STATUS revision baf2534719d8ee30c65e3b092dcc76ce15bbf714
da97433d5cd26e422a370d186f98659383c06721csovantAPACHE 2.0 STATUS: -*-text-*-
da97433d5cd26e422a370d186f98659383c06721csovantLast modified at [$Date: 2002/01/23 18:16:45 $]
da97433d5cd26e422a370d186f98659383c06721csovant 2.0.31 : In development
da97433d5cd26e422a370d186f98659383c06721csovant 2.0.30 : tagged January 8, 2002.
da97433d5cd26e422a370d186f98659383c06721csovant 2.0.29 : tagged November 27, 2001. not rolled.
da97433d5cd26e422a370d186f98659383c06721csovant 2.0.28 : released November 13, 2001
da97433d5cd26e422a370d186f98659383c06721csovant 2.0.27 : rolled November 6, 2001
da97433d5cd26e422a370d186f98659383c06721csovant 2.0.26 : tagged October 16, 2001. not rolled.
da97433d5cd26e422a370d186f98659383c06721csovant 2.0.25 : rolled August 29, 2001
da97433d5cd26e422a370d186f98659383c06721csovant 2.0.24 : rolled August 18, 2001
da97433d5cd26e422a370d186f98659383c06721csovant 2.0.23 : rolled August 9, 2001
da97433d5cd26e422a370d186f98659383c06721csovant 2.0.22 : rolled July 29, 2001
da97433d5cd26e422a370d186f98659383c06721csovant 2.0.21 : rolled July 20, 2001
da97433d5cd26e422a370d186f98659383c06721csovant 2.0.20 : rolled July 8, 2001
da97433d5cd26e422a370d186f98659383c06721csovant 2.0.19 : rolled June 27, 2001
da97433d5cd26e422a370d186f98659383c06721csovant 2.0.18 : rolled May 18, 2001
da97433d5cd26e422a370d186f98659383c06721csovant 2.0.17 : rolled April 17, 2001
da97433d5cd26e422a370d186f98659383c06721csovant 2.0.16 : rolled April 4, 2001
da97433d5cd26e422a370d186f98659383c06721csovant 2.0.15 : rolled March 21, 2001
da97433d5cd26e422a370d186f98659383c06721csovant 2.0.14 : rolled March 7, 2001
da97433d5cd26e422a370d186f98659383c06721csovant 2.0a9 : released December 12, 2000
da97433d5cd26e422a370d186f98659383c06721csovant 2.0a8 : released November 20, 2000
2b68af05b4308b32506f1d68f6357cdc458869d7ludovicp 2.0a7 : released October 8, 2000
b168b33f591684db77dc1078f77947802d2fedf4gary.williams 2.0a6 : released August 18, 2000
da97433d5cd26e422a370d186f98659383c06721csovant 2.0a5 : released August 4, 2000
da97433d5cd26e422a370d186f98659383c06721csovant 2.0a4 : released June 7, 2000
da97433d5cd26e422a370d186f98659383c06721csovant 2.0a3 : released April 28, 2000
da97433d5cd26e422a370d186f98659383c06721csovant 2.0a2 : released March 31, 2000
da97433d5cd26e422a370d186f98659383c06721csovant 2.0a1 : released March 10, 2000
da97433d5cd26e422a370d186f98659383c06721csovantPlease consult the following STATUS files for information
da97433d5cd26e422a370d186f98659383c06721csovanton related projects:
da97433d5cd26e422a370d186f98659383c06721csovantRELEASE SHOWSTOPPERS:
da97433d5cd26e422a370d186f98659383c06721csovant * win32 scoreboard should _not_ use anon with no scoreboard name, it
da97433d5cd26e422a370d186f98659383c06721csovant should revert to the old method if scoreboard name is null (use the
da97433d5cd26e422a370d186f98659383c06721csovant NOT_SHARED bit that we do in -X). At this moment Win32 doesn't
da97433d5cd26e422a370d186f98659383c06721csovant -need- to have a shared score, until we create multiple processes,
da97433d5cd26e422a370d186f98659383c06721csovant or unless a module in the parent wants scoreboard data from the child.
da97433d5cd26e422a370d186f98659383c06721csovant If such a module is installed [for the moment] they aught to configure
da97433d5cd26e422a370d186f98659383c06721csovant the ScoreboardFile directive.
da97433d5cd26e422a370d186f98659383c06721csovant * mod_rewrite's locks (win32) were broken by recent changes to the
da97433d5cd26e422a370d186f98659383c06721csovant apr_lock_t api (which we were deprecating in the first place, so
da97433d5cd26e422a370d186f98659383c06721csovant I'm far from certain why it changed.) It has to do with creating
da97433d5cd26e422a370d186f98659383c06721csovant the rewrite logs mutex. The solution is to find the brokenness in
da97433d5cd26e422a370d186f98659383c06721csovant apr_lock_t, but if it goes away, why waste the time?
da97433d5cd26e422a370d186f98659383c06721csovant * ap_directory_walk skips some per-dir config merge functions
da97433d5cd26e422a370d186f98659383c06721csovant if there is no "<Directory />" block in the configuration
da97433d5cd26e422a370d186f98659383c06721csovant Message-ID: <m3itbdiijq.fsf@rdu163-40-092.nc.rr.com>
da97433d5cd26e422a370d186f98659383c06721csovant * Test suite failures:
da97433d5cd26e422a370d186f98659383c06721csovant o perchild doesn't even build
da97433d5cd26e422a370d186f98659383c06721csovant o worker is also failing some of the 'cgi' subtests
da97433d5cd26e422a370d186f98659383c06721csovant (see <URL:http://Source-Zone.Org/Apache/regression/>):
da97433d5cd26e422a370d186f98659383c06721csovant * If any request gets to the core handler, without a flag that this
da97433d5cd26e422a370d186f98659383c06721csovant r->filename was tested by dir/file_walk, we need to 500 at the very
da97433d5cd26e422a370d186f98659383c06721csovant end of the ap_process_request_internal() processing. This provides
da97433d5cd26e422a370d186f98659383c06721csovant authors of older modules better compatibility, while still improving
da97433d5cd26e422a370d186f98659383c06721csovant the security and robustness of 2.0.
da97433d5cd26e422a370d186f98659383c06721csovant Status: still need to decide where this goes, OtherBill comments...
da97433d5cd26e422a370d186f98659383c06721csovant Message-ID: <065701c14526$495203b0$96c0b0d0@roweclan.net>
da97433d5cd26e422a370d186f98659383c06721csovant we need to look at halting this in the 'default handler' case,
da97433d5cd26e422a370d186f98659383c06721csovant and that implies pushing the 'handler election' into the request
da97433d5cd26e422a370d186f98659383c06721csovant internal processing phase from the run request phase.
da97433d5cd26e422a370d186f98659383c06721csovant * There is a bug in how we sort some hooks, at least the pre-config
da97433d5cd26e422a370d186f98659383c06721csovant hook. The first time we call the hooks, they are in the correct
da97433d5cd26e422a370d186f98659383c06721csovant order, but the second time, we don't sort them correctly. Currently,
da97433d5cd26e422a370d186f98659383c06721csovant the modules/http/config.m4 file has been renamed to
da97433d5cd26e422a370d186f98659383c06721csovant modules/http/config2.m4 to work around this problem, it should moved
da97433d5cd26e422a370d186f98659383c06721csovant back when this is fixed. rbb
da97433d5cd26e422a370d186f98659383c06721csovant * The Add...Filter and Set...Filter directives do not allow the
da97433d5cd26e422a370d186f98659383c06721csovant administrator to order filters, beyond the order of filename (mime)
da97433d5cd26e422a370d186f98659383c06721csovant extensions. It isn't clear if Set...Filter(s) should be inserted
da97433d5cd26e422a370d186f98659383c06721csovant before or after the Add...Filter(s) which are ordered by sequence of
da97433d5cd26e422a370d186f98659383c06721csovant filename extensions. At minimum, some sort of +-[0-10] syntax seems
da97433d5cd26e422a370d186f98659383c06721csovant like the quickest fix for a 2.0 gold release.
da97433d5cd26e422a370d186f98659383c06721csovant * mod_negotiation needs a new option or directive, something like
da97433d5cd26e422a370d186f98659383c06721csovant ForceLanguagePriority, to fall back to the LanguagePriority
da97433d5cd26e422a370d186f98659383c06721csovant directive instead of returning a "no acceptable variant" error.
da97433d5cd26e422a370d186f98659383c06721csovant Status: Bill has some code in his tree that accomplishes
da97433d5cd26e422a370d186f98659383c06721csovant this, and will commit it Friday after it's tested.
da97433d5cd26e422a370d186f98659383c06721csovant * Fold mod_auth_db features back into mod_auth_dbm, and depricate it.
da97433d5cd26e422a370d186f98659383c06721csovant This can't wait until we have a 2.0-gold release, if folks need
da97433d5cd26e422a370d186f98659383c06721csovant to move over to auth_dbm, we can't do that to them after 2.0 gold.
da97433d5cd26e422a370d186f98659383c06721csovant Status: Ian says.. auth_dbm can now handle multiple DBM types,
da97433d5cd26e422a370d186f98659383c06721csovant is this still an issue?
da97433d5cd26e422a370d186f98659383c06721csovant Vote: Remove mod_auth_db
b168b33f591684db77dc1078f77947802d2fedf4gary.williams +1: Justin, Ian, Lars
da97433d5cd26e422a370d186f98659383c06721csovant * Convert all instances of the old apr_lock_t type to the new
da97433d5cd26e422a370d186f98659383c06721csovant types (once they are fully supported in APR).
da97433d5cd26e422a370d186f98659383c06721csovant Status: Aaron is working on converting INTRAPROCESS
da97433d5cd26e422a370d186f98659383c06721csovant to apr_thread_mutex_t types. Full replacements for
da97433d5cd26e422a370d186f98659383c06721csovant LOCKALL and CROSS_PROCESS are not yet complete on all
da97433d5cd26e422a370d186f98659383c06721csovant platforms, and should only be used in MPMs like worker
da97433d5cd26e422a370d186f98659383c06721csovant with limited OS exposure.
da97433d5cd26e422a370d186f98659383c06721csovant * ap_create_scoreboard() can exit the process, leaving stuff like
da97433d5cd26e422a370d186f98659383c06721csovant mod_cgid's daemon process stranded. Either ap_create_scoreboard()
da97433d5cd26e422a370d186f98659383c06721csovant needs to be called at a different time or the pre-mpm hook needs
da97433d5cd26e422a370d186f98659383c06721csovant to be able to return an error code.
da97433d5cd26e422a370d186f98659383c06721csovant * A binbuild installation picks up the right libraries when
da97433d5cd26e422a370d186f98659383c06721csovant running apachectl because we set the appropriate environment
da97433d5cd26e422a370d186f98659383c06721csovant variable, but ab, htpasswd, etc. don't know how to pick up apr,
da97433d5cd26e422a370d186f98659383c06721csovant apr-util, etc.
da97433d5cd26e422a370d186f98659383c06721csovant Message-ID: <20020116000226.GA15991@ebuilt.com>
da97433d5cd26e422a370d186f98659383c06721csovant * mod_cgid leaks pipe descriptors, apparently for every request.
da97433d5cd26e422a370d186f98659383c06721csovant I would guess that this is due to recent apr_file_dup[2]
da97433d5cd26e422a370d186f98659383c06721csovant I was unable to recreate the problem with mod_cgi (yea!).
da97433d5cd26e422a370d186f98659383c06721csovant Message-ID: <20020123070827.GV28051@ebuilt.com> (dev@apr)
da97433d5cd26e422a370d186f98659383c06721csovant * Aaron is working on an update to the scoreboard creation
da97433d5cd26e422a370d186f98659383c06721csovant routines that will be more "correct" in terms of using
da97433d5cd26e422a370d186f98659383c06721csovant anonymous shared memory vs. name-based shared memory on
da97433d5cd26e422a370d186f98659383c06721csovant the platforms that support either.
da97433d5cd26e422a370d186f98659383c06721csovantRELEASE NON-SHOWSTOPPERS BUT WOULD BE REAL NICE TO WRAP THESE UP:
da97433d5cd26e422a370d186f98659383c06721csovant * Try to get libtool inter-library dependency code working on AIX.
da97433d5cd26e422a370d186f98659383c06721csovant Message-ID: <cm3n10lx555.fsf@rdu163-40-092.nc.rr.com>
da97433d5cd26e422a370d186f98659383c06721csovant Justin says: If we get it working on AIX, we can enable this
da97433d5cd26e422a370d186f98659383c06721csovant on all platforms and clean up our build system
da97433d5cd26e422a370d186f98659383c06721csovant * Handling of %2f in URIs. Currently both 1.3 and 2.0
da97433d5cd26e422a370d186f98659383c06721csovant completely disallow %2f in the request URI path (see
da97433d5cd26e422a370d186f98659383c06721csovant ap_unescape_url() in util.c). It's permitted and passed
da97433d5cd26e422a370d186f98659383c06721csovant through in the query string, however. Roy says the
da97433d5cd26e422a370d186f98659383c06721csovant original reason for disallowing it, from five years ago,
da97433d5cd26e422a370d186f98659383c06721csovant was to protect CGI scripts that applied PATH_INFO to
da97433d5cd26e422a370d186f98659383c06721csovant a filesystem location and which might be tricked by
da97433d5cd26e422a370d186f98659383c06721csovant ..%2f..%2f(...). We *should* allow path-info of the
da97433d5cd26e422a370d186f98659383c06721csovant form 'http://foo.com/index.cgi/path/to/path%2finfo'.
da97433d5cd26e422a370d186f98659383c06721csovant Since we've revamped a lot of our processing of path
da97433d5cd26e422a370d186f98659383c06721csovant segments, it would be nice to allow this, or at least
da97433d5cd26e422a370d186f98659383c06721csovant allow it conditionally with a directive.
da97433d5cd26e422a370d186f98659383c06721csovant * FreeBSD, threads, and worker MPM. All seems to work fine
da97433d5cd26e422a370d186f98659383c06721csovant if you only have one worker process with many threads. Add
da97433d5cd26e422a370d186f98659383c06721csovant a second worker process and the accept lock seems to be
da97433d5cd26e422a370d186f98659383c06721csovant lost. This might be an APR issue with how it deals with
da97433d5cd26e422a370d186f98659383c06721csovant the child_init hook (i.e. the fcntl lock needs to be resynced).
da97433d5cd26e422a370d186f98659383c06721csovant More examination and analysis is required.
da97433d5cd26e422a370d186f98659383c06721csovant Status: This has also been reported on Cygwin.
da97433d5cd26e422a370d186f98659383c06721csovant Message-ID: <3C2CC514.8EF3BED1@wapme-systems.de> (cygnus)
da97433d5cd26e422a370d186f98659383c06721csovant Justin says: So, FreeBSD-CURRENT and Cywin have the same
b168b33f591684db77dc1078f77947802d2fedf4gary.williams problem. Yum. If another platform has this
b168b33f591684db77dc1078f77947802d2fedf4gary.williams with worker, this becomes a showstopper.
da97433d5cd26e422a370d186f98659383c06721csovant Aaron says: I spent some time disecting this and have come to
da97433d5cd26e422a370d186f98659383c06721csovant the conclusion that it is not a problem in the worker MPM
da97433d5cd26e422a370d186f98659383c06721csovant (or at least, it is not isolated to a problem in worker).
da97433d5cd26e422a370d186f98659383c06721csovant I'll list some of the problems I'm seeing in case someone
da97433d5cd26e422a370d186f98659383c06721csovant else wants to pick up where I've left off:
da97433d5cd26e422a370d186f98659383c06721csovant - Delivery of just about any signal to one of the child
da97433d5cd26e422a370d186f98659383c06721csovant processes will send it into an infinite loop as well.
da97433d5cd26e422a370d186f98659383c06721csovant - Even though the parent is spinning out of control,
da97433d5cd26e422a370d186f98659383c06721csovant at first the child or children will appear to work
da97433d5cd26e422a370d186f98659383c06721csovant properly. At times it is possible to get it into a state,
da97433d5cd26e422a370d186f98659383c06721csovant however, where a request will hang until another concurrent
da97433d5cd26e422a370d186f98659383c06721csovant request "kicks" the first, at which point the second will
da97433d5cd26e422a370d186f98659383c06721csovant hang. My theory is that this has to do with the
da97433d5cd26e422a370d186f98659383c06721csovant pthread_cond_*() implementation in FreeBSD, but it's still
da97433d5cd26e422a370d186f98659383c06721csovant possible that it is in APR.
da97433d5cd26e422a370d186f98659383c06721csovant Justin adds: Oh, FreeBSD threads are implemented entirely with
da97433d5cd26e422a370d186f98659383c06721csovant select()/poll()/longjmp(). Welcome to the nightmare.
da97433d5cd26e422a370d186f98659383c06721csovant So, that means a ktrace output also has the thread
da97433d5cd26e422a370d186f98659383c06721csovant scheduling internals in it (since it is all the same to
da97433d5cd26e422a370d186f98659383c06721csovant the kernel). Which makes it hard to distinguish between
da97433d5cd26e422a370d186f98659383c06721csovant our select() calls and their select() calls.
b168b33f591684db77dc1078f77947802d2fedf4gary.williams *bangs head on wall repeatedly* But, some of the libc_r
b168b33f591684db77dc1078f77947802d2fedf4gary.williams files have a DBG_MSG #define. This is moderately helpful
da97433d5cd26e422a370d186f98659383c06721csovant when used with -DNO_DETACH. The kernel scheduler isn't
da97433d5cd26e422a370d186f98659383c06721csovant waking up the threads on a select(). Yum. And, I bet
da97433d5cd26e422a370d186f98659383c06721csovant those decrementing select calls have to do with the
da97433d5cd26e422a370d186f98659383c06721csovant scheduler. Time to brush up on our OS fundamentals.
da97433d5cd26e422a370d186f98659383c06721csovant * There is increasing demand from module writers for an API
da97433d5cd26e422a370d186f98659383c06721csovant that will allow them to control the server � la apachectl.
da97433d5cd26e422a370d186f98659383c06721csovant Reasons include sole-function servers that need to die if
da97433d5cd26e422a370d186f98659383c06721csovant an external dependency (e.g., a database) fails, et cetera.
da97433d5cd26e422a370d186f98659383c06721csovant Perhaps something in the (ever more abused) scoreboard?
da97433d5cd26e422a370d186f98659383c06721csovant rbb: I don't believe the scoreboard is the correct mechanism
da97433d5cd26e422a370d186f98659383c06721csovant for this. We already have a pipe that goes between parent
da97433d5cd26e422a370d186f98659383c06721csovant and child for graceful shutdown events, along with an API that
da97433d5cd26e422a370d186f98659383c06721csovant can be used to send a message down that pipe. In threaded MPMs,
da97433d5cd26e422a370d186f98659383c06721csovant it is easy enough to make that one pipe be used for graceful
da97433d5cd26e422a370d186f98659383c06721csovant and graceless events, and it is also easy to open that pipe
da97433d5cd26e422a370d186f98659383c06721csovant to both parent and child for writing. Then we just need to figure
945930ece8ea73e5184836bf0bf3efad2e3e6c36gary.williams out how to do graceless on non-threaded MPMs.
945930ece8ea73e5184836bf0bf3efad2e3e6c36gary.williams * revamp the input filter behavior, per discussions since
945930ece8ea73e5184836bf0bf3efad2e3e6c36gary.williams February (and especially at the hackathon last
da97433d5cd26e422a370d186f98659383c06721csovant April). Specifically, ap_get_brigade will return a brigade with
da97433d5cd26e422a370d186f98659383c06721csovant *up to* a specific number of bytes, or a "line" of data. The
da97433d5cd26e422a370d186f98659383c06721csovant read may be blocking or nonblocking. ap_getline() will be
da97433d5cd26e422a370d186f98659383c06721csovant refactored into apr_brigade_getline(), and then DECHUNK can use
da97433d5cd26e422a370d186f98659383c06721csovant f->next (ap_getline will always read "top of input stack"). Also
da97433d5cd26e422a370d186f98659383c06721csovant fix the bug where request body content will end up closing the
da97433d5cd26e422a370d186f98659383c06721csovant connection (buggering up persistent conns).
da97433d5cd26e422a370d186f98659383c06721csovant Status: Justin is working on this as fast as he can.
da97433d5cd26e422a370d186f98659383c06721csovant The core input filters, HTTP-related filters, mod_ssl, and
da97433d5cd26e422a370d186f98659383c06721csovant mod_proxy are switched to the new logic.
da97433d5cd26e422a370d186f98659383c06721csovant However, ap_getline() still needs to be refactored out. But,
da97433d5cd26e422a370d186f98659383c06721csovant there's a problem there: ap_getline() peeks ahead for MIME
da97433d5cd26e422a370d186f98659383c06721csovant continuation (first character on line is space or \t) and
da97433d5cd26e422a370d186f98659383c06721csovant stores unused data in core_request_config which violates the
da97433d5cd26e422a370d186f98659383c06721csovant abstraction. That's cheating. So, we may not be able to
da97433d5cd26e422a370d186f98659383c06721csovant implement this without setting some data aside (yuck!).
da97433d5cd26e422a370d186f98659383c06721csovant I believe this is OtherBill's main complaint with the current
da97433d5cd26e422a370d186f98659383c06721csovant AIUI (correct me if I'm wrong!), OtherBill believes we
da97433d5cd26e422a370d186f98659383c06721csovant should have a pushback option so that we can return unread
da97433d5cd26e422a370d186f98659383c06721csovant data - this would solve this case. However, my question to
da97433d5cd26e422a370d186f98659383c06721csovant him is how do we handle stuff like mod_ssl - we can't "unread"
da97433d5cd26e422a370d186f98659383c06721csovant data. So, do we have two brigades for each filter? An in
da97433d5cd26e422a370d186f98659383c06721csovant brigade and a returned brigade? That seems messy. To
da97433d5cd26e422a370d186f98659383c06721csovant everyone else, can we refactor ap_getline() without pushback
da97433d5cd26e422a370d186f98659383c06721csovant - socket bucket and core input filter changes. see end of
da97433d5cd26e422a370d186f98659383c06721csovant message ID (Feb 27): <20010227075326.S2297@lyra.org>
da97433d5cd26e422a370d186f98659383c06721csovant - fix up ap_get_brigade() semantics, fix bug in DECHUNK /
da97433d5cd26e422a370d186f98659383c06721csovant ap_getline. many messages (plus their threads) (Apr/May):
da97433d5cd26e422a370d186f98659383c06721csovant Message-ID: <20010402101207.J27539@lyra.org>
da97433d5cd26e422a370d186f98659383c06721csovant Message-ID: <3AF7F921.D2EEC41A@algroup.co.uk>
da97433d5cd26e422a370d186f98659383c06721csovant Message-ID: <20010508190029.E18404@lyra.org>
da97433d5cd26e422a370d186f98659383c06721csovant - further work with combining/tweaking the builtin filters:
da97433d5cd26e422a370d186f98659383c06721csovant Message-ID: <20010509115445.D1374@lyra.org>
b168b33f591684db77dc1078f77947802d2fedf4gary.williams - thoughts on filter modes:
da97433d5cd26e422a370d186f98659383c06721csovant Message-ID: <021b01c14dee$09782af0$93c0b0d0@roweclan.net>
da97433d5cd26e422a370d186f98659383c06721csovant - proposal for getline rewrite:
da97433d5cd26e422a370d186f98659383c06721csovant Message-ID: <20011231104019.GD3904@ebuilt.com>
da97433d5cd26e422a370d186f98659383c06721csovant - patch for getline rewrite:
da97433d5cd26e422a370d186f98659383c06721csovant Message-ID: <20020122072605.GF28051@ebuilt.com>
da97433d5cd26e422a370d186f98659383c06721csovant - Change ap_get_brigade prototype:
da97433d5cd26e422a370d186f98659383c06721csovant Message-ID: <20020120115136.GI17601@ebuilt.com>
da97433d5cd26e422a370d186f98659383c06721csovant * Allow the DocumentRoot directive within <Location > scopes? This
da97433d5cd26e422a370d186f98659383c06721csovant allows the beloved (crusty) Alias /foo/ /somepath/foo/ followed
da97433d5cd26e422a370d186f98659383c06721csovant by a <Directory /somepath/foo> to become simply
da97433d5cd26e422a370d186f98659383c06721csovant <Location /foo/> DocumentRoot /somefile/foo (IMHO a bit more legible
da97433d5cd26e422a370d186f98659383c06721csovant and in-your-face.) DocumentRoot unset would be accepted [and would
da97433d5cd26e422a370d186f98659383c06721csovant not permit content to be served, only virtual resources such as
da97433d5cd26e422a370d186f98659383c06721csovant server-info or server-status.
da97433d5cd26e422a370d186f98659383c06721csovant This proposed change would _not_ depricate Alias.
b168b33f591684db77dc1078f77947802d2fedf4gary.williams * Win32: Rotatelogs sometimes is not terminated when Apache
b168b33f591684db77dc1078f77947802d2fedf4gary.williams goes down hard. FirstBill was looking at possibly tracking the
da97433d5cd26e422a370d186f98659383c06721csovant child's-child processes in the parent process.
da97433d5cd26e422a370d186f98659383c06721csovant OtherBill asks, wasn't this fixed?
da97433d5cd26e422a370d186f98659383c06721csovant * Win32: Add a simple hold console open patch (wait for close or
da97433d5cd26e422a370d186f98659383c06721csovant the ESC key, with a nice message) if the server died a bad
da97433d5cd26e422a370d186f98659383c06721csovant death (non-zero exit code) in console mode.
da97433d5cd26e422a370d186f98659383c06721csovant Resolution: bring forward same ugly hacks from 1.3.13-.20
da97433d5cd26e422a370d186f98659383c06721csovant * Port of mod_ssl to Apache 2.0:
da97433d5cd26e422a370d186f98659383c06721csovant The current porting state is summarized in modules/ssl/README. The
da97433d5cd26e422a370d186f98659383c06721csovant remaining work includes:
da97433d5cd26e422a370d186f98659383c06721csovant (1) stablizing/optimizing the SSL filter logic
da97433d5cd26e422a370d186f98659383c06721csovant (2) Enabling the various SSL caching mechanisms (shmcb, shmht)
da97433d5cd26e422a370d186f98659383c06721csovant (3) Enabling SSL extentions
da97433d5cd26e422a370d186f98659383c06721csovant (4) Trying to seperate the https filter logic from mod_ssl -
da97433d5cd26e422a370d186f98659383c06721csovant This is to facilitate other modules that wish to use the https
945930ece8ea73e5184836bf0bf3efad2e3e6c36gary.williams filter or the mod_ssl logic or both as required.
945930ece8ea73e5184836bf0bf3efad2e3e6c36gary.williams Justin: mod_ssl filter logic is redone, so that should be fine.
945930ece8ea73e5184836bf0bf3efad2e3e6c36gary.williams Madhu has submitted a patch for SSL caching - however, I
945930ece8ea73e5184836bf0bf3efad2e3e6c36gary.williams am -0 on that patch as I *think* we could implement the
da97433d5cd26e422a370d186f98659383c06721csovant shared memory another way that is much cleaner (i.e.
da97433d5cd26e422a370d186f98659383c06721csovant treat shmem directly as a dbm via APR routines). Justin
da97433d5cd26e422a370d186f98659383c06721csovant also thinks that the https filter logic may be sufficiently
da97433d5cd26e422a370d186f98659383c06721csovant decoupled now, but isn't really sure.
da97433d5cd26e422a370d186f98659383c06721csovant * Performance: Get the SINGLE_LISTEN_UNSERIALIZED_ACCEPT
da97433d5cd26e422a370d186f98659383c06721csovant optimization working in worker. prefork's new design for how
da97433d5cd26e422a370d186f98659383c06721csovant to notice data on the pod should be sufficient.
da97433d5cd26e422a370d186f98659383c06721csovant * Performance & Debug: Eliminate most (and perhaps all) of the
da97433d5cd26e422a370d186f98659383c06721csovant malloc/free calls in the bucket brigade code. Need some
da97433d5cd26e422a370d186f98659383c06721csovant light weight memory management functions that allow freeing
da97433d5cd26e422a370d186f98659383c06721csovant memory (putting it back into a memory pool) when it is no
da97433d5cd26e422a370d186f98659383c06721csovant longer needed. Enabling simple debugging features like guard
da97433d5cd26e422a370d186f98659383c06721csovant bands, double free detection, etc. would be cool but certainly
da97433d5cd26e422a370d186f98659383c06721csovant not a hard requirement.
da97433d5cd26e422a370d186f98659383c06721csovant Status: Cliff started to implement this using SMS as has
da97433d5cd26e422a370d186f98659383c06721csovant been discussed at length for months, but since
da97433d5cd26e422a370d186f98659383c06721csovant SMS is not being used anywhere else in the server,
da97433d5cd26e422a370d186f98659383c06721csovant several people expressed the opinion that we should
da97433d5cd26e422a370d186f98659383c06721csovant get rid of it entirely, meaning that the buckets
da97433d5cd26e422a370d186f98659383c06721csovant need their own memory management (free list) functions.
da97433d5cd26e422a370d186f98659383c06721csovant Cliff will implement that this weekend so we at least
da97433d5cd26e422a370d186f98659383c06721csovant have something to look at/compare with.
da97433d5cd26e422a370d186f98659383c06721csovant * Eliminate unnecessary creation of pipes in mod_cgid
da97433d5cd26e422a370d186f98659383c06721csovant * the autoconf setup should be fixed to default to using the
da97433d5cd26e422a370d186f98659383c06721csovant "Apache" layout from config.layout, and each variable settable
da97433d5cd26e422a370d186f98659383c06721csovant in a layout should be overridable on the command line. Plus,
da97433d5cd26e422a370d186f98659383c06721csovant what we do right now just doesn't seem to fully fit into how autoconf
da97433d5cd26e422a370d186f98659383c06721csovant works, eg. AC_PREFIX_DEFAULT issues.
da97433d5cd26e422a370d186f98659383c06721csovant Message-ID: <Pine.BSF.4.20.0104031557420.20876-100000@alive.znep.com>
da97433d5cd26e422a370d186f98659383c06721csovant * Combine log_child and piped_log_spawn. Clean up http_log.c.
da97433d5cd26e422a370d186f98659383c06721csovant Common logging API.
da97433d5cd26e422a370d186f98659383c06721csovant * Document mod_file_cache.
da97433d5cd26e422a370d186f98659383c06721csovant * OS/2: Make mod_status work for spmt_os2 MPM.
da97433d5cd26e422a370d186f98659383c06721csovant * Platforms that do not support fork (primarily Win32 and AS/400)
da97433d5cd26e422a370d186f98659383c06721csovant Architect start-up code that avoids initializing all the modules
da97433d5cd26e422a370d186f98659383c06721csovant in the parent process on platforms that do not support fork.
945930ece8ea73e5184836bf0bf3efad2e3e6c36gary.williams * Win32: Migrate the MPM over to use APR thread/process calls. This
945930ece8ea73e5184836bf0bf3efad2e3e6c36gary.williams would eliminate some code in the Win32 branch that essentially
2b68af05b4308b32506f1d68f6357cdc458869d7ludovicp duplicates what is in APR.
da97433d5cd26e422a370d186f98659383c06721csovant * There are still a number of places in the code where we are
da97433d5cd26e422a370d186f98659383c06721csovant losing error status (i.e. throwing away the error returned by a
da97433d5cd26e422a370d186f98659383c06721csovant system call and replacing it with a generic error code)
da97433d5cd26e422a370d186f98659383c06721csovant * Mass vhosting version of suEXEC.
945930ece8ea73e5184836bf0bf3efad2e3e6c36gary.williams * All DBMs suffer from confusion in support/dbmmanage (perl script) since
945930ece8ea73e5184836bf0bf3efad2e3e6c36gary.williams the dbmmanage employs the first-matched dbm format. This is not
945930ece8ea73e5184836bf0bf3efad2e3e6c36gary.williams necessarily the library that Apache was built with. Aught to
da97433d5cd26e422a370d186f98659383c06721csovant rewrite dbmmanage upon installation to bin/ with the proper library
da97433d5cd26e422a370d186f98659383c06721csovant for predictable mod_auth_db/dbm administration.
da97433d5cd26e422a370d186f98659383c06721csovant Status: Mladen Turk has posted several patches and ideas.
da97433d5cd26e422a370d186f98659383c06721csovant Key question, part of htpasswd, or a seperate utility?
da97433d5cd26e422a370d186f98659383c06721csovant prefer htpasswd: Lars
da97433d5cd26e422a370d186f98659383c06721csovant prefer seperate: OtherBill
da97433d5cd26e422a370d186f98659383c06721csovant * use apu_dbm in mod_auth_dbm
da97433d5cd26e422a370d186f98659383c06721csovant Status: Greg +1 (low-priority volunteer)
da97433d5cd26e422a370d186f98659383c06721csovant Justin says: "Seems like this is already there, so should we just
da97433d5cd26e422a370d186f98659383c06721csovant remove the other DBM code in that file? If you want
da97433d5cd26e422a370d186f98659383c06721csovant to use gdbm, or dbm, etc, you should tell apr-util."
da97433d5cd26e422a370d186f98659383c06721csovant Will says: "bs - I may choose the fastest - most efficient native
da97433d5cd26e422a370d186f98659383c06721csovant dbm implementation, for shared proc caches, ssl session
da97433d5cd26e422a370d186f98659383c06721csovant caching, etc, but that has nothing to do with maintaining
da97433d5cd26e422a370d186f98659383c06721csovant a userlist via dbm, which has to remain readable between
da97433d5cd26e422a370d186f98659383c06721csovant builds/machines, etc. The use-multiple database schema
da97433d5cd26e422a370d186f98659383c06721csovant for apr-util would let us do this with just apr, though."
da97433d5cd26e422a370d186f98659383c06721csovant Ian says: "multi-dbm is in, but it still has ndbm support hardcoded
da97433d5cd26e422a370d186f98659383c06721csovant is this still required? isn't ndbm supported via gdbm?"
da97433d5cd26e422a370d186f98659383c06721csovant * Integrate mod_dav.
da97433d5cd26e422a370d186f98659383c06721csovant Some additional items remaining:
da97433d5cd26e422a370d186f98659383c06721csovant - case_preserved_filename stuff
da97433d5cd26e422a370d186f98659383c06721csovant (use the new canonical name stuff?)
da97433d5cd26e422a370d186f98659383c06721csovant - find a new home for ap_text(_header)
da97433d5cd26e422a370d186f98659383c06721csovant - is it possible to remove the DAV: namespace stuff from util_xml?
da97433d5cd26e422a370d186f98659383c06721csovant * ap_core_translate() and its use by mod_mmap_static and mod_file_cache
da97433d5cd26e422a370d186f98659383c06721csovant are a bit wonky. The function should probably be exposed as a utility
da97433d5cd26e422a370d186f98659383c06721csovant function (such as ap_translate_url2fs() or ap_validate_fs_url() or
da97433d5cd26e422a370d186f98659383c06721csovant something). Another approach would be a new hook phase after
da97433d5cd26e422a370d186f98659383c06721csovant "translate" which would allow the module to munge what the
da97433d5cd26e422a370d186f98659383c06721csovant translation has decided to do.
da97433d5cd26e422a370d186f98659383c06721csovant Status: Greg +1 (volunteers), Ryan +1
da97433d5cd26e422a370d186f98659383c06721csovant * Explore use of a post-config hook for the code in http_main.c which
da97433d5cd26e422a370d186f98659383c06721csovant calls ap_fixup_virutal_hosts(), ap_fini_vhost_config(), and
da97433d5cd26e422a370d186f98659383c06721csovant ap_sort_hooks() [to reduce the logic in main()]
da97433d5cd26e422a370d186f98659383c06721csovant * read the config tree just once, and process N times (as necessary)
da97433d5cd26e422a370d186f98659383c06721csovant * (possibly) use UUIDs in mod_unique_id and/or mod_usertrack
da97433d5cd26e422a370d186f98659383c06721csovant * (possibly) port the bug fix for PR 6942 (segv when LoadModule is put
da97433d5cd26e422a370d186f98659383c06721csovant into a VirtualHost container) to 2.0.
da97433d5cd26e422a370d186f98659383c06721csovant * shift stuff to mod_core.h
da97433d5cd26e422a370d186f98659383c06721csovant * callers of ap_run_create_request() should check the return value
da97433d5cd26e422a370d186f98659383c06721csovant for failure (Doug volunteers)
da97433d5cd26e422a370d186f98659383c06721csovant * Win32: Get Apache working on Windows 95/98. The following work
da97433d5cd26e422a370d186f98659383c06721csovant (at least) needs to be done:
da97433d5cd26e422a370d186f98659383c06721csovant - winnt MPM: Fix 95/98 code paths in the winnt MPM. There is some NT
da97433d5cd26e422a370d186f98659383c06721csovant specific code that is still not in NT only code paths
da97433d5cd26e422a370d186f98659383c06721csovant - IOL binds to APR sendfile, implemented with TransmitFile, which
da97433d5cd26e422a370d186f98659383c06721csovant is not available on 95/98.
da97433d5cd26e422a370d186f98659383c06721csovant - Document warning that OSR2 is required (for Crypt functions, in
da97433d5cd26e422a370d186f98659383c06721csovant rand.c, at least.) This could be resolved with an SSL library, or
da97433d5cd26e422a370d186f98659383c06721csovant randomization in APR itself.
da97433d5cd26e422a370d186f98659383c06721csovant - Bring the Win9xConHook.dll from 1.3 into 2.0 (no sense till it
da97433d5cd26e422a370d186f98659383c06721csovant actually works) and add in a splash of Win9x service code.
da97433d5cd26e422a370d186f98659383c06721csovant * In order to use a DSO version of mod_ssl we have to link with
da97433d5cd26e422a370d186f98659383c06721csovant -lssl and -lcrypto. A workaround is in place right now where the
da97433d5cd26e422a370d186f98659383c06721csovant entire EXTRA_LIBS macro is being appended to the objects list, but
da97433d5cd26e422a370d186f98659383c06721csovant this is a hack. We should either revamp the APACHE_CHECK_SSL_TOOLKIT
da97433d5cd26e422a370d186f98659383c06721csovant autoconf function or come up with some other autoconf checks to
da97433d5cd26e422a370d186f98659383c06721csovant search for libssl and libcrypto and properly add them to mod_ssl's
da97433d5cd26e422a370d186f98659383c06721csovant link flags.
da97433d5cd26e422a370d186f98659383c06721csovant * Make the worker MPM the default MPM for threaded Unix boxes.
da97433d5cd26e422a370d186f98659383c06721csovant +1: Justin, Jeff, Ian
da97433d5cd26e422a370d186f98659383c06721csovant -0: Aaron (premature decision, needs more discussion), Lars
da97433d5cd26e422a370d186f98659383c06721csovant -0: Cliff (I think the default config should be the safest possible)
da97433d5cd26e422a370d186f98659383c06721csovant * Fix the worker MPM to use POD to kill child processes instead
da97433d5cd26e422a370d186f98659383c06721csovant of ap_os_killpg, regardless of how they should die. (Ryan Bloom)
da97433d5cd26e422a370d186f98659383c06721csovantPRs that have been suspended forever waiting for someone to
da97433d5cd26e422a370d186f98659383c06721csovantput them into 'the next release':
b168b33f591684db77dc1078f77947802d2fedf4gary.williams * PR#76: general
b168b33f591684db77dc1078f77947802d2fedf4gary.williams missing call to "setlocale();"
da97433d5cd26e422a370d186f98659383c06721csovant * PR#78: mod_include
945930ece8ea73e5184836bf0bf3efad2e3e6c36gary.williams Additional status for XBitHack directive
945930ece8ea73e5184836bf0bf3efad2e3e6c36gary.williams * PR#362: mod_proxy
945930ece8ea73e5184836bf0bf3efad2e3e6c36gary.williams Mod_proxy doesn't allow change of error pages
945930ece8ea73e5184836bf0bf3efad2e3e6c36gary.williams * PR#370: mod_env
945930ece8ea73e5184836bf0bf3efad2e3e6c36gary.williams Modified PATH environemnt variable is not passed, instead
28eeffb99d4c7f8ff501c22d1d21501cbfe3ad32csovant system's is used
da97433d5cd26e422a370d186f98659383c06721csovant * PR#440: mod_proxy
da97433d5cd26e422a370d186f98659383c06721csovant Proxy doesn't deliver documents if not connected
2b68af05b4308b32506f1d68f6357cdc458869d7ludovicp * PR#534: mod_proxy
2b68af05b4308b32506f1d68f6357cdc458869d7ludovicp proxy converts ~name to %7Ename when name starts with a dot (.)
945930ece8ea73e5184836bf0bf3efad2e3e6c36gary.williams * PR#537: mod_access
945930ece8ea73e5184836bf0bf3efad2e3e6c36gary.williams mod_access syntax allows hosts that should be restricted
28eeffb99d4c7f8ff501c22d1d21501cbfe3ad32csovant * PR#557: mod_auth-any
da97433d5cd26e422a370d186f98659383c06721csovant ~UserHome directories are not honored in absolute pathname
da97433d5cd26e422a370d186f98659383c06721csovant requests (.htaccess)
da97433d5cd26e422a370d186f98659383c06721csovant * PR#612: mod_proxy
da97433d5cd26e422a370d186f98659383c06721csovant Proxy FTP Authentication Fails
945930ece8ea73e5184836bf0bf3efad2e3e6c36gary.williams * PR#623: mod_include
2b68af05b4308b32506f1d68f6357cdc458869d7ludovicp A smarter "Last Modified" value for SSI documents (see PR number 600)
da97433d5cd26e422a370d186f98659383c06721csovant * PR#628: config
da97433d5cd26e422a370d186f98659383c06721csovant Request of "Options SymLinksIfGroupMatch"
b168b33f591684db77dc1078f77947802d2fedf4gary.williams * PR#700: mod_proxy
da97433d5cd26e422a370d186f98659383c06721csovant Proxy doesn't do links right for OpenVMS files through ftp:
da97433d5cd26e422a370d186f98659383c06721csovant * PR#759: mod_imap
da97433d5cd26e422a370d186f98659383c06721csovant imap should read <MAP><AREA>*</MAP> too!
da97433d5cd26e422a370d186f98659383c06721csovant * PR#793: general
da97433d5cd26e422a370d186f98659383c06721csovant RLimitCPU and RLimitMEM don't apply to all children like they should
da97433d5cd26e422a370d186f98659383c06721csovant * PR#921: suexec
da97433d5cd26e422a370d186f98659383c06721csovant Uses cwd before filling it in, doesn't use syslog
da97433d5cd26e422a370d186f98659383c06721csovant * PR#922: config
da97433d5cd26e422a370d186f98659383c06721csovant it is useful to allow specifiction that root-owned symlinks
da97433d5cd26e422a370d186f98659383c06721csovant should always be followed
da97433d5cd26e422a370d186f98659383c06721csovant * PR#980: mod_proxy
da97433d5cd26e422a370d186f98659383c06721csovant Controlling Access to Remote Proxies would be nice...
da97433d5cd26e422a370d186f98659383c06721csovant * PR#994: mod_proxy
da97433d5cd26e422a370d186f98659383c06721csovant Adding authentication "on the fly" through the proxy module
da97433d5cd26e422a370d186f98659383c06721csovant * PR#1004: apache-api
da97433d5cd26e422a370d186f98659383c06721csovant request_config field in request_rec is moderately bogus
da97433d5cd26e422a370d186f98659383c06721csovant * PR#1028: other
da97433d5cd26e422a370d186f98659383c06721csovant DoS attacks involving memory consumption
da97433d5cd26e422a370d186f98659383c06721csovant * PR#1050: mod_log-any
da97433d5cd26e422a370d186f98659383c06721csovant Logging of virtual server to error_log as well
da97433d5cd26e422a370d186f98659383c06721csovant * PR#1085: mod_proxy
da97433d5cd26e422a370d186f98659383c06721csovant ProxyRemote make a dead cycle.
da97433d5cd26e422a370d186f98659383c06721csovant * PR#1117: mod_auth-any
da97433d5cd26e422a370d186f98659383c06721csovant Using NIS passwd.byname dbm files with AuthDBMUserFile
da97433d5cd26e422a370d186f98659383c06721csovant * PR#1120: suexec
da97433d5cd26e422a370d186f98659383c06721csovant suexec does not parse arguments to #exec cmd
da97433d5cd26e422a370d186f98659383c06721csovant * PR#1145: mod_include
da97433d5cd26e422a370d186f98659383c06721csovant Allow for Last-Modified: without resorting to XBitHack
da97433d5cd26e422a370d186f98659383c06721csovant * PR#1158: apache-api
da97433d5cd26e422a370d186f98659383c06721csovant improvements to child spawning API
da97433d5cd26e422a370d186f98659383c06721csovant * PR#1166: mod_proxy
da97433d5cd26e422a370d186f98659383c06721csovant ``nph-'' not honored (no buffering) for ProxyRemote mapping
da97433d5cd26e422a370d186f98659383c06721csovant * PR#1176: mod_cgi
da97433d5cd26e422a370d186f98659383c06721csovant Apache cannot handle continuation line in headers
da97433d5cd26e422a370d186f98659383c06721csovant * PR#1191: general
da97433d5cd26e422a370d186f98659383c06721csovant setlogin() is not called, causing problems with e.g. identd
da97433d5cd26e422a370d186f98659383c06721csovant * PR#1204: general
da97433d5cd26e422a370d186f98659383c06721csovant regerror() exists, use it
da97433d5cd26e422a370d186f98659383c06721csovant * PR#1233: apache-api
da97433d5cd26e422a370d186f98659383c06721csovant there is no way to keep per-connection per-module state
da97433d5cd26e422a370d186f98659383c06721csovant * PR#1263: mod_autoexec
da97433d5cd26e422a370d186f98659383c06721csovant Add frame-safe anchor attribute to mod_autoindex links
da97433d5cd26e422a370d186f98659383c06721csovant * PR#1268: suexec
da97433d5cd26e422a370d186f98659383c06721csovant CGI scripts running as Apache user: security (suexec etc.)
da97433d5cd26e422a370d186f98659383c06721csovant * PR#1285: suexec
da97433d5cd26e422a370d186f98659383c06721csovant Error messages could be easier to spot in cgi.log file for suexec.c
da97433d5cd26e422a370d186f98659383c06721csovant * PR#1287: mod_access
da97433d5cd26e422a370d186f98659383c06721csovant add allow,deny/deny,allow warning to mod_access
da97433d5cd26e422a370d186f98659383c06721csovant * PR#1290: mod_proxy
da97433d5cd26e422a370d186f98659383c06721csovant Need to know "hit-rate" on proxy cache
da97433d5cd26e422a370d186f98659383c06721csovant * PR#1358: mod_log-any
da97433d5cd26e422a370d186f98659383c06721csovant Selective url-encode of log fields (or maybe a pseudo
da97433d5cd26e422a370d186f98659383c06721csovant log_rewrite module?)
da97433d5cd26e422a370d186f98659383c06721csovant * PR#1383: mod_headers
da97433d5cd26e422a370d186f98659383c06721csovant I make mod_headers to modify request headers as well as
da97433d5cd26e422a370d186f98659383c06721csovant response ones.
da97433d5cd26e422a370d186f98659383c06721csovant * PR#1532: mod_proxy
da97433d5cd26e422a370d186f98659383c06721csovant Proxy transfer logging
da97433d5cd26e422a370d186f98659383c06721csovant * PR#1547: mod_proxy
da97433d5cd26e422a370d186f98659383c06721csovant No HTTP_X_FORWARDED_FOR set...
da97433d5cd26e422a370d186f98659383c06721csovant * PR#1567: mod_proxy
da97433d5cd26e422a370d186f98659383c06721csovant ProxyRemote proxy requests fail authentication by firewall
da97433d5cd26e422a370d186f98659383c06721csovant * PR#1582: mod_rewrite
da97433d5cd26e422a370d186f98659383c06721csovant mod_rewrite forms REQUEST_URI different than mod_cgi does
da97433d5cd26e422a370d186f98659383c06721csovant * PR#1677: mod_headers
da97433d5cd26e422a370d186f98659383c06721csovant mod_headers should allow mod_log_config-style formats in
da97433d5cd26e422a370d186f98659383c06721csovant header values
da97433d5cd26e422a370d186f98659383c06721csovant * PR#1702: mod_proxy
da97433d5cd26e422a370d186f98659383c06721csovant mod_proxy to support persistent conns?
da97433d5cd26e422a370d186f98659383c06721csovant * PR#1803: mod_include
da97433d5cd26e422a370d186f98659383c06721csovant patches to mod_include to allow for file tests
da97433d5cd26e422a370d186f98659383c06721csovant * PR#1809: mod_auth-any
da97433d5cd26e422a370d186f98659383c06721csovant Suggestion for improving authentication modules and core source
da97433d5cd26e422a370d186f98659383c06721csovant code, problem with 401 and ErrorDocument
da97433d5cd26e422a370d186f98659383c06721csovant * PR#1878: mod_proxy
da97433d5cd26e422a370d186f98659383c06721csovant listing of proxy cache content
da97433d5cd26e422a370d186f98659383c06721csovant * PR#1905: suexec
da97433d5cd26e422a370d186f98659383c06721csovant Allow modules to set user:group for execution.
da97433d5cd26e422a370d186f98659383c06721csovant * PR#2024: apache-api
da97433d5cd26e422a370d186f98659383c06721csovant adding auth_why to conn_rec
da97433d5cd26e422a370d186f98659383c06721csovant * PR#2073: mod_log-any
da97433d5cd26e422a370d186f98659383c06721csovant pipelined connections are not logged correctly
da97433d5cd26e422a370d186f98659383c06721csovant * PR#2074: mod_rewrite
da97433d5cd26e422a370d186f98659383c06721csovant mod_rewrite doesn't pass Proxy Throughput on internal subrequests
da97433d5cd26e422a370d186f98659383c06721csovant * PR#2113: config
da97433d5cd26e422a370d186f98659383c06721csovant HTTP Server Rebuild Line Needs Changing for the better
da97433d5cd26e422a370d186f98659383c06721csovant * PR#2138: mod_status
da97433d5cd26e422a370d186f98659383c06721csovant mod_status always displays 256 possible connection slots
da97433d5cd26e422a370d186f98659383c06721csovant * PR#2221: documentation
da97433d5cd26e422a370d186f98659383c06721csovant Make online documentation search link back to my installation
da97433d5cd26e422a370d186f98659383c06721csovant * PR#2284: general
da97433d5cd26e422a370d186f98659383c06721csovant Can not POST to ErrorDocument - Apache/1.3b6
da97433d5cd26e422a370d186f98659383c06721csovant * PR#2314: mod_proxy
da97433d5cd26e422a370d186f98659383c06721csovant patterns in ProxyRemote
da97433d5cd26e422a370d186f98659383c06721csovant * PR#2343: mod_status
da97433d5cd26e422a370d186f98659383c06721csovant Status module averages are for entire uptime
da97433d5cd26e422a370d186f98659383c06721csovant * PR#2360: suexec
da97433d5cd26e422a370d186f98659383c06721csovant suexec for general access of user content?
da97433d5cd26e422a370d186f98659383c06721csovant * PR#2396: general
da97433d5cd26e422a370d186f98659383c06721csovant Proposal for TimeZone directive
da97433d5cd26e422a370d186f98659383c06721csovant * PR#2415: mod_info
da97433d5cd26e422a370d186f98659383c06721csovant /server-info doesn't check for the virtual host to list the info
da97433d5cd26e422a370d186f98659383c06721csovant * PR#2421: config
da97433d5cd26e422a370d186f98659383c06721csovant problem specifying ndbm library for build ?with autoconfigure
da97433d5cd26e422a370d186f98659383c06721csovant * PR#2431: general
da97433d5cd26e422a370d186f98659383c06721csovant A small addition to rotatelogs.c to improve program functionality.
da97433d5cd26e422a370d186f98659383c06721csovant * PR#2446: config
da97433d5cd26e422a370d186f98659383c06721csovant AllowOverride FileInfo is too coarse
da97433d5cd26e422a370d186f98659383c06721csovant * PR#2460: mod_cgi
da97433d5cd26e422a370d186f98659383c06721csovant TimeOut applies to output of CGI scripts
da97433d5cd26e422a370d186f98659383c06721csovant * PR#2512: mod_access
da97433d5cd26e422a370d186f98659383c06721csovant <IfDenied> directive wanted
da97433d5cd26e422a370d186f98659383c06721csovant * PR#2573: suexec
da97433d5cd26e422a370d186f98659383c06721csovant CGI's for general use still have to be run as another user
da97433d5cd26e422a370d186f98659383c06721csovant with suExec
da97433d5cd26e422a370d186f98659383c06721csovant * PR#2648: general
da97433d5cd26e422a370d186f98659383c06721csovant Cache file names in Proxy module
da97433d5cd26e422a370d186f98659383c06721csovant * PR#2760: config
da97433d5cd26e422a370d186f98659383c06721csovant [PATCH] User/Group for <Directory> and <Location> i.e. not only
da97433d5cd26e422a370d186f98659383c06721csovant in global and <Virtual>.
da97433d5cd26e422a370d186f98659383c06721csovant * PR#2763: general
945930ece8ea73e5184836bf0bf3efad2e3e6c36gary.williams mailto tags and bundling bug report script
945930ece8ea73e5184836bf0bf3efad2e3e6c36gary.williams * PR#2785: os-aix
945930ece8ea73e5184836bf0bf3efad2e3e6c36gary.williams Support for System Resource Controller
da97433d5cd26e422a370d186f98659383c06721csovant * PR#2793: protocol
da97433d5cd26e422a370d186f98659383c06721csovant When will Apache support P3P? Any Plans?
da97433d5cd26e422a370d186f98659383c06721csovant * PR#2873: config
945930ece8ea73e5184836bf0bf3efad2e3e6c36gary.williams * PR#2889: general
da97433d5cd26e422a370d186f98659383c06721csovant Inclusion of RPM spec file in CVS/distributions
da97433d5cd26e422a370d186f98659383c06721csovant * PR#2906: general
da97433d5cd26e422a370d186f98659383c06721csovant Propose that Apache recommend $UNIQUE_ID for all "session id"
da97433d5cd26e422a370d186f98659383c06721csovant * PR#2907: config
da97433d5cd26e422a370d186f98659383c06721csovant suggestion: power up your Include directive :)
da97433d5cd26e422a370d186f98659383c06721csovant * PR#3018: general
da97433d5cd26e422a370d186f98659383c06721csovant cannot limit some HTTP methods
da97433d5cd26e422a370d186f98659383c06721csovant * PR#3143: apache-api
da97433d5cd26e422a370d186f98659383c06721csovant No module specific data hook for per-connection data
da97433d5cd26e422a370d186f98659383c06721csovant * PR#3191: mod_negotiation
da97433d5cd26e422a370d186f98659383c06721csovant no way to set global quality-of-source (qs) coneg values
da97433d5cd26e422a370d186f98659383c06721csovant with multiviews
da97433d5cd26e422a370d186f98659383c06721csovant * PR#3568: mod_proxy
da97433d5cd26e422a370d186f98659383c06721csovant Accessing URL through proxy server corrupts data.
da97433d5cd26e422a370d186f98659383c06721csovant * PR#3605: mod_proxy
da97433d5cd26e422a370d186f98659383c06721csovant Some anonymous FTP URLs ask for authentication
da97433d5cd26e422a370d186f98659383c06721csovant * PR#3677: general
da97433d5cd26e422a370d186f98659383c06721csovant New ErrorDocumentMatch directive
da97433d5cd26e422a370d186f98659383c06721csovant * PR#4241: config
da97433d5cd26e422a370d186f98659383c06721csovant Need to be able to override shebang line to make CGI scripts
da97433d5cd26e422a370d186f98659383c06721csovant more portable.
da97433d5cd26e422a370d186f98659383c06721csovant * PR#4244: config
da97433d5cd26e422a370d186f98659383c06721csovant "Files" and "FilesMatch" regexp does not recognize bang as
da97433d5cd26e422a370d186f98659383c06721csovant negation operator
da97433d5cd26e422a370d186f98659383c06721csovant * PR#4448: mod_log-any
da97433d5cd26e422a370d186f98659383c06721csovant Please allow CGI env variables (QUERY_STRING, ...) to be logged
da97433d5cd26e422a370d186f98659383c06721csovant * PR#4459: mod_include
da97433d5cd26e422a370d186f98659383c06721csovant Suggestion for better handling of Last-modified headers
945930ece8ea73e5184836bf0bf3efad2e3e6c36gary.williams * PR#4490: mod_cgi
945930ece8ea73e5184836bf0bf3efad2e3e6c36gary.williams mod_cgi prevents handling of OPTIONS requests
945930ece8ea73e5184836bf0bf3efad2e3e6c36gary.williams * PR#5713: os-windows
da97433d5cd26e422a370d186f98659383c06721csovant [PATCH] install as win32 service with domain account
da97433d5cd26e422a370d186f98659383c06721csovant * PR#5993: general
da97433d5cd26e422a370d186f98659383c06721csovant AllowOverride should have a 'CheckNone' and 'AllowNone' argument
da97433d5cd26e422a370d186f98659383c06721csovant instead of only 'None'
945930ece8ea73e5184836bf0bf3efad2e3e6c36gary.williamsOther bugs that need fixing:
da97433d5cd26e422a370d186f98659383c06721csovant * MaxRequestsPerChild measures connections, not requests.
da97433d5cd26e422a370d186f98659383c06721csovant Until someone has a better way, we'll probably just rename it
da97433d5cd26e422a370d186f98659383c06721csovant "MaxConnectionsPerChild".
da97433d5cd26e422a370d186f98659383c06721csovant * Regex containers don't work in an intutive way
da97433d5cd26e422a370d186f98659383c06721csovant Status: No one has come up with an efficient way to fix this
da97433d5cd26e422a370d186f98659383c06721csovant behavior. Dean has suggested getting rid of regex containers
da97433d5cd26e422a370d186f98659383c06721csovant completely.
da97433d5cd26e422a370d186f98659383c06721csovant * SIGSEGV on Linux (glibc 2.1.2) isn't caught properly by a
da97433d5cd26e422a370d186f98659383c06721csovant sigwaiting thread. We need to work around this, perhaps unless
da97433d5cd26e422a370d186f98659383c06721csovant there is hope soon for a fixed glibc.
da97433d5cd26e422a370d186f98659383c06721csovant * orig_ct in the byterange/multipart handling may not be
da97433d5cd26e422a370d186f98659383c06721csovant needed. Apache 1.3 just never stashed "multipart" into
da97433d5cd26e422a370d186f98659383c06721csovant r->content_type. We should probably follow suit since the
da97433d5cd26e422a370d186f98659383c06721csovant byterange stuff doesn't want the rest of the code to see the
da97433d5cd26e422a370d186f98659383c06721csovant multipart content-type; the other code should still think it is
da97433d5cd26e422a370d186f98659383c06721csovant dealing with the <orig_ct> stuff.
da97433d5cd26e422a370d186f98659383c06721csovant Status: Greg volunteers to investigate (esp. since he was most
da97433d5cd26e422a370d186f98659383c06721csovant likely the one to break it :-)
da97433d5cd26e422a370d186f98659383c06721csovantOther features that need writing:
da97433d5cd26e422a370d186f98659383c06721csovant * Finish infrastructure in core for async MPMs
da97433d5cd26e422a370d186f98659383c06721csovant Status: post 2.0
da97433d5cd26e422a370d186f98659383c06721csovant * TODO in source -- just do an egrep on "TODO" and see what's there
da97433d5cd26e422a370d186f98659383c06721csovantAvailable Patches:
da97433d5cd26e422a370d186f98659383c06721csovant * Jon Travis's <jtravis@covalent.net> patch to deal with thread-safe
da97433d5cd26e422a370d186f98659383c06721csovant issues with inet_ntoa. See message <20001201163220.A12827@covalent.net>
da97433d5cd26e422a370d186f98659383c06721csovant Status: This is being set aside until the IPv6 work is finished
da97433d5cd26e422a370d186f98659383c06721csovant so that we know exactly what is required.
da97433d5cd26e422a370d186f98659383c06721csovant * Martin Sojka <msojka@gmx.de>'s patch to add error reporting for failed
da97433d5cd26e422a370d186f98659383c06721csovant htpasswd actions due to a full /tmp volume (other programs may have
da97433d5cd26e422a370d186f98659383c06721csovant similar problems?)
da97433d5cd26e422a370d186f98659383c06721csovant * Mike Abbott's <mja@trudge.engr.sgi.com> patches to improve
da97433d5cd26e422a370d186f98659383c06721csovant performance
da97433d5cd26e422a370d186f98659383c06721csovant Status: These were written for 1.3, and are awaiting a port to
da97433d5cd26e422a370d186f98659383c06721csovant * Jim Winstead's <jimw@trainedmonkey.com> patch to add CookieDomain and
da97433d5cd26e422a370d186f98659383c06721csovant other small mod_usertrack features
da97433d5cd26e422a370d186f98659383c06721csovant * Dan Rench's <drench@xnet.com> patch to add allow the errmsg and timefmt
da97433d5cd26e422a370d186f98659383c06721csovant of SSI's to be modified in the config file. Patch is available in
da97433d5cd26e422a370d186f98659383c06721csovantOpen issues:
da97433d5cd26e422a370d186f98659383c06721csovant * Which MPMs will be included with Apache 2.0?