- Introduce php_open_temporary_file(), in place of tempnam(). Still
needs testing under UNIX (mkstemp()), works reliably under Windows now.
- Reimplement the mechanism for unlinking uploaded files at the end of the request
(was it ever tested?). Files moved with move_uploaded_file() will not be unlink()'d
again, to avoid (albeit very unlikely) race conditions.
SAPI and down into the individual SAPI modules. I have made the
appropriate changes in all the SAPI modules, but please verify these.
The reason for this change is that Apache sometimes will feed PHP
a request_method of GET but have r->header_only set to true. This happens
in an ErrorDocument redirect. In this same scenario we want to preserve
the status code as well instead of just overwriting it with a 200 and
losing this information. For now the other sapi modules act exactly as
before since they probably do not make this distinction, and they may
not even have a valid response code this early in the request.
@ Fix HEAD request bug on an Apache ErrorDocument redirect and preserve
@ the status code across the redirect as well. (Rasmus)
zend_get_parameters(), but it became invalid when someone switched it
to zend_get_parameters_ex(), and I decided to remove it anyway).
- Remove a redundant function call
[6] NEVER USE strncat(). If you're absolutely sure you know what you're doing,
check its man page again, and only then, consider using it, and even then,
try avoiding it.
strncat() is your enemy!
- Fix several SAPI services, get rid of the default_content_type (it's always
composed of the mime type and charset now).
- Win32 works again
stig, in sgi-mode i still "only" get:
(haven't touched my .ini file in ages - and don't plan to)
X-Powered-By: PHP/4.0b5-dev
Content-type: iso-8859-1 <<< where is text/html?
have in the future now obey to the variables_order directive, and there's a real way modular
way to handle POST content.
This is all untested, BEFORE_SAPI_POST_PATCH_17_FEB_2000 tagged before submission
@- Made multipart/form-data content obey to the variables_order directive (Zeev)
@ an ISAPI filter, only as an ISAPI extension, unless you wish to perform
@ authentication using PHP. This didn't yet get enough testing, but it
@ should work (Zeev)
- Fixed auth_user/auth_password memory leak (I didn't have time to test it under
Apache, feedback welcome!)
@- Finished the server abstraction layer; All of the PHP code is now shared
@ across different servers (Apache, CGI, IIS, etc.), except for thin
@ interface modules (Zeev)
@- If header information is sent after output has already been sent, the warning
@ message will now state the filename and line number at which the first output
@ was made (Zeev)
- Got rid of the old flush() implemenetation in favour of the new one
- Added implicit_flush() support to the output buffering layer.
@- Added implicit_flush() to control whether flush() should be called
@ implicitly after any output (Zeev)
- improve genif.sh to also consider all header files for inclusion
(checks for phpext_)
- use vsnprintf in main.c to avoid buffer overflows
- improve sessions's mm module to cope better with OOM situations
within the shared memory segment
- fix typo wrt session.auto_start
* Fully implement ISAPI support - POST and cookies among other things.
* Almost completely rewrote phpinfo(). Allow modules to easily display their
information in phpinfo() without modifying phpinfo() itself (prototype for
the module info function was changed, thus the large amount of updated module
files).
* Initial extended SAPI support for Apache, completely untested.
* CGI now uses SAPI fully as well.
* Changed PHP4 to compile as a DLL, both ISAPI and the the CGI run with the same DLL.
* Switched to using the DLL runtime library under Win32. PHP will NOT work if
compiled against the static library!
* Removed yesterday's php4libts project (with php4dllts, it's obsolete).
This *does* affect thread-unsafe Windows as well - the thread unsafe CGI is also
dependant on the thread-unsafe DLL.