcache/zend_accelerator_util_funcs.c
ext/session/mod_user.c
ext/spl/spl_array.c
ext/spl/spl_dllist.c
ext/standard/file.c
ext/standard/streamsfuncs.c
ext/standard/string.c
main/streams/memory.c
* PHP-5.6:
BFN for bug #67551 (php://input temp file will be located in sys_temp_dir instead of upload_tmp_dir)
reorder
restore API compatibility
finish
refactor php_stream_temp_create{,_ex} and use it for the php://input stream
refactor _php_stream_fopen_{temporary_,tmp}file()
fix length overflow of HTTP_RAW_POST_DATA
Conflicts:
main/php_content_types.c
* master: (41 commits)
Update copyright year to 2014
Update copyright year to 2014
Update copyright year to 2014
Update copyright year to 2014
Update copyright year to 2014
Update copyright year to 2014
Update copyright year to 2014
NEWS
Fix Request #67453 Allow to unserialize empty data.
Update copyright year to 2014
Update copyright year for re2c generated files
Update copyright year to 2014
Update copyright year for re2c files as well
Fix patch for bug #67436
fix failed test
Fix test on modern distro where old unsecure algo are disabled in openssl config. Testing recent algo should be enough to check this function.
Added tests for bug 67436
Fixed wrong XFAIL test - already fixed
Fix typo in Bug #67406 NEWS entry
Fix typo in Bug #67406 NEWS entry
...
Conflicts:
Zend/zend_compile.c
ext/session/session.c
ext/standard/array.c
ext/standard/http_fopen_wrapper.c
tests/classes/bug63462.phpt
* master: (41 commits)
fix test - output can be chunked
fix test
fix test
Fixed test for commit 997be125eb
Add bug fix to NEWS
Update UPGRADING according to bug fix
fix test
improve CURL tests to allow testing without separate server
improve CURL tests to allow testing without separate server
Fixed bug #67199 mb_regex_encoding mishmash
Fix bug #67248 (imageaffinematrixget missing check of parameters)
Fix bug #67247 spl_fixedarray_resize integer overflow
fix news
add tests stuff to README
Updated NEWS
Fix Linux specific fail in error traces (cherry-picked and fix for bug #67245) Linux apparently does not like memcpy in overlapping regions...
- Fixed off-by-one in phar_build (patch by crrodriguez at opensuse dot org)
- Move checking
- Fixed missing NULL check in SimpleXMLElement::xpath()
- Fixed missing NULL check
...
Conflicts:
ext/bz2/bz2.c
ext/gd/gd.c
ext/mbstring/php_mbregex.c
ext/session/tests/031.phpt
ext/simplexml/simplexml.c
ext/spl/spl_fixedarray.c
Headers: forbid \r and \n also after \0, allow CRLF followed by HT or SP and forbid \0. See bug #60227.
Conflicts:
ext/standard/tests/general_functions/bug60227.phpt
ext/standard/tests/general_functions/bug60227_1.phpt
ext/standard/tests/general_functions/bug60227_2.phpt
main/SAPI.c
to be called as all the headers are being sent and after all
of the default headers have been merged.
headers_list(), header_remove() and header() can all be used
inside the callback.
<?php
header('Content-Type: text/plain');
header('X-Test: foo');
function foo() {
foreach (headers_list() as $header) {
if (strpos($header, 'X-Powered') !== false) {
header_remove('X-Powered-By');
}
header_remove('X-Test');
}
}
$result = header_register_callback('foo');
echo "a";
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.
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.
requests--let's see what I can dig out of the bugtracker for NEWS--
and while crossing the road:
* implemented new zlib API
* fixed up ext/tidy (what was "s&" in zend_parse_parameters() supposed to do?)
Thanks to Jani and Felipe for pioneering.
* Removed ini options, safe_mode*
* Removed --enable-safe-mode --with-exec-dir configure options on Unix
* Updated extensions, SAPI's and core
* php_get_current_user() is now declared in main.c, thrus no need to include safe_mode.h anymore
I am sorry I tried fixing PHP without extensive discussion on the mailing list.
I am sorry I tried fixing PHP without extensive discussion on the mailing list.
Hope all the relevant parties are satisfied.
[DOC] proto void header_remove([string header_name])
Removes an HTTP header previously set using header()
The header_name parameter is optionnal, all headers are
removed if it is not set
[SAPIs] The header_handler callback in sapi_module_struct has
been changed, it now take a new argument.
When it is set to SAPI_HEADER_DELETE, sapi_header->header is
the name of an header, header_handler has to delete it.
When it is set to SAPI_HEADER_DELETE_ALL, header_handler has
to delete all headers.
When sapi_header_op_enum is SAPI_HEADER_ADD or _REPLACE,
sapi_header->header is in the form "Name: value", header_handler
has to add or replace the given header.
In all cases, header_handler must not free sapi_header or
sapi_header->header. SAPI_HEADER_ADD must be returned if the
header has been added or replaced, or 0 in other cases.
After the sigsetjmp change, this is patch #2 in an effort to get some
sanity restored to signal handling in PHP.
This patch does two things. First, it makes it possible to reset the
timeout without resetting the signal handlers. This is important for
cases where an extension may have deferred signals in its MINIT in order
to implement critical sections. It also lays the groundwork for cleaning
up our signal handling and perhaps eventually implementing our own
signal deferring mechanism so we can have true critical sections.
The second thing this does is to make it possible to terminate the current
child process (only for Apache1 at the moment) on a timeout. There are
a number of extensions that are unhappy about being longjmp'ed out of
and when this happens on a timeout they are left in an inconsistent state.
By turning on exit_on_timeout you can now force the process to terminate
on a timeout which will clean up any hanging locks and/or memory left
hanging after the longjmp.
