STATUS revision 7558016fa8f4776e989f9d9de7694dcb87077bfe
7f007e36bec06aba6b3a0f84a64f2abf99edfcd8gsteinAPACHE 2.0 STATUS: -*-text-*-
3bb28269556842ebf8888208fd0c7a7f3e343186jerenkrantzLast modified at [$Date: 2002/02/05 22:49:03 $]
ab71b233b3a36489e44a7b061c48293be0b17788jwoolley 2.0.32 : in development
571760de5e60c0b459cb11be45507b923cd023eejwoolley 2.0.31 : rolled Feburary 1, 2002.
571760de5e60c0b459cb11be45507b923cd023eejwoolley 2.0.30 : tagged January 8, 2002. not rolled.
9180a5933673ffb1af633c255ceee029340f3b1erbb 2.0.29 : tagged November 27, 2001. not rolled.
bcb6e1be6041dfeb549c8ea8d37f97ad4e90a0c3rbb 2.0.28 : released November 13, 2001
9bd71e35f5d26d26d23fe3a677401828e842ed72wrowe 2.0.27 : rolled November 6, 2001
2900ab946a2d76b73a14cebfe2985d253f01c967stoddard 2.0.26 : tagged October 16, 2001. not rolled.
a548c09e6a8ca1b059d0e93b5256c6ccb2b3c3cdrbb 2.0.25 : rolled August 29, 2001
a548c09e6a8ca1b059d0e93b5256c6ccb2b3c3cdrbb 2.0.24 : rolled August 18, 2001
a548c09e6a8ca1b059d0e93b5256c6ccb2b3c3cdrbb 2.0.23 : rolled August 9, 2001
b876b7bcf0ce3d232da723246d709e8dbbfe8762rbb 2.0.22 : rolled July 29, 2001
cf6bf6c34c936e6a6fe731dbce4a5c3c8bf8e9a3gstein 2.0.21 : rolled July 20, 2001
35330e0d79ceb8027223bbb8330a381b1f989d6etrawick 2.0.20 : rolled July 8, 2001
6f6f4a4bca281779d196acbdd5c017bb90858305trawick 2.0.19 : rolled June 27, 2001
8dd4618c4709236b4ea297d7250d282e463ce2d8rbb 2.0.18 : rolled May 18, 2001
09bd86d0db1114ee23eda0a6eb76ca055877a1cftrawick 2.0.17 : rolled April 17, 2001
2deb319e6b3de239f45c16a3e9e836d44f1f7108rbb 2.0.16 : rolled April 4, 2001
bd929c73ef04789b7183b840d8db6e01d03a4d86rbb 2.0.15 : rolled March 21, 2001
70f6f32765cfaadd6da8de6f0fea97ddd72d8fadmanoj 2.0.14 : rolled March 7, 2001
2a6c49cfaef5979a5a06098f3ce987cd76769409manoj 2.0a9 : released December 12, 2000
af4c982a7cf4515f124935f99a329744035fc699slive 2.0a8 : released November 20, 2000
af4c982a7cf4515f124935f99a329744035fc699slive 2.0a7 : released October 8, 2000
af4c982a7cf4515f124935f99a329744035fc699slive 2.0a6 : released August 18, 2000
af4c982a7cf4515f124935f99a329744035fc699slive 2.0a5 : released August 4, 2000
af4c982a7cf4515f124935f99a329744035fc699slive 2.0a4 : released June 7, 2000
af4c982a7cf4515f124935f99a329744035fc699slive 2.0a3 : released April 28, 2000
af4c982a7cf4515f124935f99a329744035fc699slive 2.0a2 : released March 31, 2000
2a6c49cfaef5979a5a06098f3ce987cd76769409manoj 2.0a1 : released March 10, 2000
bf9acc131271d18db51d30ace549d3c3b6a2b9fbrbbPlease consult the following STATUS files for information
bf9acc131271d18db51d30ace549d3c3b6a2b9fbrbbon related projects:
79d5106a9b65b956d646f5daae4b94bc79e315b8trawickRELEASE SHOWSTOPPERS:
79d5106a9b65b956d646f5daae4b94bc79e315b8trawick * 31 BETA STATUS:
79d5106a9b65b956d646f5daae4b94bc79e315b8trawick running on Daedalus since 02-Feb-2002 7:58 PST (need 3 days)
79d5106a9b65b956d646f5daae4b94bc79e315b8trawick Compiles on : AIX 4.3, Solaris, FreeBSD 3.4 & 4.5, Win32,
79d5106a9b65b956d646f5daae4b94bc79e315b8trawick Linux 2.2 & 2.4, HPUX
9eccc0a669a3e711629345b357c46acce5fefdb4gstein +1 : Jeff Trawick, BrianP, Aaron, Jim
750fcd3495b59eff6d4844409ae1dfb46d026059gstein -1 : BillS, BillR, Ian, gregames, Justin
750fcd3495b59eff6d4844409ae1dfb46d026059gstein bumps since original tag:
750fcd3495b59eff6d4844409ae1dfb46d026059gstein * mod-dir patch
750fcd3495b59eff6d4844409ae1dfb46d026059gstein * scoreboard x2 : 1 to fix gracefull restarts
750fcd3495b59eff6d4844409ae1dfb46d026059gstein 1 to fix netware
750fcd3495b59eff6d4844409ae1dfb46d026059gstein * win32/locks.c : to fix mod_rewrite on win32
750fcd3495b59eff6d4844409ae1dfb46d026059gstein problems with v31:
750fcd3495b59eff6d4844409ae1dfb46d026059gstein * libtool/binbuild on AIX -- possible addition of patched
9eccc0a669a3e711629345b357c46acce5fefdb4gstein binbuild.sh to 31-beta roll
ec75f189410513ab8f6e1173a9d9d277ebec9ce7gstein +1: Ian, Justin
ec75f189410513ab8f6e1173a9d9d277ebec9ce7gstein Jeff says: We can't do anything about libtool since AIX
ec75f189410513ab8f6e1173a9d9d277ebec9ce7gstein needs a version that won't work on some
ec75f189410513ab8f6e1173a9d9d277ebec9ce7gstein platforms. Handle this in the README.
ec75f189410513ab8f6e1173a9d9d277ebec9ce7gstein The binbuild issue isn't AIX. It is
ec75f189410513ab8f6e1173a9d9d277ebec9ce7gstein something that could happen anywhere that
dc96a5e6f9af3c514df4c61ab9468fcf97f9846fwrowe the binbuild-er has their own expat.
dc96a5e6f9af3c514df4c61ab9468fcf97f9846fwrowe Handle this by patching binbuild.sh for a
dc96a5e6f9af3c514df4c61ab9468fcf97f9846fwrowe beta roll or putting a patch in the README for
dc96a5e6f9af3c514df4c61ab9468fcf97f9846fwrowe use by people who want to do binbuild but
dc96a5e6f9af3c514df4c61ab9468fcf97f9846fwrowe have expat installed locally.
dc96a5e6f9af3c514df4c61ab9468fcf97f9846fwrowe gregames: why can't we roll a second tarball with the
dc96a5e6f9af3c514df4c61ab9468fcf97f9846fwrowe appropriate libtool version? doesn't
dc96a5e6f9af3c514df4c61ab9468fcf97f9846fwrowe Darwin need it as well as AIX?
dc96a5e6f9af3c514df4c61ab9468fcf97f9846fwrowe * erroneous check in an AP_DEBUG_ASSERT() call. Only happens
dc96a5e6f9af3c514df4c61ab9468fcf97f9846fwrowe in maintainer mode. Fixed in modules/http/http_protocol.c
dc96a5e6f9af3c514df4c61ab9468fcf97f9846fwrowe revision 1.391. Add item in release notes:
dc96a5e6f9af3c514df4c61ab9468fcf97f9846fwrowe +1: Justin, Cliff, Aaron, BillS, Jim, trawick
7bce59d998f2e5ca1cb60038ef6c1d0817605d62stoddard * seg faults in core_input_filter when the client goes away
7bce59d998f2e5ca1cb60038ef6c1d0817605d62stoddard before any POST body bytes are received.
7bce59d998f2e5ca1cb60038ef6c1d0817605d62stoddard Justin is tracking this down.
7bce59d998f2e5ca1cb60038ef6c1d0817605d62stoddard * FirstBill reports problem [re]starting as-a-service, shared
7bce59d998f2e5ca1cb60038ef6c1d0817605d62stoddard score is suspect. OtherBill is investigating, definately
7bce59d998f2e5ca1cb60038ef6c1d0817605d62stoddard a showstopper for a .31/.32 beta. Progress report later today.
2a6c49cfaef5979a5a06098f3ce987cd76769409manoj * mod_auth_dbm can't open a Berkely DB password file on Unix
33a70d93b992bbb63b809c24df29451f38f017c1rbb Justin postulates that this might be related to the fact
410912d9cb56a09a74f8655647ed9e81d49023f4gregames that mod_auth_dbm wasn't using apr-util. See
410912d9cb56a09a74f8655647ed9e81d49023f4gregames modules/aaa/mod_auth_dbm.c revision 1.42 and
60d567a0c2aae815ee6fc20c0d65032bea52c92cwrowe * Madhu reports that the worker MPM doesn't shutdown nicely
60d567a0c2aae815ee6fc20c0d65032bea52c92cwrowe after a stress test.
60d567a0c2aae815ee6fc20c0d65032bea52c92cwrowe * mod_autoindex displays the wrong icon for subdirectories
60d567a0c2aae815ee6fc20c0d65032bea52c92cwrowe * anon shared memory not allocating enough
7239216999e746bb4fc7671621becea33c5c1c87stoddard * ap_directory_walk skips some per-dir config merge functions
d180ec1b29106f4fec480ef7fcdb04df078010cerse if there is no "<Directory />" block in the configuration
d180ec1b29106f4fec480ef7fcdb04df078010cerse Message-ID: <m3itbdiijq.fsf@rdu163-40-092.nc.rr.com>
d180ec1b29106f4fec480ef7fcdb04df078010cerse * That is very unlikely, merges are additive. Much more likely,
d180ec1b29106f4fec480ef7fcdb04df078010cerse the default SetOutputFilter default or merge is borked.
d180ec1b29106f4fec480ef7fcdb04df078010cerse Unless it's the code that permits 'null' merges, per module.
d180ec1b29106f4fec480ef7fcdb04df078010cerse Still, it's probably in SetOutputFilter's behavior.
d180ec1b29106f4fec480ef7fcdb04df078010cerse * BrianP notes: directory_walk's handling of trailing slashes
d180ec1b29106f4fec480ef7fcdb04df078010cerse appears to be the cause: http://www.apachelabs.org/apache-mbox/200112.mbox/%3c3C1CF721.1090300@pacbell.net%3e
d180ec1b29106f4fec480ef7fcdb04df078010cerse * Test suite failures:
d180ec1b29106f4fec480ef7fcdb04df078010cerse o worker is also failing some of the 'cgi' subtests
d180ec1b29106f4fec480ef7fcdb04df078010cerse Justin says: "Worker should be fine and passes httpd-test here.
d180ec1b29106f4fec480ef7fcdb04df078010cerse If you can provide evidence that it can be reproduced
d180ec1b29106f4fec480ef7fcdb04df078010cerse outside of httpd-test, then it's a showstopper. I
d180ec1b29106f4fec480ef7fcdb04df078010cerse think it's a perl or a httpd-test problem."
d180ec1b29106f4fec480ef7fcdb04df078010cerse Not a showstopper: Justin
d180ec1b29106f4fec480ef7fcdb04df078010cerse * If any request gets to the core handler, without a flag that this
d180ec1b29106f4fec480ef7fcdb04df078010cerse r->filename was tested by dir/file_walk, we need to 500 at the very
d180ec1b29106f4fec480ef7fcdb04df078010cerse end of the ap_process_request_internal() processing. This provides
d180ec1b29106f4fec480ef7fcdb04df078010cerse authors of older modules better compatibility, while still improving
d180ec1b29106f4fec480ef7fcdb04df078010cerse the security and robustness of 2.0.
d180ec1b29106f4fec480ef7fcdb04df078010cerse Status: still need to decide where this goes, OtherBill comments...
d180ec1b29106f4fec480ef7fcdb04df078010cerse Message-ID: <065701c14526$495203b0$96c0b0d0@roweclan.net>
e32adabcbf3bf5b69ba2e8b163b971839efc94dbtrawick we need to look at halting this in the 'default handler' case,
e32adabcbf3bf5b69ba2e8b163b971839efc94dbtrawick and that implies pushing the 'handler election' into the request
e32adabcbf3bf5b69ba2e8b163b971839efc94dbtrawick internal processing phase from the run request phase.
7bce59d998f2e5ca1cb60038ef6c1d0817605d62stoddard * There is a bug in how we sort some hooks, at least the pre-config
7bce59d998f2e5ca1cb60038ef6c1d0817605d62stoddard hook. The first time we call the hooks, they are in the correct
7bce59d998f2e5ca1cb60038ef6c1d0817605d62stoddard order, but the second time, we don't sort them correctly. Currently,
7bce59d998f2e5ca1cb60038ef6c1d0817605d62stoddard the modules/http/config.m4 file has been renamed to
49facccad3f5c3e9e49311487b5069699c3bf3fdjwoolley modules/http/config2.m4 to work around this problem, it should moved
7bce59d998f2e5ca1cb60038ef6c1d0817605d62stoddard back when this is fixed. rbb
7bce59d998f2e5ca1cb60038ef6c1d0817605d62stoddard Justin says: "Is this really a showstopper? This has been here
7bce59d998f2e5ca1cb60038ef6c1d0817605d62stoddard forever. What's wrong? Does this have to do with
7bce59d998f2e5ca1cb60038ef6c1d0817605d62stoddard autoconf or m4?"
7bce59d998f2e5ca1cb60038ef6c1d0817605d62stoddard Not a showstopper: Justin, BrianP, trawick, gregames
49facccad3f5c3e9e49311487b5069699c3bf3fdjwoolley * Convert all instances of the old apr_lock_t type to the new
49facccad3f5c3e9e49311487b5069699c3bf3fdjwoolley types (once they are fully supported in APR).
49facccad3f5c3e9e49311487b5069699c3bf3fdjwoolley Status: Aaron is working on converting INTRAPROCESS
49facccad3f5c3e9e49311487b5069699c3bf3fdjwoolley to apr_thread_mutex_t types. Full replacements for
49facccad3f5c3e9e49311487b5069699c3bf3fdjwoolley LOCKALL and CROSS_PROCESS are not yet complete on all
7bce59d998f2e5ca1cb60038ef6c1d0817605d62stoddard platforms, and should only be used in MPMs like worker
19cbe4d7b7c931723e7249de6829bf965a1fee72stoddard with limited OS exposure.
19cbe4d7b7c931723e7249de6829bf965a1fee72stoddard BrianP asks: "Is this really a showstopper?"
93db592309ba9e5ab230f67611a2c74fece9cdb2marc IanH says: "If we don't do it before we go live we will have 2
93db592309ba9e5ab230f67611a2c74fece9cdb2marc different API's to support in APR"
93db592309ba9e5ab230f67611a2c74fece9cdb2marc * A binbuild installation picks up the right libraries when
93db592309ba9e5ab230f67611a2c74fece9cdb2marc running apachectl because we set the appropriate environment
93db592309ba9e5ab230f67611a2c74fece9cdb2marc variable, but ab, htpasswd, etc. don't know how to pick up apr,
93db592309ba9e5ab230f67611a2c74fece9cdb2marc apr-util, etc.
b187d568e1507d75139ebc13ca945b38fc05d55cstoddard Message-ID: <20020116000226.GA15991@ebuilt.com>
b187d568e1507d75139ebc13ca945b38fc05d55cstoddard Justin says: "You could always build all of the binaries
b187d568e1507d75139ebc13ca945b38fc05d55cstoddard statically if you don't want to let libtool
1c6fb1e726ce22694de0e9a957adb67b929e5d4fstoddard handle it. At most, I don't consider this a
1c6fb1e726ce22694de0e9a957adb67b929e5d4fstoddard a release showstopper as it is a packaging issue."
8bed76428f56e5c643174a2d6807c3f18016af5cbjh Ken sez: "Showstopper because, until it's the default, it
8bed76428f56e5c643174a2d6807c3f18016af5cbjh violates the P of LA. Binaries will give strange
a5ed555df952c85bc1b179f5981e8a6c54ba16e6stoddard errors and not run if moved."
d2f8b010487ffa990a9c268df5a25579e7291bcdrbb Status: Aaron volunteers: I'll commit a patch to allow all
d2f8b010487ffa990a9c268df5a25579e7291bcdrbb the support binaries to be built statically against
a5ed555df952c85bc1b179f5981e8a6c54ba16e6stoddard libapr/libaprutil/etc and then enable it in binbuild.
0bff2f28ef945280c17099c142126178a78e1e54manoj gregames says: what about httpd? yeah, it works at the moment
0bff2f28ef945280c17099c142126178a78e1e54manoj if you use apachectl. But I've already heard
0bff2f28ef945280c17099c142126178a78e1e54manoj complaints about not being able to execute
1e585ba09ea32272e63c4c39c35491e975d21d98stoddard httpd directly.
35330e0d79ceb8027223bbb8330a381b1f989d6etrawickRELEASE NON-SHOWSTOPPERS BUT WOULD BE REAL NICE TO WRAP THESE UP:
9c09943bad734ebd5c7cc10bd6d63b75c4c6e056stoddard * The Add...Filter and Set...Filter directives do not allow the
ff849e4163ed879288f0df15f78b6c9d278ec804fanf administrator to order filters, beyond the order of filename (mime)
ff849e4163ed879288f0df15f78b6c9d278ec804fanf extensions. It isn't clear if Set...Filter(s) should be inserted
447c6ce3ff08073c44f6785d5256271fcb877512wrowe before or after the Add...Filter(s) which are ordered by sequence of
447c6ce3ff08073c44f6785d5256271fcb877512wrowe filename extensions. At minimum, some sort of +-[0-10] syntax seems
447c6ce3ff08073c44f6785d5256271fcb877512wrowe like the quickest fix for a 2.0 gold release.
447c6ce3ff08073c44f6785d5256271fcb877512wrowe Justin says: "Could we delay this for a point release or 2.1?"
447c6ce3ff08073c44f6785d5256271fcb877512wrowe Not a showstopper: justin, wrowe, trawick, stoddard, Jim, Ian, Aaron,
cf6bf6c34c936e6a6fe731dbce4a5c3c8bf8e9a3gstein * Should we always build binaries statically unless otherwise
3bb28269556842ebf8888208fd0c7a7f3e343186jerenkrantz Message-ID: <20020129210006.B23512@Lithium.MeepZor.Com>
c03566fa0156d3a1500a42e4fe539e3e0fc8a11dgstein +1: gregames
db3ccce11afac4fc1d4f51a65424412f7480c46cgstein -1: Justin, Ian
0eb7ca6cf812d98c534661ac474e873a32bf6325gstein * If the parent process dies, should the remaining child processes
cf6bf6c34c936e6a6fe731dbce4a5c3c8bf8e9a3gstein "gracefully" self-terminate. Or maybe we should make it a runtime
8d07897b52e3b7055874501f8a499e75800db206gstein option, or have a concept of 2 parent processes (one being a
8d07897b52e3b7055874501f8a499e75800db206gstein "hot spare").
db3ccce11afac4fc1d4f51a65424412f7480c46cgstein See: Message-ID: <3C58232C.FE91F19F@Golux.Com>
79d5106a9b65b956d646f5daae4b94bc79e315b8trawick Self-destruct: Ken, Martin
79d5106a9b65b956d646f5daae4b94bc79e315b8trawick Not self-destruct: BrianP, Ian, Cliff, BillS
cf6bf6c34c936e6a6fe731dbce4a5c3c8bf8e9a3gstein Make it runtime configurable: Aaron, Jim, Justin
cf6bf6c34c936e6a6fe731dbce4a5c3c8bf8e9a3gstein Have 2 parents: +1: Jim
6fa71a1bd8c61518b05f5798a7a1594c270e78afrbb +0: Martin (while standing by, could it do
93c5cba06b623ebe8e4372e886eece12d9a80c3egstein something useful?)
14cccaddba3a9263cf0d0ddc311e18f3e3dc9b0fgstein look at accept() error handling first: gregames
14cccaddba3a9263cf0d0ddc311e18f3e3dc9b0fgstein * Make some modifications to the scoreboard creation routines,
14cccaddba3a9263cf0d0ddc311e18f3e3dc9b0fgstein prefer anonymous shared memory, and allow a configuration
14cccaddba3a9263cf0d0ddc311e18f3e3dc9b0fgstein directive to override the defaults.
14cccaddba3a9263cf0d0ddc311e18f3e3dc9b0fgstein Message-ID: <20020130080804.C16977@clove.org>
823c303d33c9e637a83d82208bcbafaf5f532d7bgstein Status: Aaron volunteers
e636eba7474e0010b5c7198af1c2fe5ad8652dbbmanoj * Get perchild to work on platforms other than Linux. This
e636eba7474e0010b5c7198af1c2fe5ad8652dbbmanoj will require a portable mechanism to pass data and file/socket
e636eba7474e0010b5c7198af1c2fe5ad8652dbbmanoj descriptors between vhost child groups. An API was proposed
281da4c02cf40c663298ded7e4e5b913a8f8b814gstein on dev@apr:
281da4c02cf40c663298ded7e4e5b913a8f8b814gstein Message-ID: <20020111115006.K1529@clove.org>
333eac96e4fb7d6901cb75e6ca7bb22b2ccb84cetrawick Vote: Is a non-portable perchild going to hold up a GA release?
2f728b2e8555fee1b7cc11e886488692f2575fbddougm No: Aaron, Justin, trawick, stoddard, Jim, Ian, BrianP, gregames
2f728b2e8555fee1b7cc11e886488692f2575fbddougm * Recent changes to ap_rgetline may have broken EBCDIC boxes.
114cc51b75ed89d5ab2f39b85ba7d472293f3dedtrawick Message-ID: <20020122072605.GF28051@ebuilt.com>
114cc51b75ed89d5ab2f39b85ba7d472293f3dedtrawick Justin says: "I don't have an EBCDIC box to test on. A potential
114cc51b75ed89d5ab2f39b85ba7d472293f3dedtrawick solution is to split out ap_rgetline into two
114cc51b75ed89d5ab2f39b85ba7d472293f3dedtrawick functions as described in this message."
60d567a0c2aae815ee6fc20c0d65032bea52c92cwrowe gregames says: I see the breakage now, and volunteer to fix it
60d567a0c2aae815ee6fc20c0d65032bea52c92cwrowe when things calm down a little. It looks OK when
60d567a0c2aae815ee6fc20c0d65032bea52c92cwrowe there are complete lines and no mime continuations.
60d567a0c2aae815ee6fc20c0d65032bea52c92cwrowe * Modify the worker MPM so that it doesn't need to create and
60d567a0c2aae815ee6fc20c0d65032bea52c92cwrowe destroy a pool for each request--possibly by adopting a
60d567a0c2aae815ee6fc20c0d65032bea52c92cwrowe leader/follower model in which each worker owns a persistent
60d567a0c2aae815ee6fc20c0d65032bea52c92cwrowe ptrans pool (like the prefork MPM) and the workers take
60d567a0c2aae815ee6fc20c0d65032bea52c92cwrowe turns acting as listeners...this approach might also help
60d567a0c2aae815ee6fc20c0d65032bea52c92cwrowe reduce context-switching
60d567a0c2aae815ee6fc20c0d65032bea52c92cwrowe * CGI single-byte reads
64ad864fa0f4493eebb181e393b40a8a90beccb9coar BrianP suggests that this is caused by the ap_scan_script_header_err()
64ad864fa0f4493eebb181e393b40a8a90beccb9coar routine, which will do single-byte reads until it finds the end
64ad864fa0f4493eebb181e393b40a8a90beccb9coar of the header, at which point it constructs a pipe-bucket (buffered)
64ad864fa0f4493eebb181e393b40a8a90beccb9coar to read from.
28d1da9ca818f831ea491f110dafcc10f7f07050coar Proposed solution in:
64ad864fa0f4493eebb181e393b40a8a90beccb9coar Message-ID: <3C36ADAF.60601@cnet.com>
64ad864fa0f4493eebb181e393b40a8a90beccb9coar * Try to get libtool inter-library dependency code working on AIX.
64ad864fa0f4493eebb181e393b40a8a90beccb9coar Message-ID: <cm3n10lx555.fsf@rdu163-40-092.nc.rr.com>
64ad864fa0f4493eebb181e393b40a8a90beccb9coar Justin says: If we get it working on AIX, we can enable this
28d1da9ca818f831ea491f110dafcc10f7f07050coar on all platforms and clean up our build system
64ad864fa0f4493eebb181e393b40a8a90beccb9coar Jeff says: I thought I tested a patch for you sometime in
64ad864fa0f4493eebb181e393b40a8a90beccb9coar January that you were going to commit within a few
64ad864fa0f4493eebb181e393b40a8a90beccb9coar * Handling of %2f in URIs. Currently both 1.3 and 2.0
64ad864fa0f4493eebb181e393b40a8a90beccb9coar completely disallow %2f in the request URI path (see
64ad864fa0f4493eebb181e393b40a8a90beccb9coar ap_unescape_url() in util.c). It's permitted and passed
28d1da9ca818f831ea491f110dafcc10f7f07050coar through in the query string, however. Roy says the
64ad864fa0f4493eebb181e393b40a8a90beccb9coar original reason for disallowing it, from five years ago,
64ad864fa0f4493eebb181e393b40a8a90beccb9coar was to protect CGI scripts that applied PATH_INFO to
64ad864fa0f4493eebb181e393b40a8a90beccb9coar a filesystem location and which might be tricked by
28d1da9ca818f831ea491f110dafcc10f7f07050coar ..%2f..%2f(...). We *should* allow path-info of the
64ad864fa0f4493eebb181e393b40a8a90beccb9coar Since we've revamped a lot of our processing of path
64ad864fa0f4493eebb181e393b40a8a90beccb9coar segments, it would be nice to allow this, or at least
28d1da9ca818f831ea491f110dafcc10f7f07050coar allow it conditionally with a directive.
64ad864fa0f4493eebb181e393b40a8a90beccb9coar * FreeBSD, threads, and worker MPM. All seems to work fine
64ad864fa0f4493eebb181e393b40a8a90beccb9coar if you only have one worker process with many threads. Add
64ad864fa0f4493eebb181e393b40a8a90beccb9coar a second worker process and the accept lock seems to be
28d1da9ca818f831ea491f110dafcc10f7f07050coar lost. This might be an APR issue with how it deals with
64ad864fa0f4493eebb181e393b40a8a90beccb9coar the child_init hook (i.e. the fcntl lock needs to be resynced).
64ad864fa0f4493eebb181e393b40a8a90beccb9coar More examination and analysis is required.
64ad864fa0f4493eebb181e393b40a8a90beccb9coar Status: This has also been reported on Cygwin.
64ad864fa0f4493eebb181e393b40a8a90beccb9coar Message-ID: <3C2CC514.8EF3BED1@wapme-systems.de> (cygnus)
64ad864fa0f4493eebb181e393b40a8a90beccb9coar Justin says: So, FreeBSD-CURRENT and Cywin have the same
28d1da9ca818f831ea491f110dafcc10f7f07050coar problem. Yum. If another platform has this
64ad864fa0f4493eebb181e393b40a8a90beccb9coar with worker, this becomes a showstopper.
64ad864fa0f4493eebb181e393b40a8a90beccb9coar Aaron says: I spent some time disecting this and have come to
64ad864fa0f4493eebb181e393b40a8a90beccb9coar the conclusion that it is not a problem in the worker MPM
28d1da9ca818f831ea491f110dafcc10f7f07050coar (or at least, it is not isolated to a problem in worker).
64ad864fa0f4493eebb181e393b40a8a90beccb9coar I'll list some of the problems I'm seeing in case someone
64ad864fa0f4493eebb181e393b40a8a90beccb9coar else wants to pick up where I've left off:
64ad864fa0f4493eebb181e393b40a8a90beccb9coar - Delivery of just about any signal to one of the child
28d1da9ca818f831ea491f110dafcc10f7f07050coar processes will send it into an infinite loop as well.
64ad864fa0f4493eebb181e393b40a8a90beccb9coar - Even though the parent is spinning out of control,
64ad864fa0f4493eebb181e393b40a8a90beccb9coar at first the child or children will appear to work
64ad864fa0f4493eebb181e393b40a8a90beccb9coar properly. At times it is possible to get it into a state,
28d1da9ca818f831ea491f110dafcc10f7f07050coar however, where a request will hang until another concurrent
64ad864fa0f4493eebb181e393b40a8a90beccb9coar request "kicks" the first, at which point the second will
64ad864fa0f4493eebb181e393b40a8a90beccb9coar hang. My theory is that this has to do with the
64ad864fa0f4493eebb181e393b40a8a90beccb9coar pthread_cond_*() implementation in FreeBSD, but it's still
28d1da9ca818f831ea491f110dafcc10f7f07050coar possible that it is in APR.
64ad864fa0f4493eebb181e393b40a8a90beccb9coar Justin adds: Oh, FreeBSD threads are implemented entirely with
64ad864fa0f4493eebb181e393b40a8a90beccb9coar select()/poll()/longjmp(). Welcome to the nightmare.
28d1da9ca818f831ea491f110dafcc10f7f07050coar So, that means a ktrace output also has the thread
64ad864fa0f4493eebb181e393b40a8a90beccb9coar scheduling internals in it (since it is all the same to
64ad864fa0f4493eebb181e393b40a8a90beccb9coar the kernel). Which makes it hard to distinguish between
64ad864fa0f4493eebb181e393b40a8a90beccb9coar our select() calls and their select() calls.
64ad864fa0f4493eebb181e393b40a8a90beccb9coar *bangs head on wall repeatedly* But, some of the libc_r
28d1da9ca818f831ea491f110dafcc10f7f07050coar files have a DBG_MSG #define. This is moderately helpful
64ad864fa0f4493eebb181e393b40a8a90beccb9coar when used with -DNO_DETACH. The kernel scheduler isn't
64ad864fa0f4493eebb181e393b40a8a90beccb9coar waking up the threads on a select(). Yum. And, I bet
64ad864fa0f4493eebb181e393b40a8a90beccb9coar those decrementing select calls have to do with the
28d1da9ca818f831ea491f110dafcc10f7f07050coar scheduler. Time to brush up on our OS fundamentals.
64ad864fa0f4493eebb181e393b40a8a90beccb9coar * There is increasing demand from module writers for an API
64ad864fa0f4493eebb181e393b40a8a90beccb9coar that will allow them to control the server � la apachectl.
28d1da9ca818f831ea491f110dafcc10f7f07050coar Reasons include sole-function servers that need to die if
64ad864fa0f4493eebb181e393b40a8a90beccb9coar an external dependency (e.g., a database) fails, et cetera.
64ad864fa0f4493eebb181e393b40a8a90beccb9coar Perhaps something in the (ever more abused) scoreboard?
64ad864fa0f4493eebb181e393b40a8a90beccb9coar rbb: I don't believe the scoreboard is the correct mechanism
28d1da9ca818f831ea491f110dafcc10f7f07050coar for this. We already have a pipe that goes between parent
64ad864fa0f4493eebb181e393b40a8a90beccb9coar and child for graceful shutdown events, along with an API that
64ad864fa0f4493eebb181e393b40a8a90beccb9coar can be used to send a message down that pipe. In threaded MPMs,
64ad864fa0f4493eebb181e393b40a8a90beccb9coar it is easy enough to make that one pipe be used for graceful
28d1da9ca818f831ea491f110dafcc10f7f07050coar and graceless events, and it is also easy to open that pipe
64ad864fa0f4493eebb181e393b40a8a90beccb9coar to both parent and child for writing. Then we just need to figure
64ad864fa0f4493eebb181e393b40a8a90beccb9coar out how to do graceless on non-threaded MPMs.
28d1da9ca818f831ea491f110dafcc10f7f07050coar * revamp the input filter behavior, per discussions since
64ad864fa0f4493eebb181e393b40a8a90beccb9coar February (and especially at the hackathon last
64ad864fa0f4493eebb181e393b40a8a90beccb9coar April). Specifically, ap_get_brigade will return a brigade with
64ad864fa0f4493eebb181e393b40a8a90beccb9coar *up to* a specific number of bytes, or a "line" of data. The
28d1da9ca818f831ea491f110dafcc10f7f07050coar read may be blocking or nonblocking. ap_getline() will be
64ad864fa0f4493eebb181e393b40a8a90beccb9coar refactored into apr_brigade_getline(), and then DECHUNK can use
64ad864fa0f4493eebb181e393b40a8a90beccb9coar f->next (ap_getline will always read "top of input stack"). Also
64ad864fa0f4493eebb181e393b40a8a90beccb9coar fix the bug where request body content will end up closing the
28d1da9ca818f831ea491f110dafcc10f7f07050coar connection (buggering up persistent conns).
64ad864fa0f4493eebb181e393b40a8a90beccb9coar Status: Justin is working on this as fast as he can.
64ad864fa0f4493eebb181e393b40a8a90beccb9coar The core input filters, HTTP-related filters, mod_ssl,
64ad864fa0f4493eebb181e393b40a8a90beccb9coar mod_proxy, and ap_[r]getline are switched to the new logic.
64ad864fa0f4493eebb181e393b40a8a90beccb9coar - socket bucket and core input filter changes. see end of
64ad864fa0f4493eebb181e393b40a8a90beccb9coar message ID (Feb 27): <20010227075326.S2297@lyra.org>
28d1da9ca818f831ea491f110dafcc10f7f07050coar - fix up ap_get_brigade() semantics, fix bug in DECHUNK /
64ad864fa0f4493eebb181e393b40a8a90beccb9coar ap_getline. many messages (plus their threads) (Apr/May):
64ad864fa0f4493eebb181e393b40a8a90beccb9coar Message-ID: <20010402101207.J27539@lyra.org>
64ad864fa0f4493eebb181e393b40a8a90beccb9coar Message-ID: <3AF7F921.D2EEC41A@algroup.co.uk>
28d1da9ca818f831ea491f110dafcc10f7f07050coar Message-ID: <20010508190029.E18404@lyra.org>
64ad864fa0f4493eebb181e393b40a8a90beccb9coar - further work with combining/tweaking the builtin filters:
64ad864fa0f4493eebb181e393b40a8a90beccb9coar Message-ID: <20010509115445.D1374@lyra.org>
64ad864fa0f4493eebb181e393b40a8a90beccb9coar - thoughts on filter modes:
64ad864fa0f4493eebb181e393b40a8a90beccb9coar Message-ID: <021b01c14dee$09782af0$93c0b0d0@roweclan.net>
28d1da9ca818f831ea491f110dafcc10f7f07050coar - proposal for getline rewrite:
64ad864fa0f4493eebb181e393b40a8a90beccb9coar Message-ID: <20011231104019.GD3904@ebuilt.com>
64ad864fa0f4493eebb181e393b40a8a90beccb9coar - patch for getline rewrite:
28d1da9ca818f831ea491f110dafcc10f7f07050coar Message-ID: <20020122072605.GF28051@ebuilt.com>
64ad864fa0f4493eebb181e393b40a8a90beccb9coar - Change ap_get_brigade prototype:
64ad864fa0f4493eebb181e393b40a8a90beccb9coar Message-ID: <20020120115136.GI17601@ebuilt.com>
64ad864fa0f4493eebb181e393b40a8a90beccb9coar * Allow the DocumentRoot directive within <Location > scopes? This
64ad864fa0f4493eebb181e393b40a8a90beccb9coar allows the beloved (crusty) Alias /foo/ /somepath/foo/ followed
64ad864fa0f4493eebb181e393b40a8a90beccb9coar by a <Directory /somepath/foo> to become simply
28d1da9ca818f831ea491f110dafcc10f7f07050coar <Location /foo/> DocumentRoot /somefile/foo (IMHO a bit more legible
64ad864fa0f4493eebb181e393b40a8a90beccb9coar and in-your-face.) DocumentRoot unset would be accepted [and would
6694e265e9a71ceaedbe1f1aa4db4d9ba42fb866wrowe not permit content to be served, only virtual resources such as
64ad864fa0f4493eebb181e393b40a8a90beccb9coar server-info or server-status.
28d1da9ca818f831ea491f110dafcc10f7f07050coar This proposed change would _not_ depricate Alias.
64ad864fa0f4493eebb181e393b40a8a90beccb9coar * Win32: Rotatelogs sometimes is not terminated when Apache
64ad864fa0f4493eebb181e393b40a8a90beccb9coar goes down hard. FirstBill was looking at possibly tracking the
28d1da9ca818f831ea491f110dafcc10f7f07050coar child's-child processes in the parent process.
64ad864fa0f4493eebb181e393b40a8a90beccb9coar OtherBill asks, wasn't this fixed?
64ad864fa0f4493eebb181e393b40a8a90beccb9coar stoddard: Not fixed. Shared scoreboard might offer a good
64ad864fa0f4493eebb181e393b40a8a90beccb9coar way for the parent to keep track of 'other child' processes
28d1da9ca818f831ea491f110dafcc10f7f07050coar and whack them if the child goes down.
64ad864fa0f4493eebb181e393b40a8a90beccb9coar * Win32: Add a simple hold console open patch (wait for close or
64ad864fa0f4493eebb181e393b40a8a90beccb9coar the ESC key, with a nice message) if the server died a bad
28d1da9ca818f831ea491f110dafcc10f7f07050coar death (non-zero exit code) in console mode.
64ad864fa0f4493eebb181e393b40a8a90beccb9coar Resolution: bring forward same ugly hacks from 1.3.13-.20
64ad864fa0f4493eebb181e393b40a8a90beccb9coar * Port of mod_ssl to Apache 2.0:
64ad864fa0f4493eebb181e393b40a8a90beccb9coar The current porting state is summarized in modules/ssl/README. The
64ad864fa0f4493eebb181e393b40a8a90beccb9coar remaining work includes:
64ad864fa0f4493eebb181e393b40a8a90beccb9coar (1) stablizing/optimizing the SSL filter logic
64ad864fa0f4493eebb181e393b40a8a90beccb9coar (2) Enabling the various SSL caching mechanisms (shmcb, shmht)
28d1da9ca818f831ea491f110dafcc10f7f07050coar (3) Enabling SSL extentions
64ad864fa0f4493eebb181e393b40a8a90beccb9coar (4) Trying to seperate the https filter logic from mod_ssl -
64ad864fa0f4493eebb181e393b40a8a90beccb9coar This is to facilitate other modules that wish to use the https
64ad864fa0f4493eebb181e393b40a8a90beccb9coar filter or the mod_ssl logic or both as required.
64ad864fa0f4493eebb181e393b40a8a90beccb9coar Justin: mod_ssl filter logic is redone, so that should be fine.
28d1da9ca818f831ea491f110dafcc10f7f07050coar Madhu has submitted a patch for SSL caching - however, I
64ad864fa0f4493eebb181e393b40a8a90beccb9coar am -0 on that patch as I *think* we could implement the
64ad864fa0f4493eebb181e393b40a8a90beccb9coar shared memory another way that is much cleaner (i.e.
64ad864fa0f4493eebb181e393b40a8a90beccb9coar treat shmem directly as a dbm via APR routines). Justin
28d1da9ca818f831ea491f110dafcc10f7f07050coar also thinks that the https filter logic may be sufficiently
64ad864fa0f4493eebb181e393b40a8a90beccb9coar decoupled now, but isn't really sure.
64ad864fa0f4493eebb181e393b40a8a90beccb9coar * Performance & Debug: Eliminate most (and perhaps all) of the
28d1da9ca818f831ea491f110dafcc10f7f07050coar malloc/free calls in the bucket brigade code. Need some
64ad864fa0f4493eebb181e393b40a8a90beccb9coar light weight memory management functions that allow freeing
64ad864fa0f4493eebb181e393b40a8a90beccb9coar memory (putting it back into a memory pool) when it is no
64ad864fa0f4493eebb181e393b40a8a90beccb9coar longer needed. Enabling simple debugging features like guard
28d1da9ca818f831ea491f110dafcc10f7f07050coar bands, double free detection, etc. would be cool but certainly
64ad864fa0f4493eebb181e393b40a8a90beccb9coar not a hard requirement.
64ad864fa0f4493eebb181e393b40a8a90beccb9coar Status: Cliff started to implement this using SMS as has
28d1da9ca818f831ea491f110dafcc10f7f07050coar been discussed at length for months, but since
64ad864fa0f4493eebb181e393b40a8a90beccb9coar SMS is not being used anywhere else in the server,
64ad864fa0f4493eebb181e393b40a8a90beccb9coar several people expressed the opinion that we should
64ad864fa0f4493eebb181e393b40a8a90beccb9coar get rid of it entirely, meaning that the buckets
64ad864fa0f4493eebb181e393b40a8a90beccb9coar need their own memory management (free list) functions.
28d1da9ca818f831ea491f110dafcc10f7f07050coar Cliff will implement that this weekend so we at least
64ad864fa0f4493eebb181e393b40a8a90beccb9coar have something to look at/compare with.
64ad864fa0f4493eebb181e393b40a8a90beccb9coar * Eliminate unnecessary creation of pipes in mod_cgid
64ad864fa0f4493eebb181e393b40a8a90beccb9coar * the autoconf setup should be fixed to default to using the
64ad864fa0f4493eebb181e393b40a8a90beccb9coar "Apache" layout from config.layout, and each variable settable
64ad864fa0f4493eebb181e393b40a8a90beccb9coar in a layout should be overridable on the command line. Plus,
28d1da9ca818f831ea491f110dafcc10f7f07050coar what we do right now just doesn't seem to fully fit into how autoconf
64ad864fa0f4493eebb181e393b40a8a90beccb9coar works, eg. AC_PREFIX_DEFAULT issues.
64ad864fa0f4493eebb181e393b40a8a90beccb9coar Message-ID: <Pine.BSF.4.20.0104031557420.20876-100000@alive.znep.com>
64ad864fa0f4493eebb181e393b40a8a90beccb9coar * Combine log_child and piped_log_spawn. Clean up http_log.c.
28d1da9ca818f831ea491f110dafcc10f7f07050coar Common logging API.
64ad864fa0f4493eebb181e393b40a8a90beccb9coar * Document mod_file_cache.
28d1da9ca818f831ea491f110dafcc10f7f07050coar * Platforms that do not support fork (primarily Win32 and AS/400)
64ad864fa0f4493eebb181e393b40a8a90beccb9coar Architect start-up code that avoids initializing all the modules
64ad864fa0f4493eebb181e393b40a8a90beccb9coar in the parent process on platforms that do not support fork.
28d1da9ca818f831ea491f110dafcc10f7f07050coar * Win32: Migrate the MPM over to use APR thread/process calls. This
64ad864fa0f4493eebb181e393b40a8a90beccb9coar would eliminate some code in the Win32 branch that essentially
64ad864fa0f4493eebb181e393b40a8a90beccb9coar duplicates what is in APR.
28d1da9ca818f831ea491f110dafcc10f7f07050coar * There are still a number of places in the code where we are
64ad864fa0f4493eebb181e393b40a8a90beccb9coar losing error status (i.e. throwing away the error returned by a
64ad864fa0f4493eebb181e393b40a8a90beccb9coar system call and replacing it with a generic error code)
28d1da9ca818f831ea491f110dafcc10f7f07050coar * Mass vhosting version of suEXEC.
64ad864fa0f4493eebb181e393b40a8a90beccb9coar * All DBMs suffer from confusion in support/dbmmanage (perl script) since
64ad864fa0f4493eebb181e393b40a8a90beccb9coar the dbmmanage employs the first-matched dbm format. This is not
28d1da9ca818f831ea491f110dafcc10f7f07050coar necessarily the library that Apache was built with. Aught to
64ad864fa0f4493eebb181e393b40a8a90beccb9coar rewrite dbmmanage upon installation to bin/ with the proper library
64ad864fa0f4493eebb181e393b40a8a90beccb9coar for predictable mod_auth_dbm administration.
64ad864fa0f4493eebb181e393b40a8a90beccb9coar Questions; htdbm exists, time to kill dbmmanage, or does it remain
28d1da9ca818f831ea491f110dafcc10f7f07050coar useful as a perl dbm management example? If we keep it,
64ad864fa0f4493eebb181e393b40a8a90beccb9coar do we address the issue above?
64ad864fa0f4493eebb181e393b40a8a90beccb9coar * use apu_dbm in mod_auth_dbm
28d1da9ca818f831ea491f110dafcc10f7f07050coar Status: Greg +1 (low-priority volunteer)
64ad864fa0f4493eebb181e393b40a8a90beccb9coar Justin says: "Seems like this is already there, so should we just
64ad864fa0f4493eebb181e393b40a8a90beccb9coar remove the other DBM code in that file? If you want
64ad864fa0f4493eebb181e393b40a8a90beccb9coar to use gdbm, or dbm, etc, you should tell apr-util."
28d1da9ca818f831ea491f110dafcc10f7f07050coar Will says: "bs - I may choose the fastest - most efficient native
64ad864fa0f4493eebb181e393b40a8a90beccb9coar dbm implementation, for shared proc caches, ssl session
64ad864fa0f4493eebb181e393b40a8a90beccb9coar caching, etc, but that has nothing to do with maintaining
64ad864fa0f4493eebb181e393b40a8a90beccb9coar a userlist via dbm, which has to remain readable between
28d1da9ca818f831ea491f110dafcc10f7f07050coar builds/machines, etc. The use-multiple database schema
64ad864fa0f4493eebb181e393b40a8a90beccb9coar for apr-util would let us do this with just apr, though."
64ad864fa0f4493eebb181e393b40a8a90beccb9coar Ian says: "multi-dbm is in, but it still has ndbm support hardcoded
64ad864fa0f4493eebb181e393b40a8a90beccb9coar is this still required? isn't ndbm supported via gdbm?"
64ad864fa0f4493eebb181e393b40a8a90beccb9coar * Integrate mod_dav.
64ad864fa0f4493eebb181e393b40a8a90beccb9coar Some additional items remaining:
28d1da9ca818f831ea491f110dafcc10f7f07050coar - case_preserved_filename stuff
64ad864fa0f4493eebb181e393b40a8a90beccb9coar (use the new canonical name stuff?)
64ad864fa0f4493eebb181e393b40a8a90beccb9coar - find a new home for ap_text(_header)
64ad864fa0f4493eebb181e393b40a8a90beccb9coar - is it possible to remove the DAV: namespace stuff from util_xml?
64ad864fa0f4493eebb181e393b40a8a90beccb9coar * ap_core_translate() and its use by mod_mmap_static and mod_file_cache
64ad864fa0f4493eebb181e393b40a8a90beccb9coar are a bit wonky. The function should probably be exposed as a utility
64ad864fa0f4493eebb181e393b40a8a90beccb9coar function (such as ap_translate_url2fs() or ap_validate_fs_url() or
28d1da9ca818f831ea491f110dafcc10f7f07050coar something). Another approach would be a new hook phase after
64ad864fa0f4493eebb181e393b40a8a90beccb9coar "translate" which would allow the module to munge what the
64ad864fa0f4493eebb181e393b40a8a90beccb9coar translation has decided to do.
64ad864fa0f4493eebb181e393b40a8a90beccb9coar Status: Greg +1 (volunteers), Ryan +1
64ad864fa0f4493eebb181e393b40a8a90beccb9coar * Explore use of a post-config hook for the code in http_main.c which
64ad864fa0f4493eebb181e393b40a8a90beccb9coar calls ap_fixup_virutal_hosts(), ap_fini_vhost_config(), and
64ad864fa0f4493eebb181e393b40a8a90beccb9coar ap_sort_hooks() [to reduce the logic in main()]
64ad864fa0f4493eebb181e393b40a8a90beccb9coar * read the config tree just once, and process N times (as necessary)
64ad864fa0f4493eebb181e393b40a8a90beccb9coar * (possibly) use UUIDs in mod_unique_id and/or mod_usertrack
64ad864fa0f4493eebb181e393b40a8a90beccb9coar * (possibly) port the bug fix for PR 6942 (segv when LoadModule is put
64ad864fa0f4493eebb181e393b40a8a90beccb9coar into a VirtualHost container) to 2.0.
28d1da9ca818f831ea491f110dafcc10f7f07050coar * shift stuff to mod_core.h
64ad864fa0f4493eebb181e393b40a8a90beccb9coar * callers of ap_run_create_request() should check the return value
64ad864fa0f4493eebb181e393b40a8a90beccb9coar for failure (Doug volunteers)
64ad864fa0f4493eebb181e393b40a8a90beccb9coar * Win32: Get Apache working on Windows 95/98. The following work
64ad864fa0f4493eebb181e393b40a8a90beccb9coar (at least) needs to be done:
64ad864fa0f4493eebb181e393b40a8a90beccb9coar - winnt MPM: Fix 95/98 code paths in the winnt MPM. There is some NT
28d1da9ca818f831ea491f110dafcc10f7f07050coar specific code that is still not in NT only code paths
64ad864fa0f4493eebb181e393b40a8a90beccb9coar - IOL binds to APR sendfile, implemented with TransmitFile, which
64ad864fa0f4493eebb181e393b40a8a90beccb9coar is not available on 95/98.
64ad864fa0f4493eebb181e393b40a8a90beccb9coar - Document warning that OSR2 is required (for Crypt functions, in
64ad864fa0f4493eebb181e393b40a8a90beccb9coar rand.c, at least.) This could be resolved with an SSL library, or
28d1da9ca818f831ea491f110dafcc10f7f07050coar randomization in APR itself.
64ad864fa0f4493eebb181e393b40a8a90beccb9coar - Bring the Win9xConHook.dll from 1.3 into 2.0 (no sense till it
64ad864fa0f4493eebb181e393b40a8a90beccb9coar actually works) and add in a splash of Win9x service code.
28d1da9ca818f831ea491f110dafcc10f7f07050coar * In order to use a DSO version of mod_ssl we have to link with
64ad864fa0f4493eebb181e393b40a8a90beccb9coar -lssl and -lcrypto. A workaround is in place right now where the
64ad864fa0f4493eebb181e393b40a8a90beccb9coar entire EXTRA_LIBS macro is being appended to the objects list, but
64ad864fa0f4493eebb181e393b40a8a90beccb9coar this is a hack. We should either revamp the APACHE_CHECK_SSL_TOOLKIT
64ad864fa0f4493eebb181e393b40a8a90beccb9coar autoconf function or come up with some other autoconf checks to
28d1da9ca818f831ea491f110dafcc10f7f07050coar search for libssl and libcrypto and properly add them to mod_ssl's
64ad864fa0f4493eebb181e393b40a8a90beccb9coar link flags.
64ad864fa0f4493eebb181e393b40a8a90beccb9coar * Make the worker MPM the default MPM for threaded Unix boxes.
28d1da9ca818f831ea491f110dafcc10f7f07050coar +1: Justin, Jeff, Ian
64ad864fa0f4493eebb181e393b40a8a90beccb9coar -0: Aaron (premature decision, needs more discussion), Lars
64ad864fa0f4493eebb181e393b40a8a90beccb9coar -0: Cliff (I think the default config should be the safest possible)
28d1da9ca818f831ea491f110dafcc10f7f07050coar * Fix the worker MPM to use POD to kill child processes instead
64ad864fa0f4493eebb181e393b40a8a90beccb9coar of ap_os_killpg, regardless of how they should die. (Ryan Bloom)
64ad864fa0f4493eebb181e393b40a8a90beccb9coarPRs that have been suspended forever waiting for someone to
28d1da9ca818f831ea491f110dafcc10f7f07050coarput them into 'the next release':
64ad864fa0f4493eebb181e393b40a8a90beccb9coar * PR#76: general
64ad864fa0f4493eebb181e393b40a8a90beccb9coar missing call to "setlocale();"
64ad864fa0f4493eebb181e393b40a8a90beccb9coar * PR#78: mod_include
64ad864fa0f4493eebb181e393b40a8a90beccb9coar Additional status for XBitHack directive
64ad864fa0f4493eebb181e393b40a8a90beccb9coar * PR#362: mod_proxy
64ad864fa0f4493eebb181e393b40a8a90beccb9coar Mod_proxy doesn't allow change of error pages
64ad864fa0f4493eebb181e393b40a8a90beccb9coar * PR#370: mod_env
64ad864fa0f4493eebb181e393b40a8a90beccb9coar Modified PATH environemnt variable is not passed, instead
64ad864fa0f4493eebb181e393b40a8a90beccb9coar system's is used
64ad864fa0f4493eebb181e393b40a8a90beccb9coar * PR#440: mod_proxy
64ad864fa0f4493eebb181e393b40a8a90beccb9coar Proxy doesn't deliver documents if not connected
64ad864fa0f4493eebb181e393b40a8a90beccb9coar * PR#534: mod_proxy
64ad864fa0f4493eebb181e393b40a8a90beccb9coar proxy converts ~name to %7Ename when name starts with a dot (.)
64ad864fa0f4493eebb181e393b40a8a90beccb9coar * PR#537: mod_access
64ad864fa0f4493eebb181e393b40a8a90beccb9coar mod_access syntax allows hosts that should be restricted
64ad864fa0f4493eebb181e393b40a8a90beccb9coar * PR#557: mod_auth-any
64ad864fa0f4493eebb181e393b40a8a90beccb9coar ~UserHome directories are not honored in absolute pathname
64ad864fa0f4493eebb181e393b40a8a90beccb9coar requests (.htaccess)
64ad864fa0f4493eebb181e393b40a8a90beccb9coar * PR#612: mod_proxy
120ff80b38fc1f72cdc911d62bb249a7c0c7262bwrowe Proxy FTP Authentication Fails
64ad864fa0f4493eebb181e393b40a8a90beccb9coar * PR#623: mod_include
64ad864fa0f4493eebb181e393b40a8a90beccb9coar A smarter "Last Modified" value for SSI documents (see PR number 600)
64ad864fa0f4493eebb181e393b40a8a90beccb9coar * PR#628: config
64ad864fa0f4493eebb181e393b40a8a90beccb9coar Request of "Options SymLinksIfGroupMatch"
64ad864fa0f4493eebb181e393b40a8a90beccb9coar * PR#700: mod_proxy
64ad864fa0f4493eebb181e393b40a8a90beccb9coar Proxy doesn't do links right for OpenVMS files through ftp:
64ad864fa0f4493eebb181e393b40a8a90beccb9coar * PR#759: mod_imap
64ad864fa0f4493eebb181e393b40a8a90beccb9coar imap should read <MAP><AREA>*</MAP> too!
28d1da9ca818f831ea491f110dafcc10f7f07050coar * PR#793: general
64ad864fa0f4493eebb181e393b40a8a90beccb9coar RLimitCPU and RLimitMEM don't apply to all children like they should
64ad864fa0f4493eebb181e393b40a8a90beccb9coar * PR#921: suexec
28d1da9ca818f831ea491f110dafcc10f7f07050coar Uses cwd before filling it in, doesn't use syslog
64ad864fa0f4493eebb181e393b40a8a90beccb9coar * PR#922: config
28d1da9ca818f831ea491f110dafcc10f7f07050coar it is useful to allow specifiction that root-owned symlinks
64ad864fa0f4493eebb181e393b40a8a90beccb9coar should always be followed
28d1da9ca818f831ea491f110dafcc10f7f07050coar * PR#980: mod_proxy
64ad864fa0f4493eebb181e393b40a8a90beccb9coar Controlling Access to Remote Proxies would be nice...
28d1da9ca818f831ea491f110dafcc10f7f07050coar * PR#994: mod_proxy
64ad864fa0f4493eebb181e393b40a8a90beccb9coar Adding authentication "on the fly" through the proxy module
64ad864fa0f4493eebb181e393b40a8a90beccb9coar * PR#1004: apache-api
28d1da9ca818f831ea491f110dafcc10f7f07050coar request_config field in request_rec is moderately bogus
2a6c49cfaef5979a5a06098f3ce987cd76769409manoj * PR#1028: other
2a6c49cfaef5979a5a06098f3ce987cd76769409manoj DoS attacks involving memory consumption
2a6c49cfaef5979a5a06098f3ce987cd76769409manoj * PR#1050: mod_log-any
2a6c49cfaef5979a5a06098f3ce987cd76769409manoj Logging of virtual server to error_log as well
2a6c49cfaef5979a5a06098f3ce987cd76769409manoj * PR#1085: mod_proxy
2a6c49cfaef5979a5a06098f3ce987cd76769409manoj ProxyRemote make a dead cycle.
0bff2f28ef945280c17099c142126178a78e1e54manoj * PR#1117: mod_auth-any
d6b3cb141f0667101c1bca883ad15b383402c93bfielding Using NIS passwd.byname dbm files with AuthDBMUserFile
ec0315cdf832eac2b78e50ad636af84fe4c9118cgstein * PR#1120: suexec
ec0315cdf832eac2b78e50ad636af84fe4c9118cgstein suexec does not parse arguments to #exec cmd
ec0315cdf832eac2b78e50ad636af84fe4c9118cgstein * PR#1145: mod_include
ec0315cdf832eac2b78e50ad636af84fe4c9118cgstein Allow for Last-Modified: without resorting to XBitHack
2a6c49cfaef5979a5a06098f3ce987cd76769409manoj * PR#1158: apache-api
2a6c49cfaef5979a5a06098f3ce987cd76769409manoj improvements to child spawning API
dbf0c7bef06259486cd2748a2d0e82f27e099d6efielding * PR#1166: mod_proxy
dbf0c7bef06259486cd2748a2d0e82f27e099d6efielding ``nph-'' not honored (no buffering) for ProxyRemote mapping
3e17185356213124b2e18ecaf1678a676f8e9ba5rbb * PR#1176: mod_cgi
3e17185356213124b2e18ecaf1678a676f8e9ba5rbb Apache cannot handle continuation line in headers
3e17185356213124b2e18ecaf1678a676f8e9ba5rbb * PR#1191: general
863ec32e13d6c9619414c48b51109f3dca99cbc6wrowe setlogin() is not called, causing problems with e.g. identd
863ec32e13d6c9619414c48b51109f3dca99cbc6wrowe * PR#1204: general
863ec32e13d6c9619414c48b51109f3dca99cbc6wrowe regerror() exists, use it
2a6c49cfaef5979a5a06098f3ce987cd76769409manoj * PR#1233: apache-api
2a6c49cfaef5979a5a06098f3ce987cd76769409manoj there is no way to keep per-connection per-module state
ae6907470ddf23ab7c6b506e6407cc5372f9c0dftrawick * PR#1263: mod_autoexec
ae6907470ddf23ab7c6b506e6407cc5372f9c0dftrawick Add frame-safe anchor attribute to mod_autoindex links
ef7ff29d74b8684761de5758423f1cc5a49f48e3rbb * PR#1268: suexec
ef7ff29d74b8684761de5758423f1cc5a49f48e3rbb CGI scripts running as Apache user: security (suexec etc.)
2a6c49cfaef5979a5a06098f3ce987cd76769409manoj * PR#1285: suexec
2a6c49cfaef5979a5a06098f3ce987cd76769409manoj Error messages could be easier to spot in cgi.log file for suexec.c
d6b3cb141f0667101c1bca883ad15b383402c93bfielding * PR#1287: mod_access
add allow,deny/deny,allow warning to mod_access
A small addition to rotatelogs.c to improve program functionality.
Feedback/Comment on APACI
Inclusion of RPM spec file in CVS/distributions
* orig_ct in the byterange/multipart handling may not be