Create the abs.dsp win32 project to support building ab with /USE_SSL,
and modify Makefile.win to install the ssl-enabled abs.exe binary
as 'bin/ab.exe' if srclib/openssl exists.
We might want to disable this, by default, when we get ready for GA.
ITMT, this provides a fantastic stress-test for interoperability of
the ap_die handling, the includes filter, the negotiated typemap
behavior and the new typemap-body functions.
Keep adding to the list of errors as they are converted from the
httpd-2.0/docs/docroot/error individual files into docs/error.
Paul J. Reder [Mon, 20 Aug 2001 19:47:52 +0000 (19:47 +0000)]
The prefork and OS/2 MPMs are overwriting the pid file when a second copy
of httpd is started and shuts down due to socket conflict. Moving the
call to ap_log_pid solves the problem.
*) Changed the late-1.3 log_config substitution %c to %X (logs the
status of the closed connection) as it conflicts with the far more
common, historical ssl logging directive %...{var}c.
My bad. We need to use the constants defined in httpd.h. The essential
reason is that these errors will grow to include 'other' things, like
custom error responses, DAV error responses, and so on. They may also
start including 'sub errors' - local langauge explanations of specific
failures. Let's use the httpd.h constants rather than numbers, they
are easier to read.
This is the Body:tag\n<Content>*tag\n patch for type map files. If the
content can be served, it sends a file bucket of the type map file
itself, forward spaced to the <Content>, set to the length of <Content>.
<Content> may contain any binary data. The end tag must exist, anything
between the tag and \n is ignored. This version (v.s. the patch sent to
the list) includes error reporting if the end tag cannot be located.
We need to change the etag code to accept an additional argument, so we
can set the cache controls properly.
Brian Havard [Mon, 20 Aug 2001 16:10:57 +0000 (16:10 +0000)]
Catch exceptions in worker threads so a segfault doesn't kill all active
connections in the process. When this occurs, a graceful shutdown of the
child is initiated so any leaked resources & hosed memory is disposed of ASAP.
- Switch to AC_SEARCH_LIBS to check multiple libraries for mod_auth_db
(the AC_* way of doing what Martin committed)
- Fix the configuration check for mod_auth_digest so that:
- Everything is on separate lines so that the preprocessor doesn't scream
- It adds the path to APR. (APR_SOURCE_DIR looks right, but I'm not sure.)
By the time the modules are configured, the CPPFLAGS and such are not
setup to point at APR. This should probably be rectified and then this
can be taken out - so you could assume that apr.h is in your CPPFLAGS
somewhere when configuring the modules.
Martin Kraemer [Mon, 20 Aug 2001 07:00:23 +0000 (07:00 +0000)]
On modern systems (e.g., FreeBSD), the db* functions reside in libc
rather than in libdb. Also, be more precise in telling what is actually
checked for (we are NOT testing for main(), but we are testing for dbopen(),
right?).
Without the first fix, configuration would fail completely for FreeBSD
Doug MacEachern [Mon, 20 Aug 2001 02:43:02 +0000 (02:43 +0000)]
change filter name to CaseFilterIn (mod_case_filter.c already uses CaseFilter)
allow filter to be configured without 'CaseFilterIn On'
PR:
Obtained from:
Submitted by:
Reviewed by:
This Variant type-map file is the authoritative reference to our
"It Worked" page variants. Of course it isn't yet (although it is
complete), please review httpd-std.conf as well as this list
for inconsistencies (I documented those I discovered).
This Variant type-map file is the authoritative reference to our
"It Worked" page variants. Of course it isn't yet (although it is
complete), please review httpd-std.conf as well as this list
for inconsistencies (I documented those I discovered).
Per discussion on new-httpd, change the logic to do the following:
--enable-modules=most will disable a module if it can not be built
--enable-modules=all will error if a module can not be built
--enable-modules=foo will error if the specified module can not be built
(even if --enable-modules=most is set)
This is Apache 2.0, which has been essentially rewritten.
This citation was inconsistent with the rest of our credits.
We still thank Ambarish Malpani of ValiCert, Inc. for his
contribution, but we have many, many others to thank, and this
one accolade was out of place.
Graham Leggett [Sun, 19 Aug 2001 01:44:43 +0000 (01:44 +0000)]
Added a default commented-out mod_ldap and mod_auth_ldap
configuration to httpd-std.conf and httpd-win.conf
PR:
Obtained from:
Submitted by:
Reviewed by:
Graham Leggett [Sat, 18 Aug 2001 21:10:03 +0000 (21:10 +0000)]
Enabled negative caching on attribute comparisons in the LDAP cache.
Fixed a problem where the default cache TTL was set in milliseconds
not microseconds causing the cache to time out almost immediately.
PR:
Obtained from:
Submitted by:
Reviewed by:
Graham Leggett [Sat, 18 Aug 2001 19:18:12 +0000 (19:18 +0000)]
Fixed LDAP cleanup on graceful restarts. LDAP connections are now
cleaned up when the connection pool pool is cleaned up.
PR:
Obtained from:
Submitted by:
Reviewed by:
Paul J. Reder [Sat, 18 Aug 2001 17:36:26 +0000 (17:36 +0000)]
Fix a minor issue with Jeff Trawick's mod_include
patch. Without this patch, the code will just allocate
more bytes in get_combined_directive than are needed.
Jeff Trawick [Fri, 17 Aug 2001 17:24:46 +0000 (17:24 +0000)]
Fix a problem parsing tags.
We're in find_end_sequence() walking over the directive and the rest
of the tag but we stop parsing so we don't walk over too much storage
(e.g., mmap) at once). When we resume later, ctx->directive_length
has the value it had when we bailed out, so directive_length is wrong
and we hit errors like:
[Fri Aug 17 11:12:06 2001] [error] [client 127.0.0.1] unknown directiv
e "include virt" in parsed doc /home/trawick/apacheinst/htdocs/manual/
mod/mod_include.html
Brian Havard [Fri, 17 Aug 2001 17:07:34 +0000 (17:07 +0000)]
New multi-process multi-threaded MPM for OS/2. Not fully polished but works
better than spmt_os2 already (graceful restarts actually work). A summary
of the process/thread structure is provided in the comments at the start
of mpmt_os2.c
Since we expect many users to start experimenting with the next release,
and the 'Apache' default server name definately conflicts with a stable
1.3.20 install, changed the AP_DEFAULT_SERVICE_NAME to "Apache2" (unless
the user overrides in the compilation.)
Don't retrieve the default service name if the -k run arg is given.
-k run represents running 'in a console', and is the default if no
-k option is given at all.
Perhaps we want to assume -k start as the default (meaning a console
user in the future would have to explicitly ask for -k run if they
want to test as a console) now that services are far more stable.
That can be debated as a seperate patch.
Ok, I didn't like the choices I made an hour or two ago. Here are (IMHO)
the most appropriate log levels (feel free to debate.) OTOH, this patch
fixes the problem with an apache -k start -n foo command simply exiting,
when foo isn't installed.
The purpose of this patch is to toggle the debugging mode (default) to
Program Database (from Program Database for Modify on the fly debugging).
The net effect of this patch is to clean up all of the irrelevant entries
associated with either the debugging or release command line switches, and
generally straighten the projects as they would be exported from VC6/SP5.
The outcome of this patch is that VC5 users -should- be able to load and
build the workspace without any errors (as they used to have no symbols
database at all, the /ZI option doesn't work, they had to use cvtdsp.pl
to toggle these to /Zi.)