]> granicus.if.org Git - php/log
php
25 years agoadd installation
Sam Ruby [Sun, 30 Jan 2000 00:42:26 +0000 (00:42 +0000)]
add installation

25 years ago@- Fixed a crash bug in strtr() working on large input strings (Zeev)
Zeev Suraski [Sat, 29 Jan 2000 17:54:30 +0000 (17:54 +0000)]
@- Fixed a crash bug in strtr() working on large input strings (Zeev)

25 years agoProgress: sapi/servlet now builds again on Unix
Sam Ruby [Sat, 29 Jan 2000 17:53:58 +0000 (17:53 +0000)]
Progress: sapi/servlet now builds again on Unix

25 years agoFix ``'s
Zeev Suraski [Sat, 29 Jan 2000 17:49:05 +0000 (17:49 +0000)]
Fix ``'s

25 years agoFix require()
Zeev Suraski [Sat, 29 Jan 2000 17:19:47 +0000 (17:19 +0000)]
Fix require()

25 years agoFirst steps towards restoring sapi/servlet on Unix.
Sam Ruby [Sat, 29 Jan 2000 13:22:47 +0000 (13:22 +0000)]
First steps towards restoring sapi/servlet on Unix.
Cosmetic improvements to allow README to be read on systems with tabstop=8

25 years agoFix buglets in the argv/argc code
Zeev Suraski [Sat, 29 Jan 2000 12:57:08 +0000 (12:57 +0000)]
Fix buglets in the argv/argc code

25 years agoNo comment - will explain next week
Rasmus Lerdorf [Sat, 29 Jan 2000 12:46:59 +0000 (12:46 +0000)]
No comment - will explain next week

25 years ago- Change the argument order of php_register_variable() to something more
Zeev Suraski [Sat, 29 Jan 2000 11:55:44 +0000 (11:55 +0000)]
- Change the argument order of php_register_variable() to something more
  intuitive.
- Make the authentication variables be a part of the HTTP_SERVER_VARS[] array

25 years ago- Get rid of another rule which isn't needed.
Andi Gutmans [Sat, 29 Jan 2000 10:20:05 +0000 (10:20 +0000)]
- Get rid of another rule which isn't needed.

25 years ago- Add parser support for string offsets. This added three shift/reduce
Andi Gutmans [Sat, 29 Jan 2000 10:16:04 +0000 (10:16 +0000)]
- Add parser support for string offsets. This added three shift/reduce
  conflicts but they all seem to be fine.
- Cleaned up the parsing rules a bit and made them much more compact and
  elegant.
- Please CVS update and see that I didn't break anything.

25 years ago- This will save some memory w/ GCC compilers on some platforms
Andi Gutmans [Sat, 29 Jan 2000 09:14:22 +0000 (09:14 +0000)]
- This will save some memory w/ GCC compilers on some platforms

25 years ago- Yet another tiny optimization.
Andi Gutmans [Sat, 29 Jan 2000 00:33:07 +0000 (00:33 +0000)]
- Yet another tiny optimization.

25 years ago- Make loop a bit faster.
Andi Gutmans [Fri, 28 Jan 2000 22:23:28 +0000 (22:23 +0000)]
- Make loop a bit faster.

25 years ago- Make sure its use is understood.
Andi Gutmans [Fri, 28 Jan 2000 21:46:07 +0000 (21:46 +0000)]
- Make sure its use is understood.

25 years ago- Double the speed of some key switch() tests for Win32.
Andi Gutmans [Fri, 28 Jan 2000 21:43:46 +0000 (21:43 +0000)]
- Double the speed of some key switch() tests for Win32.

25 years ago- A few fixes
Zeev Suraski [Fri, 28 Jan 2000 18:29:37 +0000 (18:29 +0000)]
- A few fixes
- Added register_argv_argc directive to allow disabling of argv/argc

25 years agoRename gpc_globals to register_globals
Zeev Suraski [Fri, 28 Jan 2000 17:43:41 +0000 (17:43 +0000)]
Rename gpc_globals to register_globals

25 years agoApache module compile fixes
Zeev Suraski [Fri, 28 Jan 2000 17:38:08 +0000 (17:38 +0000)]
Apache module compile fixes

25 years agoUse sizeof on the proper string.
Andrei Zmievski [Fri, 28 Jan 2000 17:29:37 +0000 (17:29 +0000)]
Use sizeof on the proper string.

25 years agoTried to centralize global variable registration as much as possible:
Zeev Suraski [Fri, 28 Jan 2000 17:24:53 +0000 (17:24 +0000)]
Tried to centralize global variable registration as much as possible:

- Added $HTTP_ENV_VARS[] and $HTTP_SERVER_VARS[] support, which similarly
  to $HTTP_GET_VARS[], contain environment and server variables.  Setting
  register_globals to Off will now also prevent registration of the
  environment and server variables into the global scope (Zeev)
- Renamed gpc_globals to register_globals (Zeev)
- Introduced variables_order that deprecates gpc_order, and allows control
  over the server and environment variables, in addition to GET/POST/Cookies
  (Zeev)

25 years agoUse a more general and descriptive name
Zeev Suraski [Fri, 28 Jan 2000 14:57:19 +0000 (14:57 +0000)]
Use a more general and descriptive name

25 years agodsp updates
Zeev Suraski [Fri, 28 Jan 2000 14:46:19 +0000 (14:46 +0000)]
dsp updates

25 years ago- Make Win32 compile again
Andi Gutmans [Fri, 28 Jan 2000 13:53:06 +0000 (13:53 +0000)]
- Make Win32 compile again

25 years agopost.c really had nothing to do with POST anymore, and it belongs to the top level...
Zeev Suraski [Fri, 28 Jan 2000 13:31:12 +0000 (13:31 +0000)]
post.c really had nothing to do with POST anymore, and it belongs to the top level directory

25 years ago- This makes the switch() statement twice as quick. Moving to enum
Andi Gutmans [Fri, 28 Jan 2000 13:25:55 +0000 (13:25 +0000)]
- This makes the switch() statement twice as quick. Moving to enum
  might make this a general speed up for other platforms too

25 years agoGet rid of the old implementation
Zeev Suraski [Fri, 28 Jan 2000 12:53:05 +0000 (12:53 +0000)]
Get rid of the old implementation

25 years agoaccomodate old versions of kaffe which don't support jar
Sam Ruby [Fri, 28 Jan 2000 10:53:06 +0000 (10:53 +0000)]
accomodate old versions of kaffe which don't support jar

25 years agoShutdown cleanly (prevents Kaffe from trapping on exit)
Sam Ruby [Fri, 28 Jan 2000 00:32:28 +0000 (00:32 +0000)]
Shutdown cleanly (prevents Kaffe from trapping on exit)

25 years agoMcal cleanup, stage 1: it compiles again (the php3 syntax to php4/zend
Chuck Hagenbuch [Thu, 27 Jan 2000 16:38:45 +0000 (16:38 +0000)]
Mcal cleanup, stage 1: it compiles again (the php3 syntax to php4/zend
syntax conversion left a lot of: PHP_FUNCTION void ... definitions, giving
two types to every function).

However, it's still segfaulting. I'll look into that next.

25 years agoPHP now starts up in CGI mode with Kaffe on RedHat Linux 6.1.
Sam Ruby [Thu, 27 Jan 2000 13:44:45 +0000 (13:44 +0000)]
PHP now starts up in CGI mode with Kaffe on RedHat Linux 6.1.
status: segfault atexit, npe in jawt - but otherwise both samples work!
Installation needs more work

25 years agoForward progress, thanks to Sascha Schumann.
Sam Ruby [Thu, 27 Jan 2000 02:13:21 +0000 (02:13 +0000)]
Forward progress, thanks to Sascha Schumann.
libphp_java.so is now built again - it just refuses to load.
At least I now have something I can debug!

25 years agoAllow shared libraries to be built everywhere
Sascha Schumann [Wed, 26 Jan 2000 21:58:46 +0000 (21:58 +0000)]
Allow shared libraries to be built everywhere

25 years ago- Keep objects as references.
Andi Gutmans [Wed, 26 Jan 2000 21:10:46 +0000 (21:10 +0000)]
- Keep objects as references.

25 years ago- add another comment to the script
Uwe Steinmann [Wed, 26 Jan 2000 06:31:29 +0000 (06:31 +0000)]
- add another comment to the script

25 years ago- test script for domxml
Uwe Steinmann [Wed, 26 Jan 2000 06:22:44 +0000 (06:22 +0000)]
- test script for domxml

25 years ago- Allow is_ref to become 0 in case the refcount is back to 1.
Andi Gutmans [Wed, 26 Jan 2000 05:03:22 +0000 (05:03 +0000)]
- Allow is_ref to become 0 in case the refcount is back to 1.

25 years ago- resumed work on domxml and found a way to circumwent a segm fault
Uwe Steinmann [Tue, 25 Jan 2000 17:00:09 +0000 (17:00 +0000)]
- resumed work on domxml and found a way to circumwent a segm fault
  (see the source). No solution to fix it, yet. Probably needs some
  understanding of zend, which I don't have.
- added more functions

25 years agoProper indentation here also.
Egon Schmid [Mon, 24 Jan 2000 20:25:57 +0000 (20:25 +0000)]
Proper indentation here also.

25 years ago- Make foreach() now copy the array but use the original array. It can
Andi Gutmans [Mon, 24 Jan 2000 19:04:07 +0000 (19:04 +0000)]
- Make foreach() now copy the array but use the original array. It can
  still be optimized A LOT but it's only a performance issue and not
  a feature issue.

25 years agoNULL out the ticks callback until Stig implements it
Zeev Suraski [Mon, 24 Jan 2000 19:03:50 +0000 (19:03 +0000)]
NULL out the ticks callback until Stig implements it

25 years ago- Implement declare() with declarables framework
Zeev Suraski [Mon, 24 Jan 2000 19:00:30 +0000 (19:00 +0000)]
- Implement declare() with declarables framework
- Implement ticks - Germany&Norway - 5 points!

25 years agoFixes
Zeev Suraski [Mon, 24 Jan 2000 17:29:15 +0000 (17:29 +0000)]
Fixes

25 years agoFix an elusive bug
Zeev Suraski [Sat, 22 Jan 2000 17:02:48 +0000 (17:02 +0000)]
Fix an elusive bug

25 years agoAdd some order...
Zeev Suraski [Thu, 20 Jan 2000 14:26:31 +0000 (14:26 +0000)]
Add some order...

25 years agoIndentation fixes
Zeev Suraski [Thu, 20 Jan 2000 05:27:57 +0000 (05:27 +0000)]
Indentation fixes

25 years ago- Optimize zend_hash_del a tiny bit.
Andi Gutmans [Wed, 19 Jan 2000 19:47:16 +0000 (19:47 +0000)]
- Optimize zend_hash_del a tiny bit.

25 years ago- Hopefully fix the hash problem.
Andi Gutmans [Wed, 19 Jan 2000 19:44:32 +0000 (19:44 +0000)]
- Hopefully fix the hash problem.

25 years agoFix for wrong library ordering with adabas
Andreas Karajannis [Wed, 19 Jan 2000 18:21:53 +0000 (18:21 +0000)]
Fix for wrong library ordering with adabas

25 years ago- Hrm I'm not concentrating
Andi Gutmans [Wed, 19 Jan 2000 18:14:20 +0000 (18:14 +0000)]
- Hrm I'm not concentrating

25 years ago- Actually the destructor should run after the data is already detached
Andi Gutmans [Wed, 19 Jan 2000 18:12:05 +0000 (18:12 +0000)]
- Actually the destructor should run after the data is already detached
  from the hash but before the bucket is freed.

25 years ago- Rollback hash_apply and friends. They assume now that hash_del is reentrant
Andi Gutmans [Wed, 19 Jan 2000 18:05:16 +0000 (18:05 +0000)]
- Rollback hash_apply and friends. They assume now that hash_del is reentrant
  as it first applies the destructor and only later nukes the bucket

25 years ago- Run destructor before the hash structure is modified, thus, making
Andi Gutmans [Wed, 19 Jan 2000 17:39:55 +0000 (17:39 +0000)]
- Run destructor before the hash structure is modified, thus, making
  hash_del, reentrant (BLOCK_INTERRUPTIONS needs to be made a counter now).

25 years ago- Undo a bug we introduced. (Another one out there).
Andi Gutmans [Wed, 19 Jan 2000 17:31:33 +0000 (17:31 +0000)]
- Undo a bug we introduced. (Another one out there).

25 years agoFirst step to restoring the ability to build Java support on Unix:
Sam Ruby [Wed, 19 Jan 2000 14:45:26 +0000 (14:45 +0000)]
First step to restoring the ability to build Java support on Unix:
1) Restore the ability to build CGI as a shared library
2) Correct make dependency rule so that the jar files are included
3) Search for jar executable, and use it, if found

Left TODO: actually build the shared libraries

25 years agoRETURN_NULL -> RETURN_NULL() // we don't want macros without an argumnet
Thies C. Arntzen [Wed, 19 Jan 2000 09:38:45 +0000 (09:38 +0000)]
RETURN_NULL -> RETURN_NULL() // we don't want macros without an argumnet

25 years agoThis doesn't seem to be necessary, and causes a leak in case of output_buffering=on
Zeev Suraski [Tue, 18 Jan 2000 22:25:48 +0000 (22:25 +0000)]
This doesn't seem to be necessary, and causes a leak in case of output_buffering=on

25 years agoLeak fix
Zeev Suraski [Tue, 18 Jan 2000 22:18:17 +0000 (22:18 +0000)]
Leak fix

25 years agoMake the phpinfo() images work properly again
Zeev Suraski [Tue, 18 Jan 2000 19:15:45 +0000 (19:15 +0000)]
Make the phpinfo() images work properly again

25 years agoRETURN_NULL & RETVAL_NULL don't need ().
Thies C. Arntzen [Tue, 18 Jan 2000 08:40:45 +0000 (08:40 +0000)]
RETURN_NULL & RETVAL_NULL don't need ().

25 years ago@- Ora_GetColumn()/Ora_FetchInto() now return NULL for NULL-Columns. (Thies)
Thies C. Arntzen [Tue, 18 Jan 2000 08:35:55 +0000 (08:35 +0000)]
@- Ora_GetColumn()/Ora_FetchInto() now return NULL for NULL-Columns. (Thies)
also changed return-type for list-dtor

25 years ago- MS-SQL fix
Zeev Suraski [Mon, 17 Jan 2000 22:41:59 +0000 (22:41 +0000)]
- MS-SQL fix
- Fix a crash bug in session_register()

25 years agouse defines
Thies C. Arntzen [Mon, 17 Jan 2000 19:17:58 +0000 (19:17 +0000)]
use defines

25 years agoGet rid of the IsPointer functionality in the hash.
Zeev Suraski [Mon, 17 Jan 2000 18:45:46 +0000 (18:45 +0000)]
Get rid of the IsPointer functionality in the hash.

25 years ago- Fixes a newly introduced bug in the hash
Zeev Suraski [Mon, 17 Jan 2000 18:09:03 +0000 (18:09 +0000)]
- Fixes a newly introduced bug in the hash

25 years agoDestructors no longer return ints, the low level problem it was intended to solve...
Zeev Suraski [Mon, 17 Jan 2000 17:33:37 +0000 (17:33 +0000)]
Destructors no longer return ints, the low level problem it was intended to solve is long gone now...

25 years ago- Remove the pre-request-shutdown hooks, they should no longer be necessary.
Zeev Suraski [Sun, 16 Jan 2000 21:03:49 +0000 (21:03 +0000)]
- Remove the pre-request-shutdown hooks, they should no longer be necessary.
The session module may not work due to an unknown (and until now, IMO, misdiagnosed) problem,
if it does, please report it!

25 years ago- Make zend_hash_apply() (and friends) reentrant and much, much quicker
Zeev Suraski [Sun, 16 Jan 2000 20:59:03 +0000 (20:59 +0000)]
- Make zend_hash_apply() (and friends) reentrant and much, much quicker
- Introduce zend_hash_graceful_destroy(), which allows the destructor functions to
  use zend_hash_apply() and/or zend_hash_graceful_destroy()
- Switch to zend_hash_graceful_destroy() in the resource list shutdowns

25 years agoFix
Zeev Suraski [Sun, 16 Jan 2000 20:37:09 +0000 (20:37 +0000)]
Fix

25 years agoRemove the post-startup callback support. Modules are now started only after
Zeev Suraski [Sun, 16 Jan 2000 19:43:34 +0000 (19:43 +0000)]
Remove the post-startup callback support.  Modules are now started only after
the entire engine is started.  People - please make sure that the session module
works properly for you.

25 years agoAllow module startup to be separate from the compiler/executor startup
Zeev Suraski [Sun, 16 Jan 2000 19:41:15 +0000 (19:41 +0000)]
Allow module startup to be separate from the compiler/executor startup

25 years agomove call to php_execute_pre_request_shutdown() into php_request_shutdown()
Thies C. Arntzen [Sun, 16 Jan 2000 14:34:28 +0000 (14:34 +0000)]
move call to php_execute_pre_request_shutdown() into php_request_shutdown()
so it will be called when script terminated due to call to exit().

25 years ago@- OCI8 now supports binding of NULL-values. Module cleanups. (Thies)
Thies C. Arntzen [Sun, 16 Jan 2000 13:30:57 +0000 (13:30 +0000)]
@- OCI8 now supports binding of NULL-values. Module cleanups. (Thies)
lots of cleanup - all invisible to the user (i hope;-)
- full refcounting of connections & statements
- cleanup of the descriptor-code (descriptors are now real resources)
- we now use SQLT_CHR instead of SQLT_STR (no trailing \0) - smarter & faster
- ocifetchinto(OCI_BOTH) will use references.

25 years agofixes #3220, #3221 finally
Thies C. Arntzen [Sun, 16 Jan 2000 12:34:18 +0000 (12:34 +0000)]
fixes #3220, #3221 finally

25 years agonother go on #3220, #3221
Thies C. Arntzen [Sun, 16 Jan 2000 11:12:44 +0000 (11:12 +0000)]
nother go on #3220, #3221

25 years agohopfully fixes #3220, #3221
Thies C. Arntzen [Sun, 16 Jan 2000 10:51:55 +0000 (10:51 +0000)]
hopfully fixes #3220, #3221

25 years agomake the ht->inconsistent stuff less ugly:)
Thies C. Arntzen [Sun, 16 Jan 2000 09:45:10 +0000 (09:45 +0000)]
make the ht->inconsistent stuff less ugly:)

25 years ago(wddx_stack_destroy) Properly free the dangling varname.
Andrei Zmievski [Sat, 15 Jan 2000 23:34:51 +0000 (23:34 +0000)]
(wddx_stack_destroy) Properly free the dangling varname.
(php_wddx_deserialize_ex) There should be only 1 value on the stack after
                          parser is done.

25 years agoFix a bug in call_user_function_ex()
Zeev Suraski [Sat, 15 Jan 2000 22:52:24 +0000 (22:52 +0000)]
Fix a bug in call_user_function_ex()

25 years agoAnother idea I had today.
Andrei Zmievski [Sat, 15 Jan 2000 21:35:50 +0000 (21:35 +0000)]
Another idea I had today.

25 years ago(php_sockwait_for_data) timeout is not assigned yet, so check the right one.
Andrei Zmievski [Sat, 15 Jan 2000 20:39:06 +0000 (20:39 +0000)]
(php_sockwait_for_data) timeout is not assigned yet, so check the right one.

25 years ago(php_sockwait_for_data) clean up
Sascha Schumann [Sat, 15 Jan 2000 19:51:44 +0000 (19:51 +0000)]
(php_sockwait_for_data) clean up

25 years agoFix prototype.
Andrei Zmievski [Sat, 15 Jan 2000 19:42:40 +0000 (19:42 +0000)]
Fix prototype.

25 years agoAdded ability to set timeout on socket read operations. New function
Andrei Zmievski [Sat, 15 Jan 2000 19:41:17 +0000 (19:41 +0000)]
Added ability to set timeout on socket read operations. New function
set_socket_timeout(socket, seconds, microseconds) sets the timeout.
@- Added ability to set timeout on socket read operations through
@ set_socket_timeout() function. (Andrei)

25 years agoFix logic
Zeev Suraski [Sat, 15 Jan 2000 17:23:18 +0000 (17:23 +0000)]
Fix logic

25 years agoAdded support for $foo->{$bar}["foobar"] notation (was supported in PHP 3)
Zeev Suraski [Sat, 15 Jan 2000 17:17:32 +0000 (17:17 +0000)]
Added support for $foo->{$bar}["foobar"] notation (was supported in PHP 3)

25 years agoInitialize flag correctly
Sascha Schumann [Sat, 15 Jan 2000 17:16:21 +0000 (17:16 +0000)]
Initialize flag correctly

25 years agoCompile fixes for ZTS
Zeev Suraski [Sat, 15 Jan 2000 16:56:30 +0000 (16:56 +0000)]
Compile fixes for ZTS

25 years agoif ZEND_DEBUG mode is on we'll now see warnings when a HashTable is accessed
Thies C. Arntzen [Sat, 15 Jan 2000 13:40:17 +0000 (13:40 +0000)]
if ZEND_DEBUG mode is on we'll now see warnings  when a HashTable is accessed
while it's inconsistent.

Zeev, Andi - you welcome to revert this patch if you don't like it - i find it
useful! accesssing inconsistent hashtables is one of the hardest things to track!

25 years agojust add a remark. zend_hash_apply can & will cause problems here unless
Thies C. Arntzen [Sat, 15 Jan 2000 13:39:19 +0000 (13:39 +0000)]
just add a remark. zend_hash_apply can & will cause problems here unless

25 years agoin an rshutdown function we should not access the regular_list like this.
Thies C. Arntzen [Sat, 15 Jan 2000 13:35:06 +0000 (13:35 +0000)]
in an rshutdown function we should not access the regular_list like this.

25 years agoInitialize/destroy hook lists implicitly.
Sascha Schumann [Sat, 15 Jan 2000 13:30:32 +0000 (13:30 +0000)]
Initialize/destroy hook lists implicitly.

25 years agoBe more verbose, if headers were already sent and remove RSHUTDOWN handler
Sascha Schumann [Sat, 15 Jan 2000 13:20:39 +0000 (13:20 +0000)]
Be more verbose, if headers were already sent and remove RSHUTDOWN handler

25 years agosessions are now written via a _php_session_shutdown which is registered
Thies C. Arntzen [Sat, 15 Jan 2000 13:09:32 +0000 (13:09 +0000)]
sessions are now written via a _php_session_shutdown which is registered
via php_register_pre_request_shutdown().

25 years ago- don't set php_errormsg on errors that will cause a zend_bailout().
Thies C. Arntzen [Sat, 15 Jan 2000 13:02:54 +0000 (13:02 +0000)]
- don't set php_errormsg on errors that will cause a zend_bailout().
  using zend_hash_update() can make things worse in this situation.
- new function php_register_pre_request_shutdown(). this way modules
  can register callbacks that will be called as soon as execution of
  the script is done but *before* any cleanup (global symbol_table etc)
  has taken place.

25 years agoComplain verbosely, if a destroy failed
Sascha Schumann [Sat, 15 Jan 2000 12:17:18 +0000 (12:17 +0000)]
Complain verbosely, if a destroy failed

25 years agoSince we're highlighting code, put <code> and </code> around the code.
Andrei Zmievski [Fri, 14 Jan 2000 19:35:29 +0000 (19:35 +0000)]
Since we're highlighting code, put <code> and </code> around the code.

25 years ago@- Added implicit_flush INI directive (Zeev)
Zeev Suraski [Thu, 13 Jan 2000 20:39:33 +0000 (20:39 +0000)]
@- Added implicit_flush INI directive (Zeev)

25 years agoMake Win32 compile again
Zeev Suraski [Thu, 13 Jan 2000 20:34:45 +0000 (20:34 +0000)]
Make Win32 compile again

25 years agoWoops, forgot to commit the right version of the file...
Zeev Suraski [Thu, 13 Jan 2000 17:54:51 +0000 (17:54 +0000)]
Woops, forgot to commit the right version of the file...