STATUS revision e302f38fd646764ce1a1e1c578d794aef514a9e5
842ae4bd224140319ae7feec1872b93dfd491143fieldingAPACHE 2.3 STATUS: -*-text-*-
842ae4bd224140319ae7feec1872b93dfd491143fieldingLast modified at [$Date$]
842ae4bd224140319ae7feec1872b93dfd491143fieldingThe current version of this file can be found at:
842ae4bd224140319ae7feec1872b93dfd491143fielding * http://svn.apache.org/repos/asf/httpd/httpd/trunk/STATUS
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholesDocumentation status is maintained seperately and can be found at:
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes * docs/STATUS in this source tree, or
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes * http://svn.apache.org/repos/asf/httpd/httpd/trunk/docs/STATUS
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholesConsult the following STATUS files for information on related projects:
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes * http://svn.apache.org/repos/asf/apr/apr/trunk/STATUS
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes * http://svn.apache.org/repos/asf/apr/apr-util/trunk/STATUS
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholesPatches considered for backport are noted in their branches' STATUS:
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes * http://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x/STATUS
e8f95a682820a599fe41b22977010636be5c2717jim * http://svn.apache.org/repos/asf/httpd/httpd/branches/2.0.x/STATUS
1747d30b98aa1bdbc43994c02cd46ab4cb9319e4fielding * http://svn.apache.org/repos/asf/httpd/httpd/branches/2.2.x/STATUS
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholesRelease history:
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes [NOTE that x.{odd}.z versions are strictly Alpha/Beta releases,
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes while x.{even}.z versions are Stable/GA releases.]
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes 2.3.9 : In Development.
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes 2.3.8 : Tagged on August 24, 2010.
11f2c481e1d57bedb3f758565307501e9a2730ddtrawick 2.3.7 : Tagged on August 19, 2010, not released.
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes 2.3.6 : Released on June 21, 2010.
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes 2.3.5 : Released on January 26, 2010.
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes 2.3.4 : Released on December 8, 2009.
5c0419d51818eb02045cf923a9fe456127a44c60wrowe 2.3.3 : Tagged on November 11, 2009, not released.
5c0419d51818eb02045cf923a9fe456127a44c60wrowe 2.3.2 : Tagged on March 23, 2009, not released.
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes 2.3.1 : Tagged on January 2, 2009, not released.
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes 2.3.0 : Tagged on December 6, 2008, not released.
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholesContributors looking for a mission:
ab86c68ce36c715e93f403dde41d0b9c1522c8b0sf * Just do an egrep on "TODO" or "XXX" in the source.
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes * Review the bug database at: http://issues.apache.org/bugzilla/
ab86c68ce36c715e93f403dde41d0b9c1522c8b0sf * Review the "PatchAvailable" bugs in the bug database:
ab86c68ce36c715e93f403dde41d0b9c1522c8b0sf https://issues.apache.org/bugzilla/buglist.cgi?bug_status=NEW&bug_status=ASSIGNED&bug_status=REOPENED&product=Apache+httpd-2&keywords=PatchAvailable
ab86c68ce36c715e93f403dde41d0b9c1522c8b0sf After testing, you can append a comment saying "Reviewed and tested".
ab86c68ce36c715e93f403dde41d0b9c1522c8b0sf * Open bugs in the bug database.
0f60998368b493f90120180a93fc2e1e74490872covenerCURRENT RELEASE NOTES:
0f60998368b493f90120180a93fc2e1e74490872covenerRELEASE SHOWSTOPPERS:
5bfaaf573bacb45c1cf290ce85ecc676587e8a64jim * Modules that are not ready for production use should be marked as
60284a9f9158baa60cc8ab4a69066404b1dcae7acovener experimental. Candidates:
87587593f1a53030e840acc0dec6cc881022ea40covener - MPM simple
a81c0c1ae464b2063a21b45f80c9da8d89bb840ecovener * Modules without documentation need to be moved to experimental or be
97cd2f98ad4abe68aaaba96b5bfc9ebf7109a2c1covener * Not all MPMs are updated to set conn_rec::current_thread correctly.
97cd2f98ad4abe68aaaba96b5bfc9ebf7109a2c1covener (Prefork, Worker, Event, Simple are updated).
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes jim sez: Then we just ship with those... mark any others as
43997561b2302d13dee973998e77743a3ddd2374trawick experimental
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes FOR NEXT ALPHA:
0568280364eb026393be492ebc732795c4934643jorton * The mod_session* modules need to be checked that their hooks respect
0568280364eb026393be492ebc732795c4934643jorton the returning of int (HTTP status codes) and apr_status_t as appropriate,
0568280364eb026393be492ebc732795c4934643jorton and any anomolies fixed.
0568280364eb026393be492ebc732795c4934643jorton OLD ISSUES THAT WERE THOUGHT TO BE SHOWSTOPPERS FOR 2.2 BUT OBVIOUSLY WEREN'T:
0568280364eb026393be492ebc732795c4934643jorton * Handling of non-trailing / config by non-default handler is broken
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes http://marc.theaimsgroup.com/?l=apache-httpd-dev&m=105451701628081&w=2
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes jerenkrantz asks: Why should this block a release?
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes wsanchez agrees: this may be a change in behavior, but isn't
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes clearly wrong, and even if so, it doesn't seem like a
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes showstopper.
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes * the edge connection filter cannot be removed
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes http://marc.theaimsgroup.com/?l=apache-httpd-dev&m=105366252619530&w=2
796e4a7141265d8ed7036e4628161c6eafb2a789jorton jerenkrantz asks: Why should this block a release?
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes stas replies: because it requires a rewrite of the filters stack
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes implementation (you have suggested that) and once 2.2 is
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes released you can't do that anymore.
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes pgollucci: this affects mod_perl I'm pretty sure.
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholesCURRENT VOTES:
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes * If the parent process dies, should the remaining child processes
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes "gracefully" self-terminate. Or maybe we should make it a runtime
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes option, or have a concept of 2 parent processes (one being a
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes "hot spare").
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes See: Message-ID: <3C58232C.FE91F19F@Golux.Com>
8113dac419143273351446c3ad653f3fe5ba5cfdwrowe Self-destruct: Ken, Martin, Lars, sctemme (parent shouldn't die, ever)
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes Not self-destruct: BrianP, Ian, Cliff, BillS
95b6fe1346805e1731e6e97c15d569c73be22cf7minfrin Make it runtime configurable: Aaron, jim, Justin, wrowe, rederpj, nd, pgollucci
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes /* The below was a concept on *how* to handle the problem */
95b6fe1346805e1731e6e97c15d569c73be22cf7minfrin Have 2 parents: +1: jim
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes -1: Justin, wrowe, rederpj, nd, pgollucci
95b6fe1346805e1731e6e97c15d569c73be22cf7minfrin +0: Lars, Martin (while standing by, could it do
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes something useful?)
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes * Make the worker MPM the default MPM for threaded Unix boxes.
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes +1: Justin, Ian, Cliff, BillS, striker, wrowe, nd, pgollucci, sctemme
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes +0: BrianP, Aaron (mutex contention is looking better with the
95b6fe1346805e1731e6e97c15d569c73be22cf7minfrin latest code, let's continue tuning and testing), rederpj, jim
a1790fb35c4b352dab721370985c623a9f8f5062rpluem pquerna: Do we want to change this for *2.4*?
713a2b68bac4aeb1e9c48785006c0732451039depquerna wrowe: Replies "yes"
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes * Name the Server (version 2.4 or 3.0, depending on the final call)
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes Recent discussion indicates we should designate a (short name).
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes This is not yet a [Vote] - Your nominations please:
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes * Apache HTTP Server (httpd)
482f676c6c19b1c5bb5cca04dad11509c1da3a4cwrowe +1: sctemme (why mess with it?)
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholesRELEASE NON-SHOWSTOPPERS BUT WOULD BE REAL NICE TO WRAP THESE UP:
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes * Add mod_allow_method or some other (usable) functionality to replace
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes * Sort out modules selections for most/all/reallyall. Maybe rename
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes all -> most, reallyall -> all, and remove the old 'most'.
4e9c24785b525d2956e6e381015c0f2bd0a72f4bcovener * Add mod_define.
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes * Patches submitted to the bug database:
f43b67c5a9d29b572eac916f8335cedc80c908bebnicholes http://issues.apache.org/bugzilla/buglist.cgi?bug_status=NEW&bug_status=ASSIGNED&bug_status=REOPENED&product=Apache+httpd-2&keywords=PatchAvailable
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes * Filter stacks and subrequests, redirects and fast redirects.
8113dac419143273351446c3ad653f3fe5ba5cfdwrowe There's at least one PR that suffers from the current unclean behaviour
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes (which lets the server send garbage): PR 17629
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes nd says: Every subrequest should get its own filter stack with the
fa123db15501821e36e513afa78e839775ad2800covener subreq_core filter as bottom-most. That filter does two things:
fa123db15501821e36e513afa78e839775ad2800covener - swallow EOS buckets
fa123db15501821e36e513afa78e839775ad2800covener - redirect the data stream to the upper request's (rr->main)
fa123db15501821e36e513afa78e839775ad2800covener filter chain directly after the subrequest's starting
fa123db15501821e36e513afa78e839775ad2800covener Once we have a clean solution, we can try to optimize
307219eca940aa30b873bfd68a44484dd3d3fa88covener it, so that the server won't be slow down too much.
5bfaaf573bacb45c1cf290ce85ecc676587e8a64jim * RFC 2616 violations.
4e9c24785b525d2956e6e381015c0f2bd0a72f4bcovener Closed PRs: 15852, 15857, 15859, 15861, 15864, 15869, 15870, 16120,
a81c0c1ae464b2063a21b45f80c9da8d89bb840ecovener 16125, 16135, 16136, 16137, 16138, 16139, 16140, 16518,
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes Open PRs: 15865, 15866, 15868, 16126, 16133, 16142, 16521
4e9c24785b525d2956e6e381015c0f2bd0a72f4bcovener jerenkrantz says: need to decide how many we need to backport and/or
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes if these rise to showstopper status.
4e9c24785b525d2956e6e381015c0f2bd0a72f4bcovener wrowe suggests: it would be nice to see "MUST" v.s. "SHOULD" v.s. "MAY"
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes out of this list, without reviewing them individually.
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes * There is a bug in how we sort some hooks, at least the pre-config
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes hook. The first time we call the hooks, they are in the correct
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes order, but the second time, we don't sort them correctly. Currently,
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes the modules/http/config.m4 file has been renamed to
4e9c24785b525d2956e6e381015c0f2bd0a72f4bcovener modules/http/config2.m4 to work around this problem, it should moved
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes back when this is fixed.
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes OtherBill offers that this is a SERIOUS problem. We do not sort
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes correctly by the ordering arguments passed to the register hook
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes functions. This was proven when I reordered the open_logs hook
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes to attempt to open the error logs prior to the access logs. Possibly
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes the entire sorting code needs to be refactored.
307219eca940aa30b873bfd68a44484dd3d3fa88covener * pipes deadlock on all platforms with limited pipe buffers (e.g. both
307219eca940aa30b873bfd68a44484dd3d3fa88covener Linux and Win32, as opposed to only Win32 on 1.3). The right solution
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes is either GStein's proposal for a "CGI Brigade", or OtherBill's proposal
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes for "Poll Buckets" for "Polling Filter Chains". Or maybe both :-)
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes * All handlers should always send content down even if r->header_only
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes is set. If not, it means that the HEAD requests don't generate the
fa123db15501821e36e513afa78e839775ad2800covener same headers as a GET which is wrong.
ab86c68ce36c715e93f403dde41d0b9c1522c8b0sf * exec cmd and suexec arg-passing enhancements
fa123db15501821e36e513afa78e839775ad2800covener Status: Patches proposed
cceddc0b6c0fdaed0c73abda39975bb1d388243acovener Message-ID: <20020526041748.A29148@prodigy.Redbrick.DCU.IE>
f2be127030aa4190033084f0a6add531c9bc41desf (see the "proc.patch" and "suexec-shell.patch" links in this message)
4e9c24785b525d2956e6e381015c0f2bd0a72f4bcovener * The 2.0.36 worker MPM graceless shutdown changes work but are
4e9c24785b525d2956e6e381015c0f2bd0a72f4bcovener a bit clunky on some platforms; eg, on Linux, the loop to
4e9c24785b525d2956e6e381015c0f2bd0a72f4bcovener join each worker thread seems to hang, and the parent ends up
60215f303c7e1ce8b6d272acb5bfa5b3d99dfd34covener killing off the child with SIGKILL. But at least it shuts down.
4e9c24785b525d2956e6e381015c0f2bd0a72f4bcovener chrisd: Has this been fixed by the changes for PR 38737?
4e9c24785b525d2956e6e381015c0f2bd0a72f4bcovener * We do not properly substitute the prefix-variables in the configuration
4e9c24785b525d2956e6e381015c0f2bd0a72f4bcovener scripts or generated-configs. (i.e. if sysconfdir is etc,
4e9c24785b525d2956e6e381015c0f2bd0a72f4bcovener httpd-std.conf points to conf.)
4e9c24785b525d2956e6e381015c0f2bd0a72f4bcovener * If any request gets through ap_process_request_internal() and is
5bfaaf573bacb45c1cf290ce85ecc676587e8a64jim scheduled to be served by the core handler, without a flag that this
4e9c24785b525d2956e6e381015c0f2bd0a72f4bcovener r->filename was tested by dir/file_walk, we need to 500 at the very
6683642c1e0032eeeed5f99e8c14880692ef84c5sf end of the ap_process_request_internal() processing so sub_req-esters
4e9c24785b525d2956e6e381015c0f2bd0a72f4bcovener know this request cannot be run. This provides authors of older
4e9c24785b525d2956e6e381015c0f2bd0a72f4bcovener modules better compatibility, while still improving the security and
4e9c24785b525d2956e6e381015c0f2bd0a72f4bcovener robustness of 2.0.
6683642c1e0032eeeed5f99e8c14880692ef84c5sf Status: still need to decide where this goes, OtherBill comments...
4e9c24785b525d2956e6e381015c0f2bd0a72f4bcovener Message-ID: <065701c14526$495203b0$96c0b0d0@roweclan.net>
4e9c24785b525d2956e6e381015c0f2bd0a72f4bcovener [Deleted comments regarding the ap_run_handler phase, as irrelevant
4e9c24785b525d2956e6e381015c0f2bd0a72f4bcovener as BillS points out that "common case will be caught in
4e9c24785b525d2956e6e381015c0f2bd0a72f4bcovener default_handler already (with the r->finfo.filetype == 0 check)"
4e9c24785b525d2956e6e381015c0f2bd0a72f4bcovener and the issue is detecting this -before- we try to run the req.]
4e9c24785b525d2956e6e381015c0f2bd0a72f4bcovener gregames says: can this happen somehow without a broken module
4e9c24785b525d2956e6e381015c0f2bd0a72f4bcovener being involved? If not, why waste cycles trying to defend against
4e9c24785b525d2956e6e381015c0f2bd0a72f4bcovener potential broken modules? It seems futile.
4e9c24785b525d2956e6e381015c0f2bd0a72f4bcovener wrowe counters: no, it shouldn't happen unless the module is broken.
5bfaaf573bacb45c1cf290ce85ecc676587e8a64jim But the right answer is to fail the request up-front in dir/file
4e9c24785b525d2956e6e381015c0f2bd0a72f4bcovener walk if the path was entirely invalid; and we can't do that either
5bfaaf573bacb45c1cf290ce85ecc676587e8a64jim UNTIL 2.1 or we break modules that haven't hooked map_to_storage.
4e9c24785b525d2956e6e381015c0f2bd0a72f4bcovener * Can a static httpd be built reliably?
4e9c24785b525d2956e6e381015c0f2bd0a72f4bcovener Message-ID: <20020207142751.T31582@clove.org>
4e9c24785b525d2956e6e381015c0f2bd0a72f4bcovener * Usage of APR_BRIGADE_NORMALIZE in core_input_filter should be
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes removed if possible.
4e9c24785b525d2956e6e381015c0f2bd0a72f4bcovener <Pine.LNX.4.33.0201202232430.318-100000@deepthought.cs.virginia.edu>
4e9c24785b525d2956e6e381015c0f2bd0a72f4bcovener Jeff wonders if we still care about this. It is no longer an
4e9c24785b525d2956e6e381015c0f2bd0a72f4bcovener API issue but simply an extra trip through the brigade.
4e9c24785b525d2956e6e381015c0f2bd0a72f4bcovener * Try to get libtool inter-library dependency code working on AIX.
4e9c24785b525d2956e6e381015c0f2bd0a72f4bcovener Message-ID: <cm3n10lx555.fsf@rdu163-40-092.nc.rr.com>
4e9c24785b525d2956e6e381015c0f2bd0a72f4bcovener Justin says: If we get it working on AIX, we can enable this
4e9c24785b525d2956e6e381015c0f2bd0a72f4bcovener on all platforms and clean up our build system somewhat.
4e9c24785b525d2956e6e381015c0f2bd0a72f4bcovener Jeff says: I thought I tested a patch for you sometime in
4e9c24785b525d2956e6e381015c0f2bd0a72f4bcovener January that you were going to commit within a few days.
4e9c24785b525d2956e6e381015c0f2bd0a72f4bcovener * Handling of %2f in URIs. Currently both 1.3 and 2.0
4e9c24785b525d2956e6e381015c0f2bd0a72f4bcovener completely disallow %2f in the request URI path (see
5bfaaf573bacb45c1cf290ce85ecc676587e8a64jim ap_unescape_url() in util.c). It's permitted and passed
5bfaaf573bacb45c1cf290ce85ecc676587e8a64jim through in the query string, however. Roy says the
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes original reason for disallowing it, from five years ago,
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes was to protect CGI scripts that applied PATH_INFO to
60215f303c7e1ce8b6d272acb5bfa5b3d99dfd34covener a filesystem location and which might be tricked by
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes ..%2f..%2f(...). We *should* allow path-info of the
d330a801b1e5d63a4b8b4fd431542ad0903fd71bbnicholes form 'http://foo.com/index.cgi/path/to/path%2finfo'.
0e05808dc59a321566303084c84b9826a4353cefrederpj Since we've revamped a lot of our processing of path
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes segments, it would be nice to allow this, or at least
b08925593f214f621161742925dcf074a8047e0acovener allow it conditionally with a directive.
ab86c68ce36c715e93f403dde41d0b9c1522c8b0sf OtherBill adds that %2f as the SECOND character of a multibyte
465bb68501690d7a47bfd2a6129580047d76d8f1rederpj sequence causes the request to fail! This happens notably in
5bfaaf573bacb45c1cf290ce85ecc676587e8a64jim the ja-jis encoding.
e8f95a682820a599fe41b22977010636be5c2717jim * There is increasing demand from module writers for an API
3dfeb02cfb853d8717ca0cc259b59fea610173f5bnicholes that will allow them to control the server à la apachectl.
3dfeb02cfb853d8717ca0cc259b59fea610173f5bnicholes Reasons include sole-function servers that need to die if
55e2e59e5910072e51c870afc68b0907f41a28e0sf an external dependency (e.g., a database) fails, et cetera.
55e2e59e5910072e51c870afc68b0907f41a28e0sf Perhaps something in the (ever more abused) scoreboard?
55e2e59e5910072e51c870afc68b0907f41a28e0sf On the other hand, we already have a pipe that goes between parent
55e2e59e5910072e51c870afc68b0907f41a28e0sf and child for graceful shutdown events, along with an API that
55e2e59e5910072e51c870afc68b0907f41a28e0sf can be used to send a message down that pipe. In threaded MPMs,
55e2e59e5910072e51c870afc68b0907f41a28e0sf it is easy enough to make that one pipe be used for graceful
55e2e59e5910072e51c870afc68b0907f41a28e0sf and graceless events, and it is also easy to open that pipe
55e2e59e5910072e51c870afc68b0907f41a28e0sf to both parent and child for writing. Then we just need to
55e2e59e5910072e51c870afc68b0907f41a28e0sf figure out how to do graceless on non-threaded MPMs.
55e2e59e5910072e51c870afc68b0907f41a28e0sf * Allow the DocumentRoot directive within <Location > scopes? This
55e2e59e5910072e51c870afc68b0907f41a28e0sf allows the beloved (crusty) Alias /foo/ /somepath/foo/ followed
55e2e59e5910072e51c870afc68b0907f41a28e0sf by a <Directory /somepath/foo> to become simply
55e2e59e5910072e51c870afc68b0907f41a28e0sf <Location /foo/> DocumentRoot /somefile/foo (IMHO a bit more legible
55e2e59e5910072e51c870afc68b0907f41a28e0sf and in-your-face.) DocumentRoot unset would be accepted [and would
55e2e59e5910072e51c870afc68b0907f41a28e0sf not permit content to be served, only virtual resources such as
54d22ed1c429b903b029bbd62621f11a9e286137minfrin server-info or server-status.
d330a801b1e5d63a4b8b4fd431542ad0903fd71bbnicholes This proposed change would _not_ depricate Alias.
d330a801b1e5d63a4b8b4fd431542ad0903fd71bbnicholes striker: See the thread starting with Message-ID:
d330a801b1e5d63a4b8b4fd431542ad0903fd71bbnicholes JLEGKKNELMHCJPNMOKHOGEEJFBAA.striker@apache.org.
d330a801b1e5d63a4b8b4fd431542ad0903fd71bbnicholes * Win32: Rotatelogs sometimes is not terminated when Apache
d330a801b1e5d63a4b8b4fd431542ad0903fd71bbnicholes goes down hard. FirstBill was looking at possibly tracking the
ab86c68ce36c715e93f403dde41d0b9c1522c8b0sf child's-child processes in the parent process.
d330a801b1e5d63a4b8b4fd431542ad0903fd71bbnicholes stoddard: Shared scoreboard might offer a good way for the parent
ab86c68ce36c715e93f403dde41d0b9c1522c8b0sf to keep track of 'other child' processes and whack them if the child
55e2e59e5910072e51c870afc68b0907f41a28e0sf goes down.
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes Other thoughts on walking the process chain using the NT kernel
513b324e774c559b579896df131fd7c8471ed529rederpj have also been proposed on APR.
513b324e774c559b579896df131fd7c8471ed529rederpj * Eliminate unnecessary creation of pipes in mod_cgid
513b324e774c559b579896df131fd7c8471ed529rederpj * Combine log_child and piped_log_spawn. Clean up http_log.c.
513b324e774c559b579896df131fd7c8471ed529rederpj Common logging API.
513b324e774c559b579896df131fd7c8471ed529rederpj * Platforms that do not support fork (primarily Win32 and AS/400)
513b324e774c559b579896df131fd7c8471ed529rederpj Architect start-up code that avoids initializing all the modules
513b324e774c559b579896df131fd7c8471ed529rederpj in the parent process on platforms that do not support fork.
02fd88c85a9850109753b87612955ad372de1575sf * There are still a number of places in the code where we are
02fd88c85a9850109753b87612955ad372de1575sf losing error status (i.e. throwing away the error returned by a
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes system call and replacing it with a generic error code)
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes * Mass vhosting version of suEXEC.
d330a801b1e5d63a4b8b4fd431542ad0903fd71bbnicholes * All DBMs suffer from confusion in support/dbmmanage (perl script) since
d330a801b1e5d63a4b8b4fd431542ad0903fd71bbnicholes the dbmmanage employs the first-matched dbm format. This is not
d330a801b1e5d63a4b8b4fd431542ad0903fd71bbnicholes necessarily the library that Apache was built with. Aught to
d330a801b1e5d63a4b8b4fd431542ad0903fd71bbnicholes rewrite dbmmanage upon installation to bin/ with the proper library
d330a801b1e5d63a4b8b4fd431542ad0903fd71bbnicholes for predictable mod_auth_dbm administration.
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes Questions; htdbm exists, time to kill dbmmanage, or does it remain
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes useful as a perl dbm management example? If we keep it,
d330a801b1e5d63a4b8b4fd431542ad0903fd71bbnicholes do we address the issue above?
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes * Integrate mod_dav.
307219eca940aa30b873bfd68a44484dd3d3fa88covener Some additional items remaining:
307219eca940aa30b873bfd68a44484dd3d3fa88covener - case_preserved_filename stuff
707f6d077f73cc948deead8df5b40ea42c1eaa78covener (use the new canonical name stuff?)
707f6d077f73cc948deead8df5b40ea42c1eaa78covener - find a new home for ap_text(_header)
ab86c68ce36c715e93f403dde41d0b9c1522c8b0sf - is it possible to remove the DAV: namespace stuff from util_xml?
185aa71728867671e105178b4c66fbc22b65ae26sf * ap_core_translate() and its use by mod_mmap_static and mod_file_cache
707f6d077f73cc948deead8df5b40ea42c1eaa78covener are a bit wonky. The function should probably be exposed as a utility
707f6d077f73cc948deead8df5b40ea42c1eaa78covener function (such as ap_translate_url2fs() or ap_validate_fs_url() or
707f6d077f73cc948deead8df5b40ea42c1eaa78covener something). Another approach would be a new hook phase after
707f6d077f73cc948deead8df5b40ea42c1eaa78covener "translate" which would allow the module to munge what the
707f6d077f73cc948deead8df5b40ea42c1eaa78covener translation has decided to do.
9ad7b260be233be7d7b5576979825cac72e15498rederpj Status: Greg +1 (volunteers)
d330a801b1e5d63a4b8b4fd431542ad0903fd71bbnicholes * Explore use of a post-config hook for the code in http_main.c which
d330a801b1e5d63a4b8b4fd431542ad0903fd71bbnicholes calls ap_fixup_virutal_hosts(), ap_fini_vhost_config(), and
54d22ed1c429b903b029bbd62621f11a9e286137minfrin ap_sort_hooks() [to reduce the logic in main()]
55e2e59e5910072e51c870afc68b0907f41a28e0sf * read the config tree just once, and process N times (as necessary)
d330a801b1e5d63a4b8b4fd431542ad0903fd71bbnicholes * (possibly) use UUIDs in mod_unique_id and/or mod_usertrack
d330a801b1e5d63a4b8b4fd431542ad0903fd71bbnicholes * (possibly) port the bug fix for PR 6942 (segv when LoadModule is put
d330a801b1e5d63a4b8b4fd431542ad0903fd71bbnicholes into a VirtualHost container) to 2.0.
54d22ed1c429b903b029bbd62621f11a9e286137minfrin * shift stuff to mod_core.h
54d22ed1c429b903b029bbd62621f11a9e286137minfrin * callers of ap_run_create_request() should check the return value
d330a801b1e5d63a4b8b4fd431542ad0903fd71bbnicholes for failure (Doug volunteers)
55e2e59e5910072e51c870afc68b0907f41a28e0sf * Fix the worker MPM to use POD to kill child processes instead
55e2e59e5910072e51c870afc68b0907f41a28e0sf of ap_os_killpg, regardless of how they should die.
55e2e59e5910072e51c870afc68b0907f41a28e0sf chrisd: Is this done, by any chance? See r92598 and r93358.
ab86c68ce36c715e93f403dde41d0b9c1522c8b0sf * Scoreboard structures could be changed in the future such that
ab86c68ce36c715e93f403dde41d0b9c1522c8b0sf proper alignment is not maintained, leading to segfaults on
d330a801b1e5d63a4b8b4fd431542ad0903fd71bbnicholes some systems. Cliff posted a patch to deal with this issue but
d330a801b1e5d63a4b8b4fd431542ad0903fd71bbnicholes later recanted. See this message to dev@apr.apache.org:
d330a801b1e5d63a4b8b4fd431542ad0903fd71bbnicholes <Pine.LNX.4.44.0203011354090.16457-200000@deepthought.cs.virginia.edu>
d330a801b1e5d63a4b8b4fd431542ad0903fd71bbnicholes * APXS either needs to be fixed completely for use when apr is out of tree,
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes or it should drop query mode altogether, and we just grow an
d330a801b1e5d63a4b8b4fd431542ad0903fd71bbnicholes httpd-config or similar arrangement.
b08925593f214f621161742925dcf074a8047e0acovener To quote a discussion in STATUS earlier:
d5b12fe8ae917e654a33247fd4e59dc9e75170aebnicholes thommay: this doesn't fix all the problems with apxs and out of
60284a9f9158baa60cc8ab4a69066404b1dcae7acovener tree apr/apr-util, but it's a good start. There's still the
707f6d077f73cc948deead8df5b40ea42c1eaa78covener query cases; but I'm beginning to think that in these cases
5f3e4e06f8e23597d2f95e2c2cff1116c522488fcovener the app should be querying ap{r,u}-config directly
707f6d077f73cc948deead8df5b40ea42c1eaa78covener deprecate -q: add htpd-config: gstein, pquerna, minfrin, pgollucci
ab86c68ce36c715e93f403dde41d0b9c1522c8b0sfTODO ISSUES REMAINING IN MOD_SSL:
707f6d077f73cc948deead8df5b40ea42c1eaa78covener * Do we need SSL_set_read_ahead()?
9ad7b260be233be7d7b5576979825cac72e15498rederpj * the ssl_expr api is NOT THREAD SAFE. race conditions exist:
185aa71728867671e105178b4c66fbc22b65ae26sf -in ssl_expr_comp() if SSLRequire is used in .htaccess
707f6d077f73cc948deead8df5b40ea42c1eaa78covener (ssl_expr_info is global)
707f6d077f73cc948deead8df5b40ea42c1eaa78covener -is ssl_expr_eval() if there is an error
707f6d077f73cc948deead8df5b40ea42c1eaa78covener (ssl_expr_error is global)
707f6d077f73cc948deead8df5b40ea42c1eaa78covener * SSLRequire directive (parsing of) leaks memory
707f6d077f73cc948deead8df5b40ea42c1eaa78covener * Diffie-Hellman-Parameters for temporary keys are hardcoded in
707f6d077f73cc948deead8df5b40ea42c1eaa78covener ssl_engine_dh.c, while the comment in ssl_engine_kernel.c says:
60284a9f9158baa60cc8ab4a69066404b1dcae7acovener "it is suggested that keys be changed daily or every 500
707f6d077f73cc948deead8df5b40ea42c1eaa78covener transactions, and more often if possible."
707f6d077f73cc948deead8df5b40ea42c1eaa78covener * ssl_var_lookup could be rewritten to be MUCH faster
185aa71728867671e105178b4c66fbc22b65ae26sf * CRL callback should be pluggable
707f6d077f73cc948deead8df5b40ea42c1eaa78covener * session cache store should be pluggable
ab86c68ce36c715e93f403dde41d0b9c1522c8b0sf * init functions should return status code rather than ssl_die()
707f6d077f73cc948deead8df5b40ea42c1eaa78covener * ssl_engine_pphrase.c needs to be reworked so it is generic enough
707f6d077f73cc948deead8df5b40ea42c1eaa78covener to also decrypt proxy keys
707f6d077f73cc948deead8df5b40ea42c1eaa78covener * output warning when allowing SSL v2.0 ? its so old
707f6d077f73cc948deead8df5b40ea42c1eaa78covener * mod_proxy: Ability to run SSL over proxy gateway connections,
707f6d077f73cc948deead8df5b40ea42c1eaa78covener encrypting (or reencrypting) at the proxy.
707f6d077f73cc948deead8df5b40ea42c1eaa78covener * mod_cache: Handle ESI tags.
9ad7b260be233be7d7b5576979825cac72e15498rederpj * mod_cache: Resolve issue of how to cache page fragments (or perhaps
9ad7b260be233be7d7b5576979825cac72e15498rederpj -if- we want to cache page fragments). Today, mod_cache/mod_mem_cache
9ad7b260be233be7d7b5576979825cac72e15498rederpj will cache #include 'virtual' requests (but not #include 'file'
ab86c68ce36c715e93f403dde41d0b9c1522c8b0sf requests). This was accomplished by making CACHE_IN a
ab86c68ce36c715e93f403dde41d0b9c1522c8b0sf CONTENT_SET-1 filter to force it to run before the SUBREQ_CORE
ab86c68ce36c715e93f403dde41d0b9c1522c8b0sf filter. But now responses cannot be cached that include the
8445dae5cc606ba8ba04efc341cc1e081d95920drpluem effects of having been run through CONTENT_SET filters
ab86c68ce36c715e93f403dde41d0b9c1522c8b0sf (mod_deflate, mod_expires, etc). We could rerun all the
ab86c68ce36c715e93f403dde41d0b9c1522c8b0sf CONTENT_SET filters on the cached response, but this will not
ab86c68ce36c715e93f403dde41d0b9c1522c8b0sf work in all cases. For example, mod_expires relies on installing
ab86c68ce36c715e93f403dde41d0b9c1522c8b0sf the EXPIRATION filter during fixups. Contents served out of
ab86c68ce36c715e93f403dde41d0b9c1522c8b0sf mod_cache (out of the quick_handler) bypass -all- the request
ab86c68ce36c715e93f403dde41d0b9c1522c8b0sf line server hooks (Ryan really hated this. It is great for
ab86c68ce36c715e93f403dde41d0b9c1522c8b0sf performance, but bad because of the complications listed above).
ab86c68ce36c715e93f403dde41d0b9c1522c8b0sf * mod_mem_cache: Consider adding a RevalidateTimeout directive to
ab86c68ce36c715e93f403dde41d0b9c1522c8b0sf specify time at which local cached content is to be revalidated
ab86c68ce36c715e93f403dde41d0b9c1522c8b0sf (ie, underlying file stat'ed to see if it has changed).
ab86c68ce36c715e93f403dde41d0b9c1522c8b0sf * mod_cache: CacheEnable/CacheDisable should accept regular expressions.
ab86c68ce36c715e93f403dde41d0b9c1522c8b0sf jerenkrantz says: Too slow. Get regexs away from speedy caches by
ab86c68ce36c715e93f403dde41d0b9c1522c8b0sf default. Introduce a new CacheEnableRegex if you want.
ab86c68ce36c715e93f403dde41d0b9c1522c8b0sf * mod_mem_cache/mod_disk_cache: Need to be able to query cache
ab86c68ce36c715e93f403dde41d0b9c1522c8b0sf status (num of entries, cache object properties, etc.).
ab86c68ce36c715e93f403dde41d0b9c1522c8b0sf mod_status could be extended to query optional hooks defined
141e1368614dc7564e1627671361b01b4869b491bnicholes by modules for the purpose of reporting module status.
3dfeb02cfb853d8717ca0cc259b59fea610173f5bnicholes mod_cache (et. al.) could define optional hooks that are called
d330a801b1e5d63a4b8b4fd431542ad0903fd71bbnicholes to collect status. Status should be queryable by
4be9c459920a7c1cfe62d654327dae5c4bb6b284sf HTTP or SNMP?
d330a801b1e5d63a4b8b4fd431542ad0903fd71bbnicholes jerenkrantz says: Yawn. Who cares.
1ae7a5fbce5d4f65f3da355792258fe5dbc4ef55covener * MaxRequestsPerChild measures connections, not requests.
ab86c68ce36c715e93f403dde41d0b9c1522c8b0sf Until someone has a better way, we'll probably just rename it
4be9c459920a7c1cfe62d654327dae5c4bb6b284sf "MaxConnectionsPerChild".
185aa71728867671e105178b4c66fbc22b65ae26sf * Regex containers don't work in an intutive way
d330a801b1e5d63a4b8b4fd431542ad0903fd71bbnicholes Status: No one has come up with an efficient way to fix this
3dfeb02cfb853d8717ca0cc259b59fea610173f5bnicholes behavior. Dean has suggested getting rid of regex containers
3dfeb02cfb853d8717ca0cc259b59fea610173f5bnicholes OtherBill suggests: We at least seem to agree on eliminating
3dfeb02cfb853d8717ca0cc259b59fea610173f5bnicholes the <Container ~ foo> forms, and using only
4be9c459920a7c1cfe62d654327dae5c4bb6b284sf <ContainerMatch foo> semantics.
4be9c459920a7c1cfe62d654327dae5c4bb6b284sf * orig_ct in the byterange/multipart handling may not be
4be9c459920a7c1cfe62d654327dae5c4bb6b284sf needed. Apache 1.3 just never stashed "multipart" into
4be9c459920a7c1cfe62d654327dae5c4bb6b284sf r->content_type. We should probably follow suit since the
4be9c459920a7c1cfe62d654327dae5c4bb6b284sf byterange stuff doesn't want the rest of the code to see the
4be9c459920a7c1cfe62d654327dae5c4bb6b284sf multipart content-type; the other code should still think it is
4be9c459920a7c1cfe62d654327dae5c4bb6b284sf dealing with the <orig_ct> stuff.
4be9c459920a7c1cfe62d654327dae5c4bb6b284sf Status: Greg volunteers to investigate (esp. since he was most
4be9c459920a7c1cfe62d654327dae5c4bb6b284sf likely the one to break it :-)
4be9c459920a7c1cfe62d654327dae5c4bb6b284sfEXPERIMENTAL MODULES:
185aa71728867671e105178b4c66fbc22b65ae26sf Experimental modules should eventually be be promoted to fully supported
4be9c459920a7c1cfe62d654327dae5c4bb6b284sf status or removed from the repository entirely (ie, the
4be9c459920a7c1cfe62d654327dae5c4bb6b284sf 'experiment' failed). This section tracks what needs to happen to
4be9c459920a7c1cfe62d654327dae5c4bb6b284sf get the modules promoted to fully supported status.