STATUS revision 608ca3face177185f0f76f899ed8428e5f05cecf
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean YeAPACHE 2.0 STATUS: -*-text-*-
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean YeLast modified at [$Date: 2002/03/04 14:44:28 $]
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye 2.0.33 : in development
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye 2.0.32 : released Feburary 16, 2002.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye 2.0.31 : rolled Feburary 1, 2002. not released.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye 2.0.30 : tagged January 8, 2002. not rolled.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye 2.0.29 : tagged November 27, 2001. not rolled.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye 2.0.28 : released November 13, 2001
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye 2.0.27 : rolled November 6, 2001
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye 2.0.26 : tagged October 16, 2001. not rolled.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye 2.0.25 : rolled August 29, 2001
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye 2.0.24 : rolled August 18, 2001
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye 2.0.23 : rolled August 9, 2001
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye 2.0.22 : rolled July 29, 2001
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye 2.0.21 : rolled July 20, 2001
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye 2.0.20 : rolled July 8, 2001
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye 2.0.19 : rolled June 27, 2001
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye 2.0.18 : rolled May 18, 2001
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye 2.0.17 : rolled April 17, 2001
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye 2.0.16 : rolled April 4, 2001
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye 2.0.15 : rolled March 21, 2001
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye 2.0.14 : rolled March 7, 2001
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye 2.0a9 : released December 12, 2000
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye 2.0a8 : released November 20, 2000
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye 2.0a7 : released October 8, 2000
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye 2.0a6 : released August 18, 2000
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye 2.0a5 : released August 4, 2000
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye 2.0a4 : released June 7, 2000
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye 2.0a3 : released April 28, 2000
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye 2.0a2 : released March 31, 2000
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye 2.0a1 : released March 10, 2000
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean YePlease consult the following STATUS files for information
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Yeon related projects:
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean YeCURRENT RELEASE NOTES:
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * 33 status: tag on 3/7/2002? Ideally, this release should have all
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye showstoppers removed. Are we dreaming? Can the horde
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye of developers actually focus on removing them in a week?
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye (Note: Win32 had its .32 yanked due to a mmap() bug.)
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean YeFINAL RELEASE SHOWSTOPPERS:
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * If any request gets to the core handler, without a flag that this
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye r->filename was tested by dir/file_walk, we need to 500 at the very
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye end of the ap_process_request_internal() processing. This provides
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye authors of older modules better compatibility, while still improving
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye the security and robustness of 2.0.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Status: still need to decide where this goes, OtherBill comments...
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Message-ID: <065701c14526$495203b0$96c0b0d0@roweclan.net>
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye we need to look at halting this in the 'default handler' case,
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye and that implies pushing the 'handler election' into the request
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye internal processing phase from the run request phase.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Jim asks: would a stopgap be something bogus like adding another
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye flag to request_rec ala eos_sent and before we OK, if not set
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * We need to find out the right place to add the AddOutputFilterByType
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye directive. mod_mime? No. core with fixups? That's what it is now.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye ap_pass_brigade() hook? As a filter that runs at HTTP_HEADER stage?
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye This needs to be resolved before the next public release.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Message-ID: <018701c1b99c$f0f24310$0a01230a@KOJ>
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * API changes planned for 2.0 that should happen before the
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * Free lists for bucket allocation
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * Pool allocator change
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * Graceful restart is broken in the worker MPM.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Message-ID: <m3g03tnrkv.fsf@rdu163-40-092.nc.rr.com>
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Justin asks: "Is this fixed? I can't reproduce now."
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Jeff: Here are some problems that remain now that the segfault
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye 1) accept mutex gets cleaned up before children are gone; with
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye some types (sysvsem), this causes immediate failure in a
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye thread blocked in the accept mutex; that thread will exit(),
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye presumably taking with it any active connections owned by other
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye threads in the process
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye 2) even without accept mutex (1 listening socket), I see some
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye evidence of a child exiting prematurely (dropped connections)
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * Address popular PRs
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * Win32 doesn't install as service correctly [9863, 9914, 9961]
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * Don't be stupid and cd to a blank directory when doing installs
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean YeCURRENT VOTES:
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * Should we always build binaries statically unless otherwise
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Message-ID: <20020129210006.B23512@Lithium.MeepZor.Com>
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye -1: Justin, Ian
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * If the parent process dies, should the remaining child processes
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye "gracefully" self-terminate. Or maybe we should make it a runtime
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye option, or have a concept of 2 parent processes (one being a
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye "hot spare").
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye See: Message-ID: <3C58232C.FE91F19F@Golux.Com>
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Self-destruct: Ken, Martin
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Not self-destruct: BrianP, Ian, Cliff, BillS
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Make it runtime configurable: Aaron, Jim, Justin
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Have 2 parents: +1: Jim
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye +0: Martin (while standing by, could it do
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye something useful?)
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * Make the worker MPM the default MPM for threaded Unix boxes.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye +1: Justin, Ian, Cliff
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye -0: Aaron (premature decision, needs more discussion), Lars
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean YeRELEASE NON-SHOWSTOPPERS BUT WOULD BE REAL NICE TO WRAP THESE UP:
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * Convert all instances of the old apr_lock_t type to the new
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Status: Aaron has converted all but the perchild MPM to
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye use the new lock API. Since perchild has been put
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye on the back burner, this is no longer a showstopper.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Aaron will patch perchild as soon as it becomes
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye buildable again.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * With AP_MODE_EXHAUSTIVE in the core, it is finally clear to me
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye how the Perchild MPM should be re-written. It hasn't worked
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye correctly since filters were added because it wasn't possible to
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye get the content that had already been written and the socket at
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye the same time. This mode lets us do that, so the MPM can be
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * htpasswd blindly processes the file you give it, and does no
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye sanity checking before totally corrupting whatever file it was
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye you thought you had. It should check the input file and bail
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye if it finds non-comment lines that do not contain exactly 1
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye ':' character.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Message-ID: <20020217150457.A31632@clove.org>
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * Can a static httpd be built reliably?
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Message-ID: <20020207142751.T31582@clove.org>
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * [Ken] Test suite failures:
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye o worker is also failing some of the 'cgi' subtests
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye (see <URL:http://Source-Zone.Org/Apache/regression/>):
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Justin says: "Worker should be fine and passes httpd-test here.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye If you can provide evidence that it can be reproduced
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye outside of httpd-test, then it's a showstopper. I
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye think it's a perl or a httpd-test problem."
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Not a showstopper: Justin
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * Usage of APR_BRIGADE_NORMALIZE in core_input_filter should be
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye removed if possible.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Message-ID: <Pine.LNX.4.33.0201202232430.318-100000@deepthought.cs.virginia.edu>
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * There is a bug in how we sort some hooks, at least the pre-config
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye hook. The first time we call the hooks, they are in the correct
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye order, but the second time, we don't sort them correctly. Currently,
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye the modules/http/config.m4 file has been renamed to
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye modules/http/config2.m4 to work around this problem, it should moved
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye back when this is fixed. rbb
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Justin says: "Is this really a showstopper? This has been here
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye forever. What's wrong? Does this have to do with
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye autoconf or m4?"
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Not a showstopper: Justin, BrianP, trawick, gregames
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * The Add...Filter and Set...Filter directives do not allow the
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye administrator to order filters, beyond the order of filename (mime)
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye extensions. It isn't clear if Set...Filter(s) should be inserted
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye before or after the Add...Filter(s) which are ordered by sequence of
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye filename extensions. At minimum, some sort of +-[0-10] syntax seems
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye like the quickest fix for a 2.0 gold release.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Justin says: "Could we delay this for a point release or 2.1?"
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Not a showstopper: justin, wrowe, trawick, stoddard, Jim, Ian, Aaron,
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * Get perchild to work on platforms other than Linux. This
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye will require a portable mechanism to pass data and file/socket
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye descriptors between vhost child groups. An API was proposed
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Message-ID: <20020111115006.K1529@clove.org>
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * Recent changes to ap_rgetline may have broken EBCDIC boxes.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Message-ID: <20020122072605.GF28051@ebuilt.com>
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Justin says: "I don't have an EBCDIC box to test on. A potential
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye solution is to split out ap_rgetline into two
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye functions as described in this message."
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye gregames says: I see the breakage now, and volunteer to fix it
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye when things calm down a little. It looks OK when
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye there are complete lines and no mime continuations.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * CGI single-byte reads
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye BrianP suggests that this is caused by the ap_scan_script_header_err()
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye routine, which will do single-byte reads until it finds the end
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye of the header, at which point it constructs a pipe-bucket (buffered)
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye to read from.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Proposed solution in:
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Message-ID: <3C36ADAF.60601@cnet.com>
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * Try to get libtool inter-library dependency code working on AIX.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Message-ID: <cm3n10lx555.fsf@rdu163-40-092.nc.rr.com>
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Justin says: If we get it working on AIX, we can enable this
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye on all platforms and clean up our build system
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Jeff says: I thought I tested a patch for you sometime in
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye January that you were going to commit within a few
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * Handling of %2f in URIs. Currently both 1.3 and 2.0
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye completely disallow %2f in the request URI path (see
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye ap_unescape_url() in util.c). It's permitted and passed
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye through in the query string, however. Roy says the
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye original reason for disallowing it, from five years ago,
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye was to protect CGI scripts that applied PATH_INFO to
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye a filesystem location and which might be tricked by
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye ..%2f..%2f(...). We *should* allow path-info of the
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye form 'http://foo.com/index.cgi/path/to/path%2finfo'.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Since we've revamped a lot of our processing of path
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye segments, it would be nice to allow this, or at least
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye allow it conditionally with a directive.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * FreeBSD, threads, and worker MPM. All seems to work fine
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye if you only have one worker process with many threads. Add
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye a second worker process and the accept lock seems to be
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye lost. This might be an APR issue with how it deals with
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye the child_init hook (i.e. the fcntl lock needs to be resynced).
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye More examination and analysis is required.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Status: This has also been reported on Cygwin.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Message-ID: <3C2CC514.8EF3BED1@wapme-systems.de> (cygnus)
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Justin says: So, FreeBSD-CURRENT and Cywin have the same
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye problem. Yum. If another platform has this
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye with worker, this becomes a showstopper.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Aaron says: I spent some time disecting this and have come to
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye the conclusion that it is not a problem in the worker MPM
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye (or at least, it is not isolated to a problem in worker).
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye I'll list some of the problems I'm seeing in case someone
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye else wants to pick up where I've left off:
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye - Delivery of just about any signal to one of the child
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye processes will send it into an infinite loop as well.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye - Even though the parent is spinning out of control,
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye at first the child or children will appear to work
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye properly. At times it is possible to get it into a state,
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye however, where a request will hang until another concurrent
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye request "kicks" the first, at which point the second will
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye hang. My theory is that this has to do with the
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye pthread_cond_*() implementation in FreeBSD, but it's still
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye possible that it is in APR.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Justin adds: Oh, FreeBSD threads are implemented entirely with
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye select()/poll()/longjmp(). Welcome to the nightmare.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye So, that means a ktrace output also has the thread
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye scheduling internals in it (since it is all the same to
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye the kernel). Which makes it hard to distinguish between
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye our select() calls and their select() calls.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye *bangs head on wall repeatedly* But, some of the libc_r
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye files have a DBG_MSG #define. This is moderately helpful
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye when used with -DNO_DETACH. The kernel scheduler isn't
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye waking up the threads on a select(). Yum. And, I bet
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye those decrementing select calls have to do with the
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye scheduler. Time to brush up on our OS fundamentals.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * There is increasing demand from module writers for an API
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye that will allow them to control the server � la apachectl.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Reasons include sole-function servers that need to die if
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye an external dependency (e.g., a database) fails, et cetera.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Perhaps something in the (ever more abused) scoreboard?
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye rbb: I don't believe the scoreboard is the correct mechanism
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye for this. We already have a pipe that goes between parent
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye and child for graceful shutdown events, along with an API that
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye can be used to send a message down that pipe. In threaded MPMs,
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye it is easy enough to make that one pipe be used for graceful
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye and graceless events, and it is also easy to open that pipe
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye to both parent and child for writing. Then we just need to
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye figure out how to do graceless on non-threaded MPMs.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * Allow the DocumentRoot directive within <Location > scopes? This
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye allows the beloved (crusty) Alias /foo/ /somepath/foo/ followed
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye by a <Directory /somepath/foo> to become simply
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye <Location /foo/> DocumentRoot /somefile/foo (IMHO a bit more legible
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye and in-your-face.) DocumentRoot unset would be accepted [and would
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye not permit content to be served, only virtual resources such as
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye server-info or server-status.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye This proposed change would _not_ depricate Alias.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * Win32: Rotatelogs sometimes is not terminated when Apache
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye goes down hard. FirstBill was looking at possibly tracking the
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye child's-child processes in the parent process.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye OtherBill asks, wasn't this fixed?
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye stoddard: Not fixed. Shared scoreboard might offer a good
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye way for the parent to keep track of 'other child' processes
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye and whack them if the child goes down.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * Win32: Add a simple hold console open patch (wait for close or
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye the ESC key, with a nice message) if the server died a bad
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye death (non-zero exit code) in console mode.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Resolution: bring forward same ugly hacks from 1.3.13-.20
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * Port of mod_ssl to Apache 2.0:
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye The current porting state is summarized in modules/ssl/README. The
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye remaining work includes:
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye (1) stablizing/optimizing the SSL filter logic
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye (2) Enabling the various SSL caching mechanisms (shmcb, shmht)
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye (3) Enabling SSL extentions
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye (4) Trying to seperate the https filter logic from mod_ssl -
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye This is to facilitate other modules that wish to use the https
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye filter or the mod_ssl logic or both as required.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Justin: mod_ssl filter logic is redone, so that should be fine.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Madhu has submitted a patch for SSL caching - however, I
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye am -0 on that patch as I *think* we could implement the
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye shared memory another way that is much cleaner (i.e.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye treat shmem directly as a dbm via APR routines). Justin
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye also thinks that the https filter logic may be sufficiently
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye decoupled now, but isn't really sure.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * Performance & Debug: Eliminate most (and perhaps all) of the
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye malloc/free calls in the bucket brigade code. Need some
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye light weight memory management functions that allow freeing
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye memory (putting it back into a memory pool) when it is no
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye longer needed. Enabling simple debugging features like guard
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye bands, double free detection, etc. would be cool but certainly
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye not a hard requirement.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Status: Cliff started to implement this using SMS as has
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye been discussed at length for months, but since
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye SMS is not being used anywhere else in the server,
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye several people expressed the opinion that we should
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye get rid of it entirely, meaning that the buckets
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye need their own memory management (free list) functions.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Cliff will implement that this weekend so we at least
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye have something to look at/compare with.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * Eliminate unnecessary creation of pipes in mod_cgid
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Status: Ken asks, didn't Jeff fix this when he fixed the queue
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * Combine log_child and piped_log_spawn. Clean up http_log.c.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Common logging API.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * Document mod_file_cache.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * Platforms that do not support fork (primarily Win32 and AS/400)
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Architect start-up code that avoids initializing all the modules
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye in the parent process on platforms that do not support fork.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * Win32: Migrate the MPM over to use APR thread/process calls. This
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye would eliminate some code in the Win32 branch that essentially
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye duplicates what is in APR.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * There are still a number of places in the code where we are
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye losing error status (i.e. throwing away the error returned by a
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye system call and replacing it with a generic error code)
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * Mass vhosting version of suEXEC.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * All DBMs suffer from confusion in support/dbmmanage (perl script) since
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye the dbmmanage employs the first-matched dbm format. This is not
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye necessarily the library that Apache was built with. Aught to
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye rewrite dbmmanage upon installation to bin/ with the proper library
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye for predictable mod_auth_dbm administration.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Questions; htdbm exists, time to kill dbmmanage, or does it remain
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye useful as a perl dbm management example? If we keep it,
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye do we address the issue above?
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * Integrate mod_dav.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Some additional items remaining:
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye - case_preserved_filename stuff
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye (use the new canonical name stuff?)
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye - find a new home for ap_text(_header)
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye - is it possible to remove the DAV: namespace stuff from util_xml?
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * ap_core_translate() and its use by mod_mmap_static and mod_file_cache
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye are a bit wonky. The function should probably be exposed as a utility
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye function (such as ap_translate_url2fs() or ap_validate_fs_url() or
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye something). Another approach would be a new hook phase after
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye "translate" which would allow the module to munge what the
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye translation has decided to do.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Status: Greg +1 (volunteers), Ryan +1
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * Explore use of a post-config hook for the code in http_main.c which
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye calls ap_fixup_virutal_hosts(), ap_fini_vhost_config(), and
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye ap_sort_hooks() [to reduce the logic in main()]
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * read the config tree just once, and process N times (as necessary)
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * (possibly) use UUIDs in mod_unique_id and/or mod_usertrack
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * (possibly) port the bug fix for PR 6942 (segv when LoadModule is put
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye into a VirtualHost container) to 2.0.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * shift stuff to mod_core.h
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * callers of ap_run_create_request() should check the return value
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye for failure (Doug volunteers)
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * Win32: Get Apache working on Windows 95/98. The following work
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye (at least) needs to be done:
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye - Document warning that OSR2 is required (for Crypt functions, in
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye rand.c, at least.) This could be resolved with an SSL library, or
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye randomization in APR itself.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye - Bring the Win9xConHook.dll from 1.3 into 2.0 (no sense till it
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye actually works) and add in a splash of Win9x service code.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * In order to use a DSO version of mod_ssl we have to link with
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye -lssl and -lcrypto. A workaround is in place right now where the
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye entire EXTRA_LIBS macro is being appended to the objects list, but
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye this is a hack. We should either revamp the APACHE_CHECK_SSL_TOOLKIT
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye autoconf function or come up with some other autoconf checks to
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye search for libssl and libcrypto and properly add them to mod_ssl's
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * Fix the worker MPM to use POD to kill child processes instead
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye of ap_os_killpg, regardless of how they should die. (Ryan Bloom)
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean YePRs that have been suspended forever waiting for someone to
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Yeput them into 'the next release':
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * PR#76: general
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye missing call to "setlocale();"
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * PR#78: mod_include
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Additional status for XBitHack directive
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * PR#362: mod_proxy
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Mod_proxy doesn't allow change of error pages
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * PR#370: mod_env
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Modified PATH environemnt variable is not passed, instead
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye system's is used
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * PR#440: mod_proxy
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Proxy doesn't deliver documents if not connected
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * PR#534: mod_proxy
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye proxy converts ~name to %7Ename when name starts with a dot (.)
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * PR#537: mod_access
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye mod_access syntax allows hosts that should be restricted
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * PR#557: mod_auth-any
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye ~UserHome directories are not honored in absolute pathname
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye requests (.htaccess)
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * PR#612: mod_proxy
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Proxy FTP Authentication Fails
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * PR#623: mod_include
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye A smarter "Last Modified" value for SSI documents (see PR number 600)
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * PR#628: config
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Request of "Options SymLinksIfGroupMatch"
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * PR#700: mod_proxy
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Proxy doesn't do links right for OpenVMS files through ftp:
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * PR#759: mod_imap
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye imap should read <MAP><AREA>*</MAP> too!
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * PR#793: general
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye RLimitCPU and RLimitMEM don't apply to all children like they should
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * PR#921: suexec
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Uses cwd before filling it in, doesn't use syslog
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * PR#922: config
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye it is useful to allow specifiction that root-owned symlinks
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye should always be followed
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * PR#980: mod_proxy
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Controlling Access to Remote Proxies would be nice...
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * PR#994: mod_proxy
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Adding authentication "on the fly" through the proxy module
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * PR#1004: apache-api
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye request_config field in request_rec is moderately bogus
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * PR#1028: other
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye DoS attacks involving memory consumption
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * PR#1050: mod_log-any
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Logging of virtual server to error_log as well
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * PR#1085: mod_proxy
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye ProxyRemote make a dead cycle.
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye * PR#1117: mod_auth-any
3f1e69bef33050bee99ea1e9992af13fc467281fCheng Sean Ye Using NIS passwd.byname dbm files with AuthDBMUserFile
setlogin() is not called, causing problems with e.g. identd
add allow,deny/deny,allow warning to mod_access
A small addition to rotatelogs.c to improve program functionality.
Feedback/Comment on APACI
Inclusion of RPM spec file in CVS/distributions
* orig_ct in the byterange/multipart handling may not be