André Malo [Mon, 7 Jul 2003 00:34:10 +0000 (00:34 +0000)]
well, (kinda) backport LimitInternalRecursion. This prevents
the server from crashing if someone configure an infinite loop of
internal redirects and subrequests. Default value is 20/20
(subsequent redirects/nested subrequests), 0 means unlimited.
The patch works fine on my box, but is required to be tested
extensively before the next release.
PR: 19753 (and probably more)
Obtained from: 2.0 patch
Reviewed by: original 2.0 port by Justin and BrianP (?)
*) Removed BIND_NOSTART from HP/UX shl_load() logic for loadable
Apache modules, so that statics are initialized when the module
is loaded (especially critical for c++ modules on HPUX.)
[William Rowe, Noah Arliss <narliss@netegrity.com>]
Fix PR 12706 for users on very fast boxen who seem to have issues with
the 'post-build' logic to create uri_delims.h and test_char.h. Also
note the .pdb debug symbol patch from last week.
Jeff Trawick [Fri, 23 May 2003 15:25:24 +0000 (15:25 +0000)]
Make sure the accept mutex is released before calling child exit
hooks and cleanups. Otherwise, modules can segfault in such code
and, with pthread mutexes, leave the server deadlocked. Even if
the module doesn't segfault, if it performs extensive processing
it can temporarily prevent the server from accepting new
connections.
Incorporate similar .pdb mechanics as Apache 2.0. Because the list is
much smaller, use explicit copy commands instead of the complex makefile
install logic we have in Apache 2.0. The generated .PDBs can be used from
VisualStudio or WinDbg, or virtually any Win32 debugging tool to either
debug the released binaries, or to perform post-mortem on crashdump files.
This has worked -very- well for Win32 Apache 2.0, moving the compilation
of buildmark.obj into a pre-link step ensures we always recompile with
the correct datestamp every rebuild, even when building in the IDE without
using makefile.win.
André Malo [Mon, 19 May 2003 02:35:31 +0000 (02:35 +0000)]
Fix mod_rewrite's abs_URI handling.
- uris were partially not correctly escaped (in particular:
ldap, news, mailto)
- not all uri schemes contain an authority component (//)
- add nntp:// scheme
- don't add a query string (and drop r->args) if it's not
http or mailto scheme
- be more efficient (think so)
For whatever reason, the install-tpf.html in *apache-1.3* (not in the
httpd-docs-1.3 repository) was in the attic. However, it was not marked
as deleted, this was noted in an rsync. Surprized it never bit anyone
else. In any case, mark this file as removed by re-removing it, one
cannot simply 'move' files into Attic/
Joshua Slive [Tue, 13 May 2003 15:19:30 +0000 (15:19 +0000)]
Reverse an earlier decision not to log EACESS results. This will
result in multiple log-entries when requesting a directory where
the DirectoryIndex contains multiple entires, but it should eliminate
a major source of confusion which results from a 403 error with
no error log entry.
The fix was pointed out by Jeff; I just adjusted the error message.
Jeff Trawick [Mon, 5 May 2003 11:45:49 +0000 (11:45 +0000)]
AIX: Change the default accept mechanism from pthread back to
fcntl. Idle child cleanup doesn't work when the child selected
for termination by the parent is waiting on a pthread mutex, and
because the AIX kernel's notion of hot process is apparently the
same as Apache's, it is common for the Apache parent to continually
select a child for termination that the kernel will leave waiting
on the mutex for extended periods of time. There are other
concerns with pthread mutexes as well, such as the ability to
deadlock the server if a child process segfaults while holding the
mutex.
Martin Kraemer [Fri, 25 Apr 2003 13:50:23 +0000 (13:50 +0000)]
A newly created child now has a start_time of 0, to prevent
mod_status from displaying a bogus value for the "time to
process most recent request" column for freshly-started children
in a previously-used scoreboard slot.
Mark J. Cox [Thu, 3 Apr 2003 13:09:37 +0000 (13:09 +0000)]
Improve consistancy of security messages in changelog
Update CVE candidate names where promoted from CAN to CVE
PR:
Obtained from:
Submitted by:
Reviewed by:
André Malo [Wed, 2 Apr 2003 21:11:41 +0000 (21:11 +0000)]
Fix the misunderstandings between local URL paths and local
system paths. Note that mod_rewrite handles _both_.
Fixed also some comments to make the explanations more clear.
André Malo [Wed, 2 Apr 2003 21:01:09 +0000 (21:01 +0000)]
Fix the prefix_stat function. (which does a stat call on the first
path segment). This function was still tailored for unix systems only.
It should work on other systems as well now.
Note that in 2.x we use the apr_filepath_root function that doesn't
exist in 1.3. It is replaced by some simple workaround stuff which
behaves similar.
Obtained from: 2.1 (r1.143)
Reviewed by: Brad Nicholes, Will Rowe
André Malo [Wed, 2 Apr 2003 20:54:09 +0000 (20:54 +0000)]
Remove _unused_ code. The condition is never true, because
"A local rewrite in per-directory context" was caught much earlier.
I'd guess this piece of code was c&p accidentally...
Obtained from: 2.1 (r1.142)
Reviewed by: Brad Nicholes, Will Rowe
André Malo [Wed, 19 Mar 2003 14:01:08 +0000 (14:01 +0000)]
backport from 2.x:
Prevent endless loops of internal redirects in mod_rewrite by
aborting after exceeding a limit of internal redirects. The
limit defaults to 10 and can be changed using the RewriteOptions
directive with the new MaxRedirects=n argument.
(The latter required some restructuring of the RewriteOptions
evaluation code).
Martin Kraemer [Wed, 5 Mar 2003 15:50:29 +0000 (15:50 +0000)]
Fix longstanding BS2000(OSD_POSIX) related bug -- a CGI program which
terminated quickly would cause an endless loop.
;-) Nobody noticed by now, so I guess few people used it.
Bradley Nicholes [Wed, 26 Feb 2003 17:38:44 +0000 (17:38 +0000)]
Since NetWare doesn't have processes the entire issue of having to close
handles before an exec is non-existent. So just bypass the close on exec
magic cleanup stuff.
Jeff Trawick [Wed, 19 Feb 2003 20:39:59 +0000 (20:39 +0000)]
Unix build: Add support for environment variable
EXTRA_LDFLAGS_SHLIB, which allows the user to add to the hard-coded
ld flags specified for DSOs. Compare with the existing LDFLAGS_SHLIB
environment variable, which allows the user to completely replace the
hard-coded ld flags specified for DSOs.
André Malo [Sun, 16 Feb 2003 04:42:24 +0000 (04:42 +0000)]
drop the guess_domain function.
Our docs say about AuthDigestDomain:
This directive should always be specified and contain at least the (set of)
root URI(s) for this space. Omitting to do so will cause the client to send
the Authorization header for every request sent to this server.
guessing the parameter is somewhat bogus. guess_domain() also resulted sometimes
in relative URIs, non-URI strings or empty strings, which caused a lot of
problems.
According to the docs, the domain parameter will be omitted now,
if not specified. This is exactly, what one would expect.
Teach win32 about magic_cleanups to toggle the INHERIT bit for handles.
This can only work for NT, since 9x cannot alter a handle's INHERIT bit
without a DuplicateHandle() and then closing the original handle, which
just doesn't fit into this API very well.
{Actually, could be done with a bit of convolutions for FILE and fd,
but this won't work by the ap_note_cleanups_for_h() api.)
Jim Jagielski [Wed, 18 Dec 2002 22:43:08 +0000 (22:43 +0000)]
Clean up the compact/verbose cookie code to prevent a lot of
overlap. Who knows, maybe a 3rd format might pop up one
day. If this breaks, note who to blame for it
PR:
Obtained from:
Submitted by:
Reviewed by:
Apply a fix already in Apache 2.0 (and add quality weights) to prevent
us from attempting to serve .gif or any other flavor of negotiated
resources.
This patch fixes the Accept: header of the autoindex request so that
it's subrequests look for text/ only, with weights applied to recover
.html, then .txt, then some other flavor of text/* in their absense.
Resolve 80% of the HEADER/README not recognized on Win32 bug (PR 7300).
Since Apache 1.3 does NOT normalize file names, but instead always lower
cases the names, HEADER.txt or HEADER.html would never, ever be noted
as a match. This is different from Apache 2.0, which can deal with the
proper names and match only HEADER.* or header.* files based on the
origin filename.
Jim Jagielski [Mon, 16 Dec 2002 13:17:38 +0000 (13:17 +0000)]
Add in the new CookiePrefix and CookieFormat directives that allow
for the tracking cookie to have an arbitrary string prepended or
to choose a normal or "compact" cookie.
PR:
Obtained from: Fast Search & Transfer (contributed)
Submitted by: P�l L�berg <pallo@initio.no>
Reviewed by: Rasmus, Jim
Jim Jagielski [Thu, 12 Dec 2002 16:09:52 +0000 (16:09 +0000)]
Where the OS allows, we now proactively use the various _ex and
magic cleanups to close fds (lock files, log files and sockets)
to prevent them from leaking into 3rd party modules that don't
call ap_cleanup_for_exec() before forking off subprocesses.
Bill Stoddard [Tue, 10 Dec 2002 20:18:25 +0000 (20:18 +0000)]
Fix condition that arises when the server is configured as a proxy gateway (aka reverse
proxy) and multiple processes are serving the same backend file at the same time.
Only one of the processes will actually succeed at saving the file, the others will
fail with the message "[error] (17)File exists: proxy: error linking cache file ...".
Changing the message level from APLOG_ERR to APLOG_INFO will keep this message
out of the logs w/o major overhaul of this code.
Jim Jagielski [Mon, 9 Dec 2002 20:21:00 +0000 (20:21 +0000)]
Get rid of somewhat long-standing issue regarding large values
of precision causing a buffer to be clobbered in the vformatter
function (eg: ap_snprintf)
PR:
Obtained from:
Submitted by:
Reviewed by:
Implemented ap_os_default_port() to allow NetWare to resolve the correct
default port based on the request method. This fixes a problem with URL
reconstruction on a redirect.
projects / thirdparty / apache / httpd.git / log
