2N/AAPACHE 2.0 STATUS: -*-text-*-
2N/ALast modified at [$Date: 2002/02/06 08:55:12 $]
2N/A 2.0.32 : in development
2N/A 2.0.31 : rolled Feburary 1, 2002.
2N/A 2.0.30 : tagged January 8, 2002. not rolled.
2N/A 2.0.29 : tagged November 27, 2001. not rolled.
2N/A 2.0.28 : released November 13, 2001
2N/A 2.0.27 : rolled November 6, 2001
2N/A 2.0.26 : tagged October 16, 2001. not rolled.
2N/A 2.0.25 : rolled August 29, 2001
2N/A 2.0.24 : rolled August 18, 2001
2N/A 2.0.23 : rolled August 9, 2001
2N/A 2.0.22 : rolled July 29, 2001
2N/A 2.0.21 : rolled July 20, 2001
2N/A 2.0.20 : rolled July 8, 2001
2N/A 2.0.19 : rolled June 27, 2001
2N/A 2.0.18 : rolled May 18, 2001
2N/A 2.0.17 : rolled April 17, 2001
2N/A 2.0.16 : rolled April 4, 2001
2N/A 2.0.15 : rolled March 21, 2001
2N/A 2.0.14 : rolled March 7, 2001
2N/A 2.0a9 : released December 12, 2000
2N/A 2.0a8 : released November 20, 2000
2N/A 2.0a7 : released October 8, 2000
2N/A 2.0a6 : released August 18, 2000
2N/A 2.0a5 : released August 4, 2000
2N/A 2.0a4 : released June 7, 2000
2N/A 2.0a3 : released April 28, 2000
2N/A 2.0a2 : released March 31, 2000
2N/A 2.0a1 : released March 10, 2000
2N/APlease consult the following STATUS files for information
2N/ARELEASE SHOWSTOPPERS:
2N/A running on Daedalus since 02-Feb-2002 7:58 PST (need 3 days)
2N/A Compiles on : AIX 4.3, Solaris, FreeBSD 3.4 & 4.5, Win32,
2N/A Linux 2.2 & 2.4, HPUX
2N/A Broken on: Win32 [no error logging within service, other bugs]
2N/A +0 : Lars, Justin, trawick
2N/A -1 : BillS, Ian, gregames, BrianP
2N/A bumps since original tag:
2N/A * scoreboard x2 : 1 to fix gracefull restarts
2N/A Jeff says: We can't do anything about libtool since AIX
2N/A needs a version that won't work on some
2N/A platforms. Handle this in the README.
2N/A The binbuild issue isn't AIX. It is
2N/A something that could happen anywhere that
2N/A the binbuild-er has their own expat.
2N/A beta roll or putting a patch in the README for
2N/A use by people who want to do binbuild but
2N/A have expat installed locally.
2N/A gregames: why can't we roll a second tarball with the
2N/A appropriate libtool version? doesn't
2N/A Darwin need it as well as AIX?
2N/A * erroneous check in an AP_DEBUG_ASSERT() call. Only happens
2N/A revision 1.391. Add item in release notes:
2N/A +1: Justin, Cliff, Aaron, BillS, Jim, trawick
2N/A * seg faults in core_input_filter when the client goes away
2N/A before any POST body bytes are received.
2N/A annoying error message)
2N/A * FirstBill reports problem [re]starting as-a-service, shared
2N/A score is suspect. OtherBill is investigating... has found
2N/A . created restart and shutdown events, only restart
2N/A survived initialization on XP, although breaking into
2N/A the debugger interferes with reproducing the bug.
2N/A Perhaps in FirstBill's example _restart didn't survive.
2N/A . Scoreboard appears irrelevant to the problem.
2N/A . This looks like handle corruption in
NT/XP 2N/A . Args are not initialized correctly when the -k install
2N/A OtherBill will not create Win32 binaries due to this bug.
2N/A * mod_auth_dbm can't open a Berkeley DB password file on Unix
2N/A Justin postulates that this might be related to the fact
2N/A that mod_auth_dbm wasn't using apr-util. See
2N/A variant was not properly detected. See
2N/A * Madhu reports that the worker MPM doesn't shutdown nicely
2N/A after a stress test.
2N/A * mod_autoindex displays the wrong icon for subdirectories
2N/A * anon shared memory not allocating enough
2N/A * other BETA possibilities
2N/A 31 + fixes for all non-cosmetic problems seen on daedalus
2N/A * ap_directory_walk skips some per-dir config merge functions
2N/A if there is no "<Directory />" block in the configuration
2N/A Message-ID: <m3itbdiijq.fsf@rdu163-40-092.nc.rr.com>
2N/A * That is very unlikely, merges are additive. Much more likely,
2N/A the default SetOutputFilter default or merge is borked.
2N/A Unless it's the code that permits 'null' merges, per module.
2N/A Still, it's probably in SetOutputFilter's behavior.
2N/A * BrianP notes: directory_walk's handling of trailing slashes
2N/A * Test suite failures:
2N/A o worker is also failing some of the 'cgi' subtests
2N/A Justin says: "Worker should be fine and passes httpd-test here.
2N/A If you can provide evidence that it can be reproduced
2N/A outside of httpd-test, then it's a showstopper. I
2N/A think it's a perl or a httpd-test problem."
2N/A Not a showstopper: Justin
2N/A * If any request gets to the core handler, without a flag that this
2N/A end of the ap_process_request_internal() processing. This provides
2N/A authors of older modules better compatibility, while still improving
2N/A the security and robustness of 2.0.
2N/A Status: still need to decide where this goes, OtherBill comments...
2N/A Message-ID: <065701c14526$495203b0$96c0b0d0@roweclan.net>
2N/A we need to look at halting this in the 'default handler' case,
2N/A and that implies pushing the 'handler election' into the request
2N/A internal processing phase from the run request phase.
2N/A * There is a bug in how we sort some hooks, at least the pre-config
2N/A hook. The first time we call the hooks, they are in the correct
2N/A order, but the second time, we don't sort them correctly. Currently,
2N/A back when this is fixed. rbb
2N/A Justin says: "Is this really a showstopper? This has been here
2N/A forever. What's wrong? Does this have to do with
2N/A Not a showstopper: Justin, BrianP, trawick, gregames
2N/A * Convert all instances of the old apr_lock_t type to the new
2N/A types (once they are fully supported in APR).
2N/A Status: Aaron is working on converting INTRAPROCESS
2N/A to apr_thread_mutex_t types. Full replacements for
2N/A LOCKALL and CROSS_PROCESS are not yet complete on all
2N/A platforms, and should only be used in MPMs like worker
2N/A with limited OS exposure.
2N/A BrianP asks: "Is this really a showstopper?"
2N/A IanH says: "If we don't do it before we go live we will have 2
2N/A different API's to support in APR"
2N/A * A binbuild installation picks up the right libraries when
2N/A running apachectl because we set the appropriate environment
2N/A variable, but ab, htpasswd, etc. don't know how to pick up apr,
2N/A Message-ID: <20020116000226.GA15991@ebuilt.com>
2N/A Justin says: "You could always build all of the binaries
2N/A statically if you don't want to let libtool
2N/A handle it. At most, I don't consider this a
2N/A a release showstopper as it is a packaging issue."
2N/A Ken sez: "Showstopper because, until it's the default, it
2N/A violates the P of LA. Binaries will give strange
2N/A errors and not run if moved."
2N/A Status: Aaron volunteers: I'll commit a patch to allow all
2N/A the support binaries to be built statically against
2N/A gregames says: what about httpd? yeah, it works at the moment
2N/A if you use apachectl. But I've already heard
2N/A complaints about not being able to execute
2N/ARELEASE NON-SHOWSTOPPERS BUT WOULD BE REAL NICE TO WRAP THESE UP:
2N/A administrator to order filters, beyond the order of filename (mime)
2N/A filename extensions. At minimum, some sort of +-[0-10] syntax seems
2N/A like the quickest fix for a 2.0 gold release.
2N/A Justin says: "Could we delay this for a point release or 2.1?"
2N/A Not a showstopper: justin, wrowe, trawick, stoddard, Jim, Ian, Aaron,
2N/A * Should we always build binaries statically unless otherwise
2N/A Message-ID: <20020129210006.B23512@Lithium.MeepZor.Com>
2N/A * If the parent process dies, should the remaining child processes
2N/A "gracefully" self-terminate. Or maybe we should make it a runtime
2N/A option, or have a concept of 2 parent processes (one being a
2N/A See: Message-ID: <3C58232C.FE91F19F@Golux.Com>
2N/A Self-destruct: Ken, Martin
2N/A Not self-destruct: BrianP, Ian, Cliff, BillS
2N/A Make it runtime configurable: Aaron, Jim, Justin
2N/A Have 2 parents: +1: Jim
2N/A +0: Martin (while standing by, could it do
2N/A look at accept() error handling first: gregames
2N/A * Make some modifications to the scoreboard creation routines,
2N/A prefer anonymous shared memory, and allow a configuration
2N/A directive to override the defaults.
2N/A Message-ID: <20020130080804.C16977@clove.org>
2N/A Status: Aaron volunteers
2N/A * Get perchild to work on platforms other than Linux. This
2N/A descriptors between vhost child groups. An API was proposed
2N/A Message-ID: <20020111115006.K1529@clove.org>
2N/A Vote: Is a non-portable perchild going to hold up a GA release?
2N/A No: Aaron, Justin, trawick, stoddard, Jim, Ian, BrianP, gregames
2N/A * Recent changes to ap_rgetline may have broken EBCDIC boxes.
2N/A Message-ID: <20020122072605.GF28051@ebuilt.com>
2N/A Justin says: "I don't have an EBCDIC box to test on. A potential
2N/A solution is to split out ap_rgetline into two
2N/A functions as described in this message."
2N/A gregames says: I see the breakage now, and volunteer to fix it
2N/A when things calm down a little. It looks OK when
2N/A there are complete lines and no mime continuations.
2N/A * Modify the worker MPM so that it doesn't need to create and
2N/A destroy a pool for each request--possibly by adopting a
2N/A ptrans pool (like the prefork MPM) and the workers take
2N/A reduce context-switching
2N/A * CGI single-byte reads
2N/A BrianP suggests that this is caused by the ap_scan_script_header_err()
2N/A routine, which will do single-byte reads until it finds the end
2N/A of the header, at which point it constructs a pipe-bucket (buffered)
2N/A Proposed solution in:
2N/A Message-ID: <3C36ADAF.60601@cnet.com>
2N/A * Try to get libtool inter-library dependency code working on AIX.
2N/A Message-ID: <cm3n10lx555.fsf@rdu163-40-092.nc.rr.com>
2N/A Justin says: If we get it working on AIX, we can enable this
2N/A on all platforms and clean up our build system
2N/A Jeff says: I thought I tested a patch for you sometime in
2N/A January that you were going to commit within a few
2N/A * Handling of %2f in URIs. Currently both 1.3 and 2.0
2N/A completely disallow %2f in the request URI path (see
2N/A ap_unescape_url() in
util.c). It's permitted and passed
2N/A through in the query string, however. Roy says the
2N/A original reason for disallowing it, from five years ago,
2N/A was to protect CGI scripts that applied PATH_INFO to
2N/A a filesystem location and which might be tricked by
2N/A ..%2f..%2f(...). We *should* allow path-info of the
2N/A Since we've revamped a lot of our processing of path
2N/A segments, it would be nice to allow this, or at least
2N/A allow it conditionally with a directive.
2N/A * FreeBSD, threads, and worker MPM. All seems to work fine
2N/A if you only have one worker process with many threads. Add
2N/A a second worker process and the accept lock seems to be
2N/A lost. This might be an APR issue with how it deals with
2N/A the child_init hook (
i.e. the fcntl lock needs to be resynced).
2N/A More examination and analysis is required.
2N/A Status: This has also been reported on Cygwin.
2N/A Message-ID: <3C2CC514.8EF3BED1@wapme-systems.de> (cygnus)
2N/A Justin says: So, FreeBSD-CURRENT and Cywin have the same
2N/A problem. Yum. If another platform has this
2N/A with worker, this becomes a showstopper.
2N/A Aaron says: I spent some time disecting this and have come to
2N/A the conclusion that it is not a problem in the worker MPM
2N/A (or at least, it is not isolated to a problem in worker).
2N/A I'll list some of the problems I'm seeing in case someone
2N/A else wants to pick up where I've left off:
2N/A - Delivery of just about any signal to one of the child
2N/A processes will send it into an infinite loop as well.
2N/A - Even though the parent is spinning out of control,
2N/A at first the child or children will appear to work
2N/A properly. At times it is possible to get it into a state,
2N/A however, where a request will hang until another concurrent
2N/A request "kicks" the first, at which point the second will
2N/A hang. My theory is that this has to do with the
2N/A pthread_cond_*() implementation in FreeBSD, but it's still
2N/A possible that it is in APR.
2N/A Justin adds: Oh, FreeBSD threads are implemented entirely with
2N/A select()/poll()/longjmp(). Welcome to the nightmare.
2N/A So, that means a ktrace output also has the thread
2N/A scheduling internals in it (since it is all the same to
2N/A the kernel). Which makes it hard to distinguish between
2N/A our select() calls and their select() calls.
2N/A *bangs head on wall repeatedly* But, some of the libc_r
2N/A files have a DBG_MSG #define. This is moderately helpful
2N/A when used with -DNO_DETACH. The kernel scheduler isn't
2N/A waking up the threads on a select(). Yum. And, I bet
2N/A those decrementing select calls have to do with the
2N/A scheduler. Time to brush up on our OS fundamentals.
2N/A * There is increasing demand from module writers for an API
2N/A that will allow them to control the server � la apachectl.
2N/A Reasons include sole-function servers that need to die if
2N/A an external dependency (
e.g., a database) fails, et cetera.
2N/A Perhaps something in the (ever more abused) scoreboard?
2N/A rbb: I don't believe the scoreboard is the correct mechanism
2N/A for this. We already have a pipe that goes between parent
2N/A and child for graceful shutdown events, along with an API that
2N/A can be used to send a message down that pipe. In threaded MPMs,
2N/A it is easy enough to make that one pipe be used for graceful
2N/A and graceless events, and it is also easy to open that pipe
2N/A to both parent and child for writing. Then we just need to figure
2N/A out how to do graceless on non-threaded MPMs.
2N/A * revamp the input filter behavior, per discussions since
2N/A February (and especially at the hackathon last
2N/A April). Specifically, ap_get_brigade will return a brigade with
2N/A *up to* a specific number of bytes, or a "line" of data. The
2N/A read may be blocking or nonblocking. ap_getline() will be
2N/A refactored into apr_brigade_getline(), and then DECHUNK can use
2N/A f->next (ap_getline will always read "top of input stack"). Also
2N/A fix the bug where request body content will end up closing the
2N/A connection (buggering up persistent conns).
2N/A Status: Justin is working on this as fast as he can.
2N/A The core input filters, HTTP-related filters, mod_ssl,
2N/A mod_proxy, and ap_[r]getline are switched to the new logic.
2N/A - socket bucket and core input filter changes. see end of
2N/A message ID (Feb 27): <20010227075326.S2297@lyra.org>
2N/A - fix up ap_get_brigade() semantics, fix bug in DECHUNK /
2N/A ap_getline. many messages (plus their threads) (
Apr/May):
2N/A Message-ID: <20010402101207.J27539@lyra.org>
2N/A Message-ID: <3AF7F921.D2EEC41A@algroup.co.uk>
2N/A Message-ID: <20010508190029.E18404@lyra.org>
2N/A Message-ID: <20010509115445.D1374@lyra.org>
2N/A - thoughts on filter modes:
2N/A Message-ID: <021b01c14dee$09782af0$93c0b0d0@roweclan.net>
2N/A - proposal for getline rewrite:
2N/A Message-ID: <20011231104019.GD3904@ebuilt.com>
2N/A - patch for getline rewrite:
2N/A Message-ID: <20020122072605.GF28051@ebuilt.com>
2N/A - Change ap_get_brigade prototype:
2N/A Message-ID: <20020120115136.GI17601@ebuilt.com>
2N/A * Allow the DocumentRoot directive within <Location > scopes? This
2N/A and in-your-face.) DocumentRoot unset would be accepted [and would
2N/A not permit content to be served, only virtual resources such as
2N/A server-info or server-status.
2N/A This proposed change would _not_ depricate Alias.
2N/A * Win32: Rotatelogs sometimes is not terminated when Apache
2N/A goes down hard. FirstBill was looking at possibly tracking the
2N/A child's-child processes in the parent process.
2N/A OtherBill asks, wasn't this fixed?
2N/A stoddard: Not fixed. Shared scoreboard might offer a good
2N/A way for the parent to keep track of 'other child' processes
2N/A and whack them if the child goes down.
2N/A * Win32: Add a simple hold console open patch (wait for close or
2N/A the ESC key, with a nice message) if the server died a bad
2N/A death (non-zero exit code) in console mode.
2N/A Resolution: bring forward same ugly hacks from 1.3.13-.20
2N/A * Port of mod_ssl to Apache 2.0:
2N/A remaining work includes:
2N/A (2) Enabling the various SSL caching mechanisms (shmcb, shmht)
2N/A (3) Enabling SSL extentions
2N/A (4) Trying to seperate the https filter logic from mod_ssl -
2N/A This is to facilitate other modules that wish to use the https
2N/A filter or the mod_ssl logic or both as required.
2N/A Justin: mod_ssl filter logic is redone, so that should be fine.
2N/A Madhu has submitted a patch for SSL caching - however, I
2N/A am -0 on that patch as I *think* we could implement the
2N/A shared memory another way that is much cleaner (
i.e. 2N/A treat shmem directly as a dbm via APR routines). Justin
2N/A also thinks that the https filter logic may be sufficiently
2N/A decoupled now, but isn't really sure.
2N/A * Performance & Debug: Eliminate most (and perhaps all) of the
2N/A light weight memory management functions that allow freeing
2N/A memory (putting it back into a memory pool) when it is no
2N/A longer needed. Enabling simple debugging features like guard
2N/A bands, double free detection, etc. would be cool but certainly
2N/A not a hard requirement.
2N/A Status: Cliff started to implement this using SMS as has
2N/A been discussed at length for months, but since
2N/A SMS is not being used anywhere else in the server,
2N/A several people expressed the opinion that we should
2N/A get rid of it entirely, meaning that the buckets
2N/A need their own memory management (free list) functions.
2N/A Cliff will implement that this weekend so we at least
2N/A * Eliminate unnecessary creation of pipes in mod_cgid
2N/A * the autoconf setup should be fixed to default to using the
2N/A in a layout should be overridable on the command line. Plus,
2N/A what we do right now just doesn't seem to fully fit into how autoconf
2N/A works, eg. AC_PREFIX_DEFAULT issues.
2N/A Message-ID: <Pine.BSF.4.20.0104031557420.20876-100000@alive.znep.com>
2N/A * Document mod_file_cache.
2N/A * Platforms that do not support fork (primarily Win32 and AS/400)
2N/A Architect start-up code that avoids initializing all the modules
2N/A in the parent process on platforms that do not support fork.
2N/A would eliminate some code in the Win32 branch that essentially
2N/A duplicates what is in APR.
2N/A * There are still a number of places in the code where we are
2N/A losing error status (
i.e. throwing away the error returned by a
2N/A system call and replacing it with a generic error code)
2N/A * Mass vhosting version of suEXEC.
2N/A the dbmmanage employs the first-matched dbm format. This is not
2N/A necessarily the library that Apache was built with. Aught to
2N/A rewrite dbmmanage upon installation to bin/ with the proper library
2N/A for predictable mod_auth_dbm administration.
2N/A Questions; htdbm exists, time to kill dbmmanage, or does it remain
2N/A useful as a perl dbm management example? If we keep it,
2N/A do we address the issue above?
2N/A * use apu_dbm in mod_auth_dbm
2N/A Status: Greg +1 (low-priority volunteer)
2N/A Justin says: "Seems like this is already there, so should we just
2N/A remove the other DBM code in that file? If you want
2N/A to use gdbm, or dbm, etc, you should tell apr-util."
2N/A Will says: "bs - I may choose the fastest - most efficient native
2N/A dbm implementation, for shared proc caches, ssl session
2N/A caching, etc, but that has nothing to do with maintaining
2N/A a userlist via dbm, which has to remain readable between
2N/A for apr-util would let us do this with just apr, though."
2N/A Ian says: "multi-dbm is in, but it still has ndbm support hardcoded
2N/A is this still required? isn't ndbm supported via gdbm?"
2N/A * Integrate mod_dav.
2N/A Some additional items remaining:
2N/A - case_preserved_filename stuff
2N/A (use the new canonical name stuff?)
2N/A - find a new home for ap_text(_header)
2N/A - is it possible to remove the DAV: namespace stuff from util_xml?
2N/A * ap_core_translate() and its use by mod_mmap_static and mod_file_cache
2N/A are a bit wonky. The function should probably be exposed as a utility
2N/A function (such as ap_translate_url2fs() or ap_validate_fs_url() or
2N/A something). Another approach would be a new hook phase after
2N/A "translate" which would allow the module to munge what the
2N/A translation has decided to do.
2N/A Status: Greg +1 (volunteers), Ryan +1
2N/A calls ap_fixup_virutal_hosts(), ap_fini_vhost_config(), and
2N/A ap_sort_hooks() [to reduce the logic in main()]
2N/A * read the config tree just once, and process N times (as necessary)
2N/A * (possibly) use UUIDs in mod_unique_id
and/or mod_usertrack
2N/A * (possibly) port the bug fix for PR 6942 (segv when LoadModule is put
2N/A into a VirtualHost container) to 2.0.
2N/A * callers of ap_run_create_request() should check the return value
2N/A for failure (Doug volunteers)
2N/A * Win32: Get Apache working on Windows 95/98. The following work
2N/A (at least) needs to be done:
2N/A - winnt MPM: Fix 95/98 code paths in the winnt MPM. There is some NT
2N/A specific code that is still not in NT only code paths
2N/A - IOL binds to APR sendfile, implemented with TransmitFile, which
2N/A is not available on 95/98.
2N/A - Document warning that OSR2 is required (for Crypt functions, in
2N/A rand.c, at least.) This could be resolved with an SSL library, or
2N/A randomization in APR itself.
2N/A actually works) and add in a splash of Win9x service code.
2N/A * In order to use a DSO version of mod_ssl we have to link with
2N/A -lssl and -lcrypto. A workaround is in place right now where the
2N/A entire EXTRA_LIBS macro is being appended to the objects list, but
2N/A this is a hack. We should either revamp the APACHE_CHECK_SSL_TOOLKIT
2N/A autoconf function or come up with some other autoconf checks to
2N/A search for libssl and libcrypto and properly add them to mod_ssl's
2N/A * Make the worker MPM the default MPM for threaded Unix boxes.
2N/A +1: Justin, Jeff, Ian
2N/A -0: Aaron (premature decision, needs more discussion), Lars
2N/A -0: Cliff (I think the default config should be the safest possible)
2N/A * Fix the worker MPM to use POD to kill child processes instead
2N/A of ap_os_killpg, regardless of how they should die. (Ryan Bloom)
2N/APRs that have been suspended forever waiting for someone to
2N/Aput them into 'the next release':
2N/A missing call to "setlocale();"
2N/A * PR#78: mod_include
2N/A Additional status for XBitHack directive
2N/A Mod_proxy doesn't allow change of error pages
2N/A Modified PATH environemnt variable is not passed, instead
2N/A Proxy doesn't deliver documents if not connected
2N/A proxy converts ~name to %7Ename when name starts with a dot (.)
2N/A * PR#537: mod_access
2N/A mod_access syntax allows hosts that should be restricted
2N/A * PR#557: mod_auth-any
2N/A ~UserHome directories are not honored in absolute pathname
2N/A requests (.htaccess)
2N/A Proxy FTP Authentication Fails
2N/A * PR#623: mod_include
2N/A A smarter "Last Modified" value for SSI documents (see PR number 600)
2N/A Request of "Options SymLinksIfGroupMatch"
2N/A Proxy doesn't do links right for OpenVMS files through ftp:
2N/A imap should read <MAP><AREA>*</MAP> too!
2N/A RLimitCPU and RLimitMEM don't apply to all children like they should
2N/A Uses cwd before filling it in, doesn't use syslog
2N/A it is useful to allow specifiction that root-owned symlinks
2N/A should always be followed
2N/A Controlling Access to Remote Proxies would be nice...
2N/A Adding authentication "on the fly" through the proxy module
2N/A * PR#1004: apache-api
2N/A request_config field in request_rec is moderately bogus
2N/A DoS attacks involving memory consumption
2N/A * PR#1050: mod_log-any
2N/A Logging of virtual server to error_log as well
2N/A * PR#1085: mod_proxy
2N/A ProxyRemote make a dead cycle.
2N/A * PR#1117: mod_auth-any
2N/A suexec does not parse arguments to #exec cmd
2N/A * PR#1145: mod_include
2N/A Allow for Last-Modified: without resorting to XBitHack
2N/A * PR#1158: apache-api
2N/A improvements to child spawning API
2N/A * PR#1166: mod_proxy
2N/A ``nph-'' not honored (no buffering) for ProxyRemote mapping
2N/A Apache cannot handle continuation line in headers
2N/A setlogin() is not called, causing problems with
e.g. identd
2N/A regerror() exists, use it
2N/A * PR#1233: apache-api
2N/A there is no way to keep per-connection per-module state
2N/A * PR#1263: mod_autoexec
2N/A Add frame-safe anchor attribute to mod_autoindex links
2N/A CGI scripts running as Apache user: security (suexec etc.)
2N/A * PR#1287: mod_access
2N/A * PR#1290: mod_proxy
2N/A Need to know "hit-rate" on proxy cache
2N/A * PR#1358: mod_log-any
2N/A Selective url-encode of log fields (or maybe a pseudo
2N/A log_rewrite module?)
2N/A * PR#1383: mod_headers
2N/A I make mod_headers to modify request headers as well as
2N/A * PR#1532: mod_proxy
2N/A Proxy transfer logging
2N/A * PR#1547: mod_proxy
2N/A No HTTP_X_FORWARDED_FOR set...
2N/A * PR#1567: mod_proxy
2N/A ProxyRemote proxy requests fail authentication by firewall
2N/A * PR#1582: mod_rewrite
2N/A mod_rewrite forms REQUEST_URI different than mod_cgi does
2N/A * PR#1677: mod_headers
2N/A mod_headers should allow mod_log_config-style formats in
2N/A * PR#1702: mod_proxy
2N/A mod_proxy to support persistent conns?
2N/A * PR#1803: mod_include
2N/A patches to mod_include to allow for file tests
2N/A * PR#1809: mod_auth-any
2N/A Suggestion for improving authentication modules and core source
2N/A code, problem with 401 and ErrorDocument
2N/A * PR#1878: mod_proxy
2N/A listing of proxy cache content
2N/A Allow modules to set user:group for execution.
2N/A * PR#2024: apache-api
2N/A adding auth_why to conn_rec
2N/A * PR#2073: mod_log-any
2N/A pipelined connections are not logged correctly
2N/A * PR#2074: mod_rewrite
2N/A mod_rewrite doesn't pass Proxy Throughput on internal subrequests
2N/A HTTP Server Rebuild Line Needs Changing for the better
2N/A * PR#2138: mod_status
2N/A mod_status always displays 256 possible connection slots
2N/A * PR#2221: documentation
2N/A Make online documentation search link back to my installation
2N/A Can not POST to ErrorDocument - Apache/1.3b6
2N/A * PR#2314: mod_proxy
2N/A patterns in ProxyRemote
2N/A * PR#2343: mod_status
2N/A Status module averages are for entire uptime
2N/A suexec for general access of user content?
2N/A Proposal for TimeZone directive
2N/A /server-info doesn't check for the virtual host to list the info
2N/A problem specifying ndbm library for build ?with autoconfigure
2N/A AllowOverride FileInfo is too coarse
2N/A TimeOut applies to output of CGI scripts
2N/A * PR#2512: mod_access
2N/A <IfDenied> directive wanted
2N/A CGI's for general use still have to be run as another user
2N/A Cache file names in Proxy module
2N/A in global and <Virtual>.
2N/A mailto tags and bundling bug report script
2N/A Support for System Resource Controller
2N/A When will Apache support P3P? Any Plans?
2N/A Propose that Apache recommend $UNIQUE_ID for all "session id"
2N/A suggestion: power up your Include directive :)
2N/A cannot limit some HTTP methods
2N/A * PR#3143: apache-api
2N/A No module specific data hook for per-connection data
2N/A * PR#3191: mod_negotiation
2N/A no way to set global quality-of-source (qs) coneg values
2N/A * PR#3568: mod_proxy
2N/A Accessing URL through proxy server corrupts data.
2N/A * PR#3605: mod_proxy
2N/A Some anonymous FTP URLs ask for authentication
2N/A New ErrorDocumentMatch directive
2N/A Need to be able to override shebang line to make CGI scripts
2N/A "Files" and "FilesMatch" regexp does not recognize bang as
2N/A * PR#4448: mod_log-any
2N/A Please allow CGI env variables (QUERY_STRING, ...) to be logged
2N/A * PR#4459: mod_include
2N/A Suggestion for better handling of Last-modified headers
2N/A mod_cgi prevents handling of OPTIONS requests
2N/A * PR#5713: os-windows
2N/A [PATCH] install as win32 service with domain account
2N/A Status: Cannot accept password-as-arg, we should prompt the
2N/A user when -k install/-k config with a user argument.
2N/A AllowOverride should have a 'CheckNone' and 'AllowNone' argument
2N/A instead of only 'None'
2N/AOther bugs that need fixing:
2N/A * MaxRequestsPerChild measures connections, not requests.
2N/A Until someone has a better way, we'll probably just rename it
2N/A "MaxConnectionsPerChild".
2N/A * Regex containers don't work in an intutive way
2N/A Status: No one has come up with an efficient way to fix this
2N/A behavior. Dean has suggested getting rid of regex containers
2N/A OtherBill suggests: We at least seem to agree on eliminating
2N/A the <Container ~ foo> forms, and using only
2N/A <ContainerMatch foo> semantics.
2N/A * SIGSEGV on Linux (glibc 2.1.2) isn't caught properly by a
2N/A sigwaiting thread. We need to work around this, perhaps unless
2N/A there is hope soon for a fixed glibc.
2N/A needed. Apache 1.3 just never stashed "multipart" into
2N/A r->content_type. We should probably follow suit since the
2N/A byterange stuff doesn't want the rest of the code to see the
2N/A multipart content-type; the other code should still think it is
2N/A dealing with the <orig_ct> stuff.
2N/A Status: Greg volunteers to investigate (esp. since he was most
2N/A likely the one to break it :-)
2N/AOther features that need writing:
2N/A * Finish infrastructure in core for async MPMs
2N/A * TODO in source -- just do an egrep on "TODO" and see what's there
2N/A * Jon Travis's <jtravis@covalent.net> patch to deal with thread-safe
2N/A issues with inet_ntoa. See message <20001201163220.A12827@covalent.net>
2N/A Status: This is being set aside until the IPv6 work is finished
2N/A so that we know exactly what is required.
2N/A * Martin Sojka <msojka@gmx.de>'s patch to add error reporting for failed
2N/A htpasswd actions due to a full /tmp volume (other programs may have
2N/A * Mike Abbott's <mja@trudge.engr.sgi.com> patches to improve
2N/A Status: These were written for 1.3, and are awaiting a port to
2N/A * Jim Winstead's <jimw@trainedmonkey.com> patch to add CookieDomain and
2N/A other small mod_usertrack features
2N/A * Dan Rench's <drench@xnet.com> patch to add allow the errmsg and timefmt
2N/A of SSI's to be modified in the config file. Patch is available in
2N/A * Which MPMs will be included with Apache 2.0?