This is documented to not be case sensitive and it is not case sensitive
in PHP 3, so let's not make it case sensitive.
@Make the special Header("http/...") response be case insensitive like 3.0 (Rasmus)
Dan Kalowsky [Fri, 30 Jun 2000 19:22:52 +0000 (19:22 +0000)]
updates for the SolidEE to work for 3.5 with PHP4.0.1. Still causes warning of "integer from pointer used without typecast" but does not seem to effect the final build any. have not found a correction yet. has not shown the SQLParamData bug either, but unknown if it is closed as i was unable to reproduce it.
basically fixed the config.m4 file, and added a #define of HAVE_SOLID_35, due to the solidee 3.5 vs all other solidee systems being drastically different. and solidtech providing no means for testing what version you are using.
does not include fixes for the linux users of glibc2 (automatically assumes libc5). temporary work around, link your /usr/local/solid/lib files from a scll2x23 or scll2x30 to a scllux30 or scllux23 name... suggestions to automate this are welcome.
Chuck Hagenbuch [Fri, 30 Jun 2000 19:20:03 +0000 (19:20 +0000)]
DB::parseDSN() now checks the value it is passed, and if it is already an
array, returns it. This lets you pass arrays to the connect functions and
they'll work transparently.
Rasmus Lerdorf [Fri, 30 Jun 2000 17:39:21 +0000 (17:39 +0000)]
Some user-agents send 'basic' instead of 'Basic' in their Authorization
header. ApacheBench is an example. PHP's HTTP Auth would not work with
these.
Some user-agents send 'basic' instead of 'Basic' in their Authorization
header. ApacheBench is an example. PHP's HTTP Auth would not work with
these. (Rasmus)
Zeev Suraski [Fri, 30 Jun 2000 13:23:08 +0000 (13:23 +0000)]
- Revert bogus time.h commit
- Fixed a problem under some versions of Windows that could cause PHP to hang
on startup if errors occured, e.g. in the php.ini file
Try to fix #5271 - crash when $HTTP_SESSION_VARS is already defined
via "global"
# I'm not sure it's fastest fix possible, but at least it should remove the
# crash
Zeev Suraski [Wed, 28 Jun 2000 18:27:13 +0000 (18:27 +0000)]
I'm not sure whether this is necessary or not. Is it guarenteed that get_module_conf()
will always return a value? Even if there were no relevant directives at all?
Jouni Ahto [Tue, 27 Jun 2000 21:36:26 +0000 (21:36 +0000)]
- dba_[p]open accepts now a new parameter, which kind of database to create
(DBA_BTREE or DBA_HASH), if the handler is either 'db2' or 'db3' and
mode 'c' or 'n'. It is ignored if mode is 'c' and the db already exists.
# Asked on the list 4 1/2 hours ago if anyone's got to say something to this,
# no response, so I'm assuming it's ok. And yes, I did test this.
Zeev Suraski [Mon, 26 Jun 2000 17:40:24 +0000 (17:40 +0000)]
- Please NEVER copy fopen_wrappers.c functions to modules. They are ugly
- enough as is and need cleaning up. Duplicating the mess makes things just
- worse and prevents bug fixes in the functions from propogating throughout
- PHP.
- This fix is untested so someone who uses zlib, please test it and report
- back.
If we freed it, we might not touch it
# BTW, this is not a real fix - we still might use it
# when reporting memory leaks, but at least we won't reference
# unitialized memory
Jouni Ahto [Mon, 26 Jun 2000 12:23:02 +0000 (12:23 +0000)]
- Test gdImagePaletteCopy and define HAVE_LIBGD15 if found.
- s/HAVE_LIBGD13/HAVE_LIBGD15/ in ImagePaletteCopy, ImageCopyMerge.
# This really should go into 4.0.1 or we'll get a lot of complaints from
# people using older versions of gd (because it still supports gif).