Commit Graph

140 Commits

Author SHA1 Message Date
Sascha Schumann
ee619493b3 Use a simple IFS; set step to interprete the httpd -v output 2002-07-11 08:44:42 +00:00
foobar
2ca9ae687d - Fix the version check again. 2002-07-07 15:43:45 +00:00
foobar
999d9aa10f Better fix it here too.. 2002-06-29 00:33:02 +00:00
Sebastian Bergmann
836d9c3e9f Mention Apache 2.0.40. 2002-06-28 15:16:28 +00:00
foobar
21debac333 Require 2.0.40 2002-06-28 14:48:37 +00:00
Derick Rethans
2fda2eb4a8 - YAAC: I hope this is the last one! 2002-06-28 14:45:10 +00:00
foobar
f4699170fc Fixed bug: #17491 (honor INSTALL_ROOT) 2002-06-28 09:01:56 +00:00
Rasmus Lerdorf
890e01bebe @ Add runtime Apache2 thread check to ensure we don't run a non-threaded
@ PHP inside a threaded Apache2 MPM.  (Rasmus)
2002-06-24 16:04:20 +00:00
Rasmus Lerdorf
8a2ab01744 @ Turn off ZTS if Apache2 is using the prefork MPM. (Rasmus) 2002-06-24 14:51:01 +00:00
Sander Roobol
d279a92242 Made php_admin_value work outside <location> and <directory> blocks. 2002-06-22 15:31:44 +00:00
Stefan Esser
44a5b43b79 Fixing the same possible memory leak. 2002-06-21 14:47:59 +00:00
Sascha Schumann
5e148e3875 too many flushes are bad 2002-06-08 18:11:03 +00:00
Cliff Woolley
46eeabccf7 That macro is and always was hopelessly broken, which is why it's
now deprecated.
2002-05-30 06:07:28 +00:00
Cliff Woolley
8283979afc Allow the version checks for --with-apxs= and --with-apxs2= to work
with development version of Apache, whose version strings end in "-dev",
eg "Apache/2.0.37-dev".

PR: 17233
Submitted by: Dale Ghent <daleg@elemental.org>
2002-05-17 08:07:14 +00:00
Cliff Woolley
57a4a7ed39 apache 2.0's apache_lookup_uri() was returning an array rather than an
object, which contradicted both the documentation and the behavior of the
same function under apache 1.3.

PR: 14999
2002-05-17 06:32:04 +00:00
Cliff Woolley
78cac2a42e * restore the php_flag and php_admin_flag Apache directives which for
some mysterious reason never made their way from sapi/apache to
  sapi/apache2filter when it was first written  PR: 16629
* change the allowed locations of php_admin_value (and php_admin_flag to
  match) to ACCESS_CONF instead of OR_NONE to match sapi/apache.  No
  idea why it was ever OR_NONE.  PR: 16489
2002-05-17 05:16:18 +00:00
Marko Karppinen
5a5767e1d6 Patch by Justin Erenkrantz <jerenkrantz@apache.org> for enabling
--with-apxs2 build on Darwin. Omitting the change to start linking
with libtool for now, though.
2002-05-13 21:37:52 +00:00
Jim Jagielski
8006ee79a3 More verbose but more generic error message when we spot multiple
PHPINIDir directives
2002-05-13 19:22:27 +00:00
Sebastian Bergmann
5192f6432c Patch by Justin Erenkrantz <jerenkrantz@apache.org>. This should be MFHed. 2002-05-11 08:58:20 +00:00
Derick Rethans
dedf22f5ad - Added Aaron Bannert as maintainer 2002-05-08 05:17:59 +00:00
Aaron Bannert
8e176a10fc Fix a graceful restart SEGV. We no longer only perform initialization
on the second pass through the post_config. Now we only avoid the
initialization only on the first DSO load, and on all subsequent loads
we rerun the init code.
2002-05-05 18:15:33 +00:00
Aaron Bannert
b759322411 Fix an elusive and intermittent startup SEGV. The problem was
the static string we were using to set an initialization flag
would get remapped to a different location when Apache reloaded
the DSO, causing us to not run our initialization routines.

Submitted by: Justin Erenkrantz <jerenkrantz@apache.org>
Reviewed by: Aaron Bannert (I added the big comment too)
2002-05-05 18:11:41 +00:00
Marko Karppinen
b7e50782b7 Well that didn't last long! (Reverting previous.) 2002-05-05 09:52:39 +00:00
Marko Karppinen
32ff990c68 Prelim. patch to enable apache2filter to link on Darwin.
Submitted by: Justin Erenkrantz <jerenkrantz@apache.org>
Reviewed by: markonen
# A stopgap measure while we try to find a "real"
# solution, if one exists.
2002-05-05 09:40:31 +00:00
Sascha Schumann
12ba44827a be a bit more verbose about what is wrong 2002-04-26 21:26:47 +00:00
foobar
492efe963f - Fix for bug: #16791. (more reliable test) 2002-04-26 03:07:16 +00:00
Aaron Bannert
b2f9b6fb16 Apache does a full load, unload, load cycle for each DSO module.
This patch makes sure that any startup actions that are performed
for PHP don't happen until the second load (the second call to
the post_config hook), and it also prevents subsequent calls
to the initialization routines.

Suggested By:   Cliff Woolley

PR: 16475, 16754
2002-04-24 00:55:38 +00:00
Jim Jagielski
a6117a874e Typo in error string 2002-04-22 20:00:41 +00:00
Marko Karppinen
42555b189e Refine the OS X support a bit 2002-04-22 15:37:54 +00:00
Marko Karppinen
2cd85ac084 Merged the Mac OS X compile changes from the Apache 1.3 SAPI.
--with-apxs2 should now work on Mac OS X / Darwin.
# Not tested, though. So shoot me! :)
2002-04-22 15:25:17 +00:00
Edin Kadribasic
6c491a5654 Reverting my previous patch. 2002-04-22 12:00:53 +00:00
Edin Kadribasic
8509772d93 Added ..\..\..\php_build\lib\apache2 to libpath and
..\..\..\php_build\include\apache2 to includepath
2002-04-21 20:36:08 +00:00
Aaron Bannert
5ab179e830 Document the new PHPINIDir directive in the apache2filter/README. 2002-04-18 22:47:13 +00:00
Aaron Bannert
3aa8a9b62b This patch implements a new Apache2 directive called PHPINIDir that
allows the specification of the php.ini directory from within the Apache
configuration. If left unset, the default is to defer to the hard-coded
php paths. When set, the supplied path is made relative to Apache's
internal ServerRoot setting.

Example:
PHPINIDir "conf"
# PHP will now look in the ServerRoot/conf directory for the php.ini file
2002-04-18 22:10:57 +00:00
Aaron Bannert
d11ee7c1ea Fix an intermittent SEGV when an error bubbled up from PHP before our
server context was set. Now if that happens we simply don't log against
any particular server config (vhost).

Obtained from bug report by:  Balazs Nagy <js@iksz.hu>
2002-04-18 16:34:06 +00:00
Aaron Bannert
2e3ce44c80 It makes more sense to do the null-pointer check *before* trying to use it.
(Also fix a typo that Cliff pointed out: "safe" --> "save".)

Obtained from:  Ryan Morgan <rmorgan@covalent.net>
2002-04-16 17:41:28 +00:00
foobar
b21931e6c3 ws fixes 2002-04-14 12:38:33 +00:00
foobar
f81166f34e - Added checks to prevent building the DSO with wrong configure option. 2002-04-12 22:59:07 +00:00
Aaron Bannert
bf8bb929ac Fix a typo and a build error detected by the lovely HPUX11 ANSI C compiler. 2002-04-12 22:25:29 +00:00
Aaron Bannert
22fb507325 Fix a problem where php-generated data was pushed down the entire output
filter chain instead of just down the rest of the chain. This fix will
speed up some unnecessary overhead introduced in the last patch.

Suggested by:  Cliff Woolley <jwoolley@apache.org>
2002-04-11 20:34:56 +00:00
Aaron Bannert
27a5b380cf PHP filters and Apache 2 aren't quite a perfect match yet, so we have
to do some trickery with the server_context to make sure it is always
valid within the current thread.

This patch makes sure the server_context is created in apache's
post_read_request hook phase, and then registeres a cleanup that
will NULL out the server context when the request goes out of scope.
Then, inside the output filters, if the server_context is null we
throw an error. Finally, instead of saving the output filter in
the server_context, now we store the entire request_rec pointer
in there.

POST bodies appear to be working now, although they are very inefficient.
The input filter is still just realloc()ing for whatever data comes
down the input pipe, and then sending this to PHP. This means that
we are doing some really nasty memory management on big POST bodies.
For now this it allows for unlimited input bodies, which means that
a big POST could potentially DoS a box by making it run out of memory.
We might want to put a limit on here just in case, at least until
we figure out how to consume input data more efficiently into php.
2002-04-11 19:25:08 +00:00
Aaron Bannert
caf7f3d4d2 Don't depend on the context provided by the filter (f->ctx) anymore. In
Apache 2 the input and output filter contexts are kept unique. We now
only depend on SG(server_context) for each request, and assume that
the same thread will process the entire request. At some point it
would be wise to separate the input and output contexts.
2002-04-11 17:34:17 +00:00
Aaron Bannert
d19fd6a0ad Return the number of bytes consumed, not the number of bytes left.
Suggested by:  Brian Havard <brianh@kheldar.apana.org.au>
2002-04-11 17:22:11 +00:00
Sebastian Bergmann
bdc27671cd Update README. 2002-04-11 17:01:51 +00:00
Sebastian Bergmann
38af983c80 Patch by Aaron Bannert <aaron@clove.org> and Cliff Woolley <jwoolley@virginia.edu>. 2002-04-11 06:01:54 +00:00
foobar
6958369295 Let people know this stuff is experimental. 2002-04-10 09:42:38 +00:00
Sterling Hughes
53fdf83f09 eekk -- and ssb thought my macro's were icky :) 2002-04-09 17:25:13 +00:00
Sterling Hughes
11acdb097d fix compile warnings 2002-04-09 06:13:00 +00:00
Sebastian Bergmann
b0eb69ecc0 Fix output directory. 2002-04-06 15:37:08 +00:00
Sebastian Bergmann
db840b8001 Fix Debug_TS build. 2002-04-06 08:57:06 +00:00