STATUS revision 80464b326874ee15d74742ae39708ec3f2eae1d7
f743002678eb67b99bbc29fee116b65d9530fec0wroweAPACHE 2.3 STATUS: -*-text-*-
80833bb9a1bf25dcf19e814438a4b311d2e1f4cffuankgLast modified at [$Date$]
1337c7673efc1f80f634139fbad7cbb98a0dc657ylavicThe current version of this file can be found at:
1337c7673efc1f80f634139fbad7cbb98a0dc657ylavic * http://svn.apache.org/repos/asf/httpd/httpd/trunk/STATUS
4da61833a1cbbca94094f9653fd970582b97a72etrawickDocumentation status is maintained seperately and can be found at:
4da61833a1cbbca94094f9653fd970582b97a72etrawick * docs/STATUS in this source tree, or
4da61833a1cbbca94094f9653fd970582b97a72etrawick * http://svn.apache.org/repos/asf/httpd/httpd/trunk/docs/STATUS
4789804be088bcd86ae637a29cdb7fda25169521jailletcConsult the following STATUS files for information on related projects:
4789804be088bcd86ae637a29cdb7fda25169521jailletc * http://svn.apache.org/repos/asf/apr/apr/trunk/STATUS
4789804be088bcd86ae637a29cdb7fda25169521jailletc * http://svn.apache.org/repos/asf/apr/apr-util/trunk/STATUS
e50c3026198fd496f183cda4c32a202925476778covenerPatches considered for backport are noted in their branches' STATUS:
5b88c8507d5ef6d0c4cfbc78230294968175b638minfrin * http://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x/STATUS
5b88c8507d5ef6d0c4cfbc78230294968175b638minfrin * http://svn.apache.org/repos/asf/httpd/httpd/branches/2.0.x/STATUS
6c3b9cebb551140fbb25d58bae08b539b3802133ylavic * http://svn.apache.org/repos/asf/httpd/httpd/branches/2.2.x/STATUS
4f29b65ab4b547ad5dbe506e2d0ff5d12ead9247ylavicRelease history:
4f29b65ab4b547ad5dbe506e2d0ff5d12ead9247ylavic [NOTE that x.{odd}.z versions are strictly Alpha/Beta releases,
0a0df13b7f1f4f1a74fe295253d89ca3911b301aylavic while x.{even}.z versions are Stable/GA releases.]
0a0df13b7f1f4f1a74fe295253d89ca3911b301aylavic 2.3.0 : in development
69301145375a889e7e37caf7cc7321ac0f91801erpluemContributors looking for a mission:
506bfe33206b2fece40ef25f695af39dd4130facjkaluza * Just do an egrep on "TODO" or "XXX" in the source.
506bfe33206b2fece40ef25f695af39dd4130facjkaluza * Review the bug database at: http://issues.apache.org/bugzilla/
d58a848a016d401b965111e50ef829e1641f7834minfrin * Review the "PatchAvailable" bugs in the bug database:
d58a848a016d401b965111e50ef829e1641f7834minfrin http://issues.apache.org/bugzilla/buglist.cgi?bug_status=NEW&bug_status=ASSIGNED&bug_status=REOPENED&product=Apache+httpd-2.0&keywords=PatchAvailable
2e6f4d654c96c98b761fb012fd25c5d5b1558c44sf After testing, you can append a comment saying "Reviewed and tested".
17e6c95f3b22d18acdf8380fb26a8d0e10c80767ylavic * Open bugs in the bug database.
17e6c95f3b22d18acdf8380fb26a8d0e10c80767ylavicCURRENT RELEASE NOTES:
e8bd80a4bb88199d2f9a24a50345688e52d9c116ylavicRELEASE SHOWSTOPPERS:
330e16bea8fe9cace4de90c349750c03dfb1fe64ylavic * Handling of non-trailing / config by non-default handler is broken
330e16bea8fe9cace4de90c349750c03dfb1fe64ylavic http://marc.theaimsgroup.com/?l=apache-httpd-dev&m=105451701628081&w=2
330e16bea8fe9cace4de90c349750c03dfb1fe64ylavic jerenkrantz asks: Why should this block a release?
330e16bea8fe9cace4de90c349750c03dfb1fe64ylavic wsanchez agrees: this may be a change in behavior, but isn't
330e16bea8fe9cace4de90c349750c03dfb1fe64ylavic clearly wrong, and even if so, it doesn't seem like a
330e16bea8fe9cace4de90c349750c03dfb1fe64ylavic showstopper.
d7205b1a86c51c27b71a2c458dc453fd53a261c1covener * the edge connection filter cannot be removed
d7205b1a86c51c27b71a2c458dc453fd53a261c1covener http://marc.theaimsgroup.com/?l=apache-httpd-dev&m=105366252619530&w=2
d7205b1a86c51c27b71a2c458dc453fd53a261c1covener jerenkrantz asks: Why should this block a release?
44ff304057225e944e220e981d434a046d14cf06covener stas replies: because it requires a rewrite of the filters stack
44ff304057225e944e220e981d434a046d14cf06covener implementation (you have suggested that) and once 2.2 is
44ff304057225e944e220e981d434a046d14cf06covener released you can't do that anymore.
5d1ba75b8794925e67591c209085a49279791de9covenerCURRENT VOTES:
032982212dbcc7c3cce95bf89c503bb56e185ac7kbrand * If the parent process dies, should the remaining child processes
032982212dbcc7c3cce95bf89c503bb56e185ac7kbrand "gracefully" self-terminate. Or maybe we should make it a runtime
032982212dbcc7c3cce95bf89c503bb56e185ac7kbrand option, or have a concept of 2 parent processes (one being a
032982212dbcc7c3cce95bf89c503bb56e185ac7kbrand "hot spare").
caad2986f81ab263f7af41467dd622dc9add17f3ylavic See: Message-ID: <3C58232C.FE91F19F@Golux.Com>
caad2986f81ab263f7af41467dd622dc9add17f3ylavic Self-destruct: Ken, Martin, Lars
caad2986f81ab263f7af41467dd622dc9add17f3ylavic Not self-destruct: BrianP, Ian, Cliff, BillS
45a10d38e6051fd7bdf9d742aaae633d97ff02abjailletc Make it runtime configurable: Aaron, jim, Justin, wrowe, rederpj, nd
f7317ff316c2b141feea31bddb74d5d3fa1584edjorton /* The below was a concept on *how* to handle the problem */
2165214331e4afafca4048f66f303d0253d7b001covener Have 2 parents: +1: jim
a34684a59b60a4173c25035d0c627ef17e6dc215rpluem -1: Justin, wrowe, rederpj, nd
a34684a59b60a4173c25035d0c627ef17e6dc215rpluem +0: Lars, Martin (while standing by, could it do
1e2d421a36999d292042a5539971070d54aa6c63ylavic something useful?)
1e2d421a36999d292042a5539971070d54aa6c63ylavic * Make the worker MPM the default MPM for threaded Unix boxes.
fa7ed98b9dc94c5845cf845aea0a44ecacd290c9humbedooh +1: Justin, Ian, Cliff, BillS, striker, wrowe, nd
fa7ed98b9dc94c5845cf845aea0a44ecacd290c9humbedooh +0: BrianP, Aaron (mutex contention is looking better with the
fa7ed98b9dc94c5845cf845aea0a44ecacd290c9humbedooh latest code, let's continue tuning and testing), rederpj, jim
0b67eb8568cd58bb77082703951679b42cf098actrawick pquerna: Do we want to change this for 2.2?
fb1985a97912b25ec6564c73e610a31e5fc6e25fcovenerRELEASE NON-SHOWSTOPPERS BUT WOULD BE REAL NICE TO WRAP THESE UP:
6502b7b32f980cc2093bb3ebce37e5e4dc68fba4ylavic * Patches submitted to the bug database:
6502b7b32f980cc2093bb3ebce37e5e4dc68fba4ylavic http://issues.apache.org/bugzilla/buglist.cgi?bug_status=NEW&bug_status=ASSIGNED&bug_status=REOPENED&product=Apache+httpd-2.0&keywords=PatchAvailable
c1a63b8fad09c419c1a64f75993feb8a343a6801ylavic * Filter stacks and subrequests, redirects and fast redirects.
c1a63b8fad09c419c1a64f75993feb8a343a6801ylavic There's at least one PR that suffers from the current unclean behaviour
c1a63b8fad09c419c1a64f75993feb8a343a6801ylavic (which lets the server send garbage): PR 17629
e6b4bd1113567627ab6bb6c6a7105e1e01a7d889jailletc nd says: Every subrequest should get its own filter stack with the
e6b4bd1113567627ab6bb6c6a7105e1e01a7d889jailletc subreq_core filter as bottom-most. That filter does two things:
e466c40e1801982602ee0200c9e8b61cc148742djailletc - swallow EOS buckets
e466c40e1801982602ee0200c9e8b61cc148742djailletc - redirect the data stream to the upper request's (rr->main)
457468b82e59d01eba00dd9d0817309c8f5e414ejim filter chain directly after the subrequest's starting
457468b82e59d01eba00dd9d0817309c8f5e414ejim Once we have a clean solution, we can try to optimize
04983e3bd1754764eec7d6bb772fe3b0bf391771jorton it, so that the server won't be slow down too much.
15890c9306ba98f6fc243e15a3c4778ddc7d773erpluem * RFC 2616 violations.
15660979a30d251681463de2e0584853890082accovener Closed PRs: 15857.
15660979a30d251681463de2e0584853890082accovener Open PRs: 15852, 15859, 15861, 15864, 15865, 15866, 15868, 15869,
49dacedb6c387b786b7911082ff35121a45f414bcovener 15870, 16120, 16125, 16126, 16133, 16135, 16136, 16137,
49dacedb6c387b786b7911082ff35121a45f414bcovener 16138, 16139, 16140, 16142, 16518, 16520, 16521,
cfd9415521847b2f9394fad04fb701cfb955f503rjung jerenkrantz says: need to decide how many we need to backport and/or
cfd9415521847b2f9394fad04fb701cfb955f503rjung if these rise to showstopper status.
cfd9415521847b2f9394fad04fb701cfb955f503rjung wrowe suggests: it would be nice to see "MUST" v.s. "SHOULD" v.s. "MAY"
28c31fb73c1264bd1d0ff932573677030b024c7dwrowe out of this list, without reviewing them individually.
28c31fb73c1264bd1d0ff932573677030b024c7dwrowe * There is a bug in how we sort some hooks, at least the pre-config
28c31fb73c1264bd1d0ff932573677030b024c7dwrowe hook. The first time we call the hooks, they are in the correct
28c31fb73c1264bd1d0ff932573677030b024c7dwrowe order, but the second time, we don't sort them correctly. Currently,
8491e0600f69b0405e156ea8a419653c065c645bcovener the modules/http/config.m4 file has been renamed to
63b9f1f5880391261705f696d7d65507bbe9ace3covener modules/http/config2.m4 to work around this problem, it should moved
63b9f1f5880391261705f696d7d65507bbe9ace3covener back when this is fixed.
49dacedb6c387b786b7911082ff35121a45f414bcovener OtherBill offers that this is a SERIOUS problem. We do not sort
49dacedb6c387b786b7911082ff35121a45f414bcovener correctly by the ordering arguments passed to the register hook
49dacedb6c387b786b7911082ff35121a45f414bcovener functions. This was proven when I reordered the open_logs hook
49dacedb6c387b786b7911082ff35121a45f414bcovener to attempt to open the error logs prior to the access logs. Possibly
3c990331fc6702119e4f5b8ba9eae3021aea5265jim the entire sorting code needs to be refactored.
3c990331fc6702119e4f5b8ba9eae3021aea5265jim * pipes deadlock on all platforms with limited pipe buffers (e.g. both
3c990331fc6702119e4f5b8ba9eae3021aea5265jim Linux and Win32, as opposed to only Win32 on 1.3). The right solution
fc42512879dd0504532f52fe5d0d0383dda96a1eniq is either GStein's proposal for a "CGI Brigade", or OtherBill's proposal
fc42512879dd0504532f52fe5d0d0383dda96a1eniq for "Poll Buckets" for "Polling Filter Chains". Or maybe both :-)
0451df5dc50fa5d8b3e07d92ee6a92e36a1181a5niq * All handlers should always send content down even if r->header_only
0451df5dc50fa5d8b3e07d92ee6a92e36a1181a5niq is set. If not, it means that the HEAD requests don't generate the
0451df5dc50fa5d8b3e07d92ee6a92e36a1181a5niq same headers as a GET which is wrong.
983528026996668ea295be95aedb9c7a346af470ylavic * exec cmd and suexec arg-passing enhancements
da0442c0440caef34706e2c2f3af05cb65921cc0jailletc Status: Patches proposed
da0442c0440caef34706e2c2f3af05cb65921cc0jailletc Message-ID: <20020526041748.A29148@prodigy.Redbrick.DCU.IE>
06b8f183140c8e02e0974e938a05078b511d1603covener (see the "proc.patch" and "suexec-shell.patch" links in this message)
06b8f183140c8e02e0974e938a05078b511d1603covener * The 2.0.36 worker MPM graceless shutdown changes work but are
15890c9306ba98f6fc243e15a3c4778ddc7d773erpluem a bit clunky on some platforms; eg, on Linux, the loop to
259878293a997ff49f5ddfc53d3739cbdc25444ecovener join each worker thread seems to hang, and the parent ends up
259878293a997ff49f5ddfc53d3739cbdc25444ecovener killing off the child with SIGKILL. But at least it shuts down.
259878293a997ff49f5ddfc53d3739cbdc25444ecovener * --enable-mods-shared="foo1 foo2" is busted on Darwin. Pier
15890c9306ba98f6fc243e15a3c4778ddc7d773erpluem posted a patch (Message-ID: <B8DBBE8D.575A%pier@betaversion.org>).
b54b024c06a19926832d77d40ba35ad8c41e4d3dminfrin * We do not properly substitute the prefix-variables in the configuration
b54b024c06a19926832d77d40ba35ad8c41e4d3dminfrin scripts or generated-configs. (i.e. if sysconfdir is etc,
65967d05f839dbf27cf91d91fa79585eeae19660minfrin httpd-std.conf points to conf.)
65967d05f839dbf27cf91d91fa79585eeae19660minfrin * If any request gets through ap_process_request_internal() and is
65967d05f839dbf27cf91d91fa79585eeae19660minfrin scheduled to be served by the core handler, without a flag that this
8152945ae46857b170cb227e79bb799f4fc7710dminfrin r->filename was tested by dir/file_walk, we need to 500 at the very
8152945ae46857b170cb227e79bb799f4fc7710dminfrin end of the ap_process_request_internal() processing so sub_req-esters
8152945ae46857b170cb227e79bb799f4fc7710dminfrin know this request cannot be run. This provides authors of older
8152945ae46857b170cb227e79bb799f4fc7710dminfrin modules better compatibility, while still improving the security and
75f5c2db254c0167a0e396254460de09b775d203trawick robustness of 2.0.
75f5c2db254c0167a0e396254460de09b775d203trawick Status: still need to decide where this goes, OtherBill comments...
4f0358189bfa57b8e75bd6b94db264302a8f336amrumph Message-ID: <065701c14526$495203b0$96c0b0d0@roweclan.net>
4f0358189bfa57b8e75bd6b94db264302a8f336amrumph [Deleted comments regarding the ap_run_handler phase, as irrelevant
4f0358189bfa57b8e75bd6b94db264302a8f336amrumph as BillS points out that "common case will be caught in
5716f9c6daa92dde5f2f9d11ed63f7c9549c223atrawick default_handler already (with the r->finfo.filetype == 0 check)"
5716f9c6daa92dde5f2f9d11ed63f7c9549c223atrawick and the issue is detecting this -before- we try to run the req.]
5716f9c6daa92dde5f2f9d11ed63f7c9549c223atrawick gregames says: can this happen somehow without a broken module
54d750a84a175d8e338880514d440773eb986b50covener being involved? If not, why waste cycles trying to defend against
54d750a84a175d8e338880514d440773eb986b50covener potential broken modules? It seems futile.
54d750a84a175d8e338880514d440773eb986b50covener wrowe counters: no, it shouldn't happen unless the module is broken.
54d750a84a175d8e338880514d440773eb986b50covener But the right answer is to fail the request up-front in dir/file
54d750a84a175d8e338880514d440773eb986b50covener walk if the path was entirely invalid; and we can't do that either
54d750a84a175d8e338880514d440773eb986b50covener UNTIL 2.1 or we break modules that haven't hooked map_to_storage.
54d750a84a175d8e338880514d440773eb986b50covener * With AP_MODE_EXHAUSTIVE in the core, it is finally clear to me
7a3aa12f0eda24793ee26d6a179bd53132e9dae8covener how the Perchild MPM should be re-written. It hasn't worked
54d750a84a175d8e338880514d440773eb986b50covener correctly since filters were added because it wasn't possible to
54d750a84a175d8e338880514d440773eb986b50covener get the content that had already been written and the socket at
83b50288fa7d306324bba68832011ea08f5c7832covener the same time. This mode lets us do that, so the MPM can be
5f066f496cd9f20a2a701255bc67d44e7cb46daetrawick * Can a static httpd be built reliably?
5f066f496cd9f20a2a701255bc67d44e7cb46daetrawick Message-ID: <20020207142751.T31582@clove.org>
2e15620d724fb8e3a5be183b917359a2fd6e9468covener * Usage of APR_BRIGADE_NORMALIZE in core_input_filter should be
2e15620d724fb8e3a5be183b917359a2fd6e9468covener removed if possible.
2e15620d724fb8e3a5be183b917359a2fd6e9468covener Message-ID: <Pine.LNX.4.33.0201202232430.318-100000@deepthought.cs.virginia.edu>
2e15620d724fb8e3a5be183b917359a2fd6e9468covener Jeff wonders if we still care about this. It is no longer an
1b988c41ee505962781d110a3e4c2c90f1ea0aa4covener API issue but simply an extra trip through the brigade.
1b988c41ee505962781d110a3e4c2c90f1ea0aa4covener * Get perchild to work on platforms other than Linux. This
1b988c41ee505962781d110a3e4c2c90f1ea0aa4covener will require a portable mechanism to pass data and file/socket
b8efdc95bec9cf089aa1be0bfd07d46aa1137a7acovener descriptors between vhost child groups. An API was proposed
b8efdc95bec9cf089aa1be0bfd07d46aa1137a7acovener on dev@apr:
b8efdc95bec9cf089aa1be0bfd07d46aa1137a7acovener Message-ID: <20020111115006.K1529@clove.org>
f06e7c4b1bce6b6491e5de0b7998d3f5696b293dchrisd * Try to get libtool inter-library dependency code working on AIX.
f06e7c4b1bce6b6491e5de0b7998d3f5696b293dchrisd Message-ID: <cm3n10lx555.fsf@rdu163-40-092.nc.rr.com>
179565be4043d7e5f9161aa75271fa0a001866d9covener Justin says: If we get it working on AIX, we can enable this
179565be4043d7e5f9161aa75271fa0a001866d9covener on all platforms and clean up our build system
fce4949fb0b309a5744afcd503c6ed2d35621ee2covener Jeff says: I thought I tested a patch for you sometime in
fce4949fb0b309a5744afcd503c6ed2d35621ee2covener January that you were going to commit within a few
7b7430e701e9a31ce809da7c220bb8dfcf68c86etrawick * Handling of %2f in URIs. Currently both 1.3 and 2.0
7b7430e701e9a31ce809da7c220bb8dfcf68c86etrawick completely disallow %2f in the request URI path (see
7b7430e701e9a31ce809da7c220bb8dfcf68c86etrawick ap_unescape_url() in util.c). It's permitted and passed
ccc20788c1e5fc973f36df634399c89acb70deaejerenkrantz through in the query string, however. Roy says the
ccc20788c1e5fc973f36df634399c89acb70deaejerenkrantz original reason for disallowing it, from five years ago,
ccc20788c1e5fc973f36df634399c89acb70deaejerenkrantz was to protect CGI scripts that applied PATH_INFO to
273e512f20f262e5e2aa8e0e83371d1929fb76adjkaluza a filesystem location and which might be tricked by
273e512f20f262e5e2aa8e0e83371d1929fb76adjkaluza ..%2f..%2f(...). We *should* allow path-info of the
273e512f20f262e5e2aa8e0e83371d1929fb76adjkaluza form 'http://foo.com/index.cgi/path/to/path%2finfo'.
efe780dcf13b2b95effabf897d694d8f23feac74trawick Since we've revamped a lot of our processing of path
fe83f60b41477b14a37edcfcd1f7f5c5a1ebfe44minfrin segments, it would be nice to allow this, or at least
fe83f60b41477b14a37edcfcd1f7f5c5a1ebfe44minfrin allow it conditionally with a directive.
993d1261a278d7322bccef219101220b7b4fb8c5jkaluza OtherBill adds that %2f as the SECOND character of a multibyte
993d1261a278d7322bccef219101220b7b4fb8c5jkaluza sequence causes the request to fail! This happens notably in
993d1261a278d7322bccef219101220b7b4fb8c5jkaluza the ja-jis encoding.
ba050a6f942b9fa0e81ed73437588005c569655ccovener * FreeBSD, threads, and worker MPM. All seems to work fine
ba050a6f942b9fa0e81ed73437588005c569655ccovener if you only have one worker process with many threads. Add
ba050a6f942b9fa0e81ed73437588005c569655ccovener a second worker process and the accept lock seems to be
135ddda3a989215d2bedbcf1529bfb269c3eda23niq lost. This might be an APR issue with how it deals with
135ddda3a989215d2bedbcf1529bfb269c3eda23niq the child_init hook (i.e. the fcntl lock needs to be resynced).
135ddda3a989215d2bedbcf1529bfb269c3eda23niq More examination and analysis is required.
001a44c352f89c9ec332ffd3e0a6927dcd19432chumbedooh Status: Works with FreeBSD 5.3. Does not work in previous versions.
001a44c352f89c9ec332ffd3e0a6927dcd19432chumbedooh This has also been reported on Cygwin.
efe780dcf13b2b95effabf897d694d8f23feac74trawick * There is increasing demand from module writers for an API
793214f67dede32edfd9ee96c664ead04d175cbbjfclere that will allow them to control the server � la apachectl.
cc5a4a08dc9783fcbc52ce86f11e01c281a43810minfrin Reasons include sole-function servers that need to die if
9b0076ddd1103e5fa9c1f9bafde4b06ce244fbaecovener an external dependency (e.g., a database) fails, et cetera.
9b0076ddd1103e5fa9c1f9bafde4b06ce244fbaecovener Perhaps something in the (ever more abused) scoreboard?
249d09d51808cb7981af99762c3b3736ca126cd5jkaluza On the other hand, we already have a pipe that goes between parent
249d09d51808cb7981af99762c3b3736ca126cd5jkaluza and child for graceful shutdown events, along with an API that
249d09d51808cb7981af99762c3b3736ca126cd5jkaluza can be used to send a message down that pipe. In threaded MPMs,
249d09d51808cb7981af99762c3b3736ca126cd5jkaluza it is easy enough to make that one pipe be used for graceful
56589be3d7a3e9343370df240010c6928cc78b39jkaluza and graceless events, and it is also easy to open that pipe
56589be3d7a3e9343370df240010c6928cc78b39jkaluza to both parent and child for writing. Then we just need to
56589be3d7a3e9343370df240010c6928cc78b39jkaluza figure out how to do graceless on non-threaded MPMs.
77ca16c5676da23155311e13cee61e7eaba9fa3ejailletc * Allow the DocumentRoot directive within <Location > scopes? This
77ca16c5676da23155311e13cee61e7eaba9fa3ejailletc allows the beloved (crusty) Alias /foo/ /somepath/foo/ followed
77ca16c5676da23155311e13cee61e7eaba9fa3ejailletc by a <Directory /somepath/foo> to become simply
f87299dab99bc04b51a6b8cad51b6795db862c0atrawick <Location /foo/> DocumentRoot /somefile/foo (IMHO a bit more legible
f87299dab99bc04b51a6b8cad51b6795db862c0atrawick and in-your-face.) DocumentRoot unset would be accepted [and would
f87299dab99bc04b51a6b8cad51b6795db862c0atrawick not permit content to be served, only virtual resources such as
4d12805e6c18253040223ea637acd6b3b3c18f60jorton server-info or server-status.
4d12805e6c18253040223ea637acd6b3b3c18f60jorton This proposed change would _not_ depricate Alias.
4d12805e6c18253040223ea637acd6b3b3c18f60jorton striker: See the thread starting with Message-ID:
85eacfc96a04547ef25aabbc06440039715084c2jorton JLEGKKNELMHCJPNMOKHOGEEJFBAA.striker@apache.org.
e5d909f2b06bd880fb3675cd49363df981caa631trawick * Win32: Rotatelogs sometimes is not terminated when Apache
a4df2cd1e1391575a327c2a90ba4315f805a0a78covener goes down hard. FirstBill was looking at possibly tracking the
a4df2cd1e1391575a327c2a90ba4315f805a0a78covener child's-child processes in the parent process.
a4df2cd1e1391575a327c2a90ba4315f805a0a78covener stoddard: Shared scoreboard might offer a good way for the parent
cb666b29f81df1d11d65002250153353568021fccovener to keep track of 'other child' processes and whack them if the child
cb666b29f81df1d11d65002250153353568021fccovener Other thoughts on walking the process chain using the NT kernel
6a80c3c6f4b8ea7ba5e89402b8b779b09ce020e0covener have also been proposed on APR.
6a80c3c6f4b8ea7ba5e89402b8b779b09ce020e0covener * Eliminate unnecessary creation of pipes in mod_cgid
75a230a728338d84dcfe81edd375352f34de22d0covener * Combine log_child and piped_log_spawn. Clean up http_log.c.
75a230a728338d84dcfe81edd375352f34de22d0covener Common logging API.
1f50dc34ae069adeed20b2986e5ffdefa5c410e0covener * Platforms that do not support fork (primarily Win32 and AS/400)
1f50dc34ae069adeed20b2986e5ffdefa5c410e0covener Architect start-up code that avoids initializing all the modules
63a5ea80bddcc84a462e40f402b4f330e0e05411covener in the parent process on platforms that do not support fork.
63a5ea80bddcc84a462e40f402b4f330e0e05411covener * There are still a number of places in the code where we are
63a5ea80bddcc84a462e40f402b4f330e0e05411covener losing error status (i.e. throwing away the error returned by a
65a4e663b82f8bce28ac22ab2edfd7502de36998sf system call and replacing it with a generic error code)
65a4e663b82f8bce28ac22ab2edfd7502de36998sf * Mass vhosting version of suEXEC.
c7de1955eb0eaeabf7042902476397692672d549sf * All DBMs suffer from confusion in support/dbmmanage (perl script) since
74e7f6c55fd67b10cb400b3f6d1dc718a303d944minfrin the dbmmanage employs the first-matched dbm format. This is not
74e7f6c55fd67b10cb400b3f6d1dc718a303d944minfrin necessarily the library that Apache was built with. Aught to
74e7f6c55fd67b10cb400b3f6d1dc718a303d944minfrin rewrite dbmmanage upon installation to bin/ with the proper library
74e7f6c55fd67b10cb400b3f6d1dc718a303d944minfrin for predictable mod_auth_dbm administration.
a511a29faf2ff7ead3b67680154a624effb31aafminfrin Questions; htdbm exists, time to kill dbmmanage, or does it remain
a511a29faf2ff7ead3b67680154a624effb31aafminfrin useful as a perl dbm management example? If we keep it,
a511a29faf2ff7ead3b67680154a624effb31aafminfrin do we address the issue above?
a511a29faf2ff7ead3b67680154a624effb31aafminfrin * Integrate mod_dav.
63921358ef93fcb41bc71d9894221ba3d7fbb87bminfrin Some additional items remaining:
63921358ef93fcb41bc71d9894221ba3d7fbb87bminfrin - case_preserved_filename stuff
63921358ef93fcb41bc71d9894221ba3d7fbb87bminfrin (use the new canonical name stuff?)
deec48c67d4786bc77112ffbf3a4e70b931097edminfrin - find a new home for ap_text(_header)
6d601599d3d65df0410eae6e573e75b2dbfb1fb4minfrin - is it possible to remove the DAV: namespace stuff from util_xml?
6d601599d3d65df0410eae6e573e75b2dbfb1fb4minfrin * ap_core_translate() and its use by mod_mmap_static and mod_file_cache
6d601599d3d65df0410eae6e573e75b2dbfb1fb4minfrin are a bit wonky. The function should probably be exposed as a utility
684e0cfc200f66287a93bbd1708d1dd8a92a7eefcovener function (such as ap_translate_url2fs() or ap_validate_fs_url() or
684e0cfc200f66287a93bbd1708d1dd8a92a7eefcovener something). Another approach would be a new hook phase after
5c43d2fb853f84497b5ece2d414ef9484aa87e5fsf "translate" which would allow the module to munge what the
05a5a9c3e16f21566e1b61f4bd68025ce1b741ccjoes translation has decided to do.
05a5a9c3e16f21566e1b61f4bd68025ce1b741ccjoes Status: Greg +1 (volunteers)
26c5829347f6a355c00f1ba0301d575056b69536niq * Explore use of a post-config hook for the code in http_main.c which
ef82e8fa164e0a1f8b813f7deb6b7ead96018c94niq calls ap_fixup_virutal_hosts(), ap_fini_vhost_config(), and
ef82e8fa164e0a1f8b813f7deb6b7ead96018c94niq ap_sort_hooks() [to reduce the logic in main()]
ef82e8fa164e0a1f8b813f7deb6b7ead96018c94niq * read the config tree just once, and process N times (as necessary)
ef82e8fa164e0a1f8b813f7deb6b7ead96018c94niq * (possibly) use UUIDs in mod_unique_id and/or mod_usertrack
c12917da693bae4028a1d5a5e8224bceed8c739dsf * (possibly) port the bug fix for PR 6942 (segv when LoadModule is put
c12917da693bae4028a1d5a5e8224bceed8c739dsf into a VirtualHost container) to 2.0.
eafcc0ebf263d0ba69855b6e10958c4c1a2361bdsf * shift stuff to mod_core.h
eafcc0ebf263d0ba69855b6e10958c4c1a2361bdsf * callers of ap_run_create_request() should check the return value
eafcc0ebf263d0ba69855b6e10958c4c1a2361bdsf for failure (Doug volunteers)
d7ffd2da16d58b1a0de212e4d56f7aebb72bef26sf * Fix the worker MPM to use POD to kill child processes instead
d7ffd2da16d58b1a0de212e4d56f7aebb72bef26sf of ap_os_killpg, regardless of how they should die.
4576c1a9ef54cd1e5555ee07d016a7f559f80338sf * Scoreboard structures could be changed in the future such that
4576c1a9ef54cd1e5555ee07d016a7f559f80338sf proper alignment is not maintained, leading to segfaults on
9811aed12bbc71783d2e544ccb5fecd193843eadsf some systems. Cliff posted a patch to deal with this issue but
9811aed12bbc71783d2e544ccb5fecd193843eadsf later recanted. See this message to dev@apr.apache.org:
9811aed12bbc71783d2e544ccb5fecd193843eadsf Message-ID: <Pine.LNX.4.44.0203011354090.16457-200000@deepthought
bd3f5647b96d378d9c75c954e3f13582af32c643sf * APXS either needs to be fixed completely for use when apr is out of tree,
bd3f5647b96d378d9c75c954e3f13582af32c643sf or it should drop query mode altogether, and we just grow an
bd3f5647b96d378d9c75c954e3f13582af32c643sf httpd-config or similar arrangement.
bd3f5647b96d378d9c75c954e3f13582af32c643sf To quote a discussion in STATUS earlier:
2a7beea91d46beb41f043a84eaad060047ee04aafabien thommay: this doesn't fix all the problems with apxs and out of
2a7beea91d46beb41f043a84eaad060047ee04aafabien tree apr/apr-util, but it's a good start. There's still the
2a7beea91d46beb41f043a84eaad060047ee04aafabien query cases; but I'm beginning to think that in these cases
2a7beea91d46beb41f043a84eaad060047ee04aafabien the app should be querying ap{r,u}-config directly
584a85dd4047e38d3ed3a29b6662fcc9d100ae4csf gstein: agreed. apxs should deprecate the -q flag
584a85dd4047e38d3ed3a29b6662fcc9d100ae4csf pquerna: I vote for a httpd-config, and to deprecate the -q flag.
584a85dd4047e38d3ed3a29b6662fcc9d100ae4csf minfrin: +1 for httpd-config, and to deprecate -q.
f21e9e3d0bfb7a507ecc5bc963f2159d693503d1sfTODO ISSUES REMAINING IN MOD_SSL:
f6b9c755a0b793e8a3a3aebd327ca20a86478117sf * In order to use a DSO version of mod_ssl we have to link with
f6b9c755a0b793e8a3a3aebd327ca20a86478117sf -lssl and -lcrypto. A workaround is in place right now where the
132ee6ac1c26d6e8953836316ba50734eefab47bsf entire EXTRA_LIBS macro is being appended to the objects list, but
132ee6ac1c26d6e8953836316ba50734eefab47bsf this is a hack. We should either revamp the APACHE_CHECK_SSL_TOOLKIT
132ee6ac1c26d6e8953836316ba50734eefab47bsf autoconf function or come up with some other autoconf checks to
85eacfc96a04547ef25aabbc06440039715084c2jorton search for libssl and libcrypto and properly add them to mod_ssl's
85eacfc96a04547ef25aabbc06440039715084c2jorton link flags.
536d2e7cd1fdec1255b8c3bdf41fdc714c506a54trawick * SSL renegotiations in combination with POST request
536d2e7cd1fdec1255b8c3bdf41fdc714c506a54trawick * Port or dispose all code inside #if 0...#endif blocks that remain
536d2e7cd1fdec1255b8c3bdf41fdc714c506a54trawick from the porting effort.
79c5787b92ac5f0e1cc82393816c77a006399316trawick * Do we need SSL_set_read_ahead()?
79c5787b92ac5f0e1cc82393816c77a006399316trawick * the ssl_expr api is NOT THREAD SAFE. race conditions exist:
c967bf3bc89e8aa60dbd30d9da388e448ddc1cc4trawick -in ssl_expr_comp() if SSLRequire is used in .htaccess
79c5787b92ac5f0e1cc82393816c77a006399316trawick (ssl_expr_info is global)
79c5787b92ac5f0e1cc82393816c77a006399316trawick -is ssl_expr_eval() if there is an error
79c5787b92ac5f0e1cc82393816c77a006399316trawick (ssl_expr_error is global)
79c5787b92ac5f0e1cc82393816c77a006399316trawick * SSLRequire directive (parsing of) leaks memory
7b395e4e878c28a4784919cfd2e704ddd14a3390jorton * Diffie-Hellman-Parameters for temporary keys are hardcoded in
7b395e4e878c28a4784919cfd2e704ddd14a3390jorton ssl_engine_dh.c, while the comment in ssl_engine_kernel.c says:
7b395e4e878c28a4784919cfd2e704ddd14a3390jorton "it is suggested that keys be changed daily or every 500
536e48c08d674acac5d44929318f2ad928edc361jorton transactions, and more often if possible."
e81785da447b469da66f218b3f0244aab507958djorton * ssl_var_lookup could be rewritten to be MUCH faster
3e4e54d4e3fc0123c63d57aa84ac7ad7a8c73ff8jorton * CRL callback should be pluggable
3e4e54d4e3fc0123c63d57aa84ac7ad7a8c73ff8jorton * session cache store should be pluggable
53e9b27aba029b18be814df40bcf6f0428771d1efuankg * init functions should return status code rather than ssl_die()
53e9b27aba029b18be814df40bcf6f0428771d1efuankg * ssl_engine_pphrase.c needs to be reworked so it is generic enough
53e9b27aba029b18be814df40bcf6f0428771d1efuankg to also decrypt proxy keys
6bb524f1895f30265a1431afc460977d391cb36bsf * the shmcb code should just align its memory segment rather than
ca61ccd0c306c2c72df153688ba1b49f3eceed80sf jumping through all the "safe" memcpy and memset hoops
e6dd71992459d05a676b98b7963423dc5dc1e24aminfrin * mod_proxy: Ability to run SSL over proxy gateway connections,
e6dd71992459d05a676b98b7963423dc5dc1e24aminfrin encrypting (or reencrypting) at the proxy.
23f1535d6a60817d2846bac0aea230ea475d7dccminfrin * mod_cache: Handle ESI tags.
23f1535d6a60817d2846bac0aea230ea475d7dccminfrin * mod_cache: Resolve issue of how to cache page fragements (or perhaps
ec7520b24cd80d34d82bbcaca153cbb23cc04bc0rjung -if- we want to cache page fragements). Today, mod_cache/mod_mem_cache
ec7520b24cd80d34d82bbcaca153cbb23cc04bc0rjung will cache #include 'virtual' requests (but not #include 'file'
ec7520b24cd80d34d82bbcaca153cbb23cc04bc0rjung requests). This was accomplished by making CACHE_IN a
ec7520b24cd80d34d82bbcaca153cbb23cc04bc0rjung CONTENT_SET-1 filter to force it to run before the SUBREQ_CORE
ec7520b24cd80d34d82bbcaca153cbb23cc04bc0rjung filter. But now responses cannot be cached that include the
ec7520b24cd80d34d82bbcaca153cbb23cc04bc0rjung effects of having been run through CONTENT_SET filters
ec7520b24cd80d34d82bbcaca153cbb23cc04bc0rjung (mod_deflate, mod_expires, etc). We could rerun all the
ec7520b24cd80d34d82bbcaca153cbb23cc04bc0rjung CONTENT_SET filters on the cached response, but this will not
6249dfa569d3b4f1f539665b979a80c6e335d93etrawick work in all cases. For example, mod_expires relies on installing
6249dfa569d3b4f1f539665b979a80c6e335d93etrawick the EXPIRATION filter during fixups. Contents served out of
0827cb14e550f6f65018431c22c2c913631c8f25kbrand mod_cache (out of the quick_handler) bypass -all- the request
6249dfa569d3b4f1f539665b979a80c6e335d93etrawick line server hooks (Ryan really hated this. It is great for
ae600ca541efc686b34f8b1f21bd3d0741d37674covener performance, but bad because of the complications listed above).
cfa64348224b66dd1c9979b809406c4d15b1c137fielding * mod_mem_cache: Consider adding a RevalidateTimeout directive to
74499a117b3b2cd9666715a14f90c0e5d1a4ee8ajim specify time at which local cached content is to be revalidated
cfa64348224b66dd1c9979b809406c4d15b1c137fielding (ie, underlying file stat'ed to see if it has changed).
cfa64348224b66dd1c9979b809406c4d15b1c137fielding * mod_cache: CacheEnable/CacheDisable should accept regular expressions.
74499a117b3b2cd9666715a14f90c0e5d1a4ee8ajim jerenkrantz says: Too slow. Get regexs away from speedy caches by
cfa64348224b66dd1c9979b809406c4d15b1c137fielding default. Introduce a new CacheEnableRegex if you want.
* mod_mem_cache/mod_disk_cache: Need to be able to query cache
* orig_ct in the byterange/multipart handling may not be