STATUS revision 0bcb1fe39dfaacf9745b6633f5cc9ebc8e2596ca
9174efb969475801d0dc88eee35aae40c748d450ndAPACHE 2.0 STATUS: -*-text-*-
ec79b29695b183f794264bbb578c51e93d1f9b1emartinLast modified at [$Date: 2002/02/01 06:05:51 $]
95817edd05387a5276f51fcd5db79fc21b89b55brooneg 2.0.32 : in development
95817edd05387a5276f51fcd5db79fc21b89b55brooneg 2.0.31 : tagged January 30, 2002.
63689d77e084e36b8194fb6df5adfc0344965e01trawick 2.0.30 : tagged January 8, 2002. not rolled.
63689d77e084e36b8194fb6df5adfc0344965e01trawick 2.0.29 : tagged November 27, 2001. not rolled.
63689d77e084e36b8194fb6df5adfc0344965e01trawick 2.0.28 : released November 13, 2001
63689d77e084e36b8194fb6df5adfc0344965e01trawick 2.0.27 : rolled November 6, 2001
148c51524ef39a89416ea52d58afacb5d839a060trawick 2.0.26 : tagged October 16, 2001. not rolled.
148c51524ef39a89416ea52d58afacb5d839a060trawick 2.0.25 : rolled August 29, 2001
148c51524ef39a89416ea52d58afacb5d839a060trawick 2.0.24 : rolled August 18, 2001
148c51524ef39a89416ea52d58afacb5d839a060trawick 2.0.23 : rolled August 9, 2001
148c51524ef39a89416ea52d58afacb5d839a060trawick 2.0.22 : rolled July 29, 2001
5714cdd83e23557d801437daa5e3ab8ba78ae595jorton 2.0.21 : rolled July 20, 2001
5714cdd83e23557d801437daa5e3ab8ba78ae595jorton 2.0.20 : rolled July 8, 2001
5714cdd83e23557d801437daa5e3ab8ba78ae595jorton 2.0.19 : rolled June 27, 2001
784e72b9673ae72f981ae5ee062330bc30ad9efcjim 2.0.18 : rolled May 18, 2001
d43d9b67e5fad1d183febeb2ccc09dfd52e9bd54rpluem 2.0.17 : rolled April 17, 2001
784e72b9673ae72f981ae5ee062330bc30ad9efcjim 2.0.16 : rolled April 4, 2001
784e72b9673ae72f981ae5ee062330bc30ad9efcjim 2.0.15 : rolled March 21, 2001
784e72b9673ae72f981ae5ee062330bc30ad9efcjim 2.0.14 : rolled March 7, 2001
d43d9b67e5fad1d183febeb2ccc09dfd52e9bd54rpluem 2.0a9 : released December 12, 2000
784e72b9673ae72f981ae5ee062330bc30ad9efcjim 2.0a8 : released November 20, 2000
dace23289120f3af62555d6754d80e9e51d4249erooneg 2.0a7 : released October 8, 2000
dace23289120f3af62555d6754d80e9e51d4249erooneg 2.0a6 : released August 18, 2000
dace23289120f3af62555d6754d80e9e51d4249erooneg 2.0a5 : released August 4, 2000
53412cc2f0c1676b1df5cc65f54e1784e0dc42carpluem 2.0a4 : released June 7, 2000
53412cc2f0c1676b1df5cc65f54e1784e0dc42carpluem 2.0a3 : released April 28, 2000
53412cc2f0c1676b1df5cc65f54e1784e0dc42carpluem 2.0a2 : released March 31, 2000
53412cc2f0c1676b1df5cc65f54e1784e0dc42carpluem 2.0a1 : released March 10, 2000
1aa8ab8152a5b85f2e0e0cfa0cc4226416685123trawickPlease consult the following STATUS files for information
a1a615ca49b162d71d88089210395c9a9cfeb539rpluemon related projects:
69c36bbae91de0e99a682aaae9d6fa61fceb2771bnicholesRELEASE SHOWSTOPPERS:
69c36bbae91de0e99a682aaae9d6fa61fceb2771bnicholes * Runtime startup failures don't reliably set the exit status.
69c36bbae91de0e99a682aaae9d6fa61fceb2771bnicholes For instance, try listening to port 80 and starting as non-root;
a1a615ca49b162d71d88089210395c9a9cfeb539rpluem server fails with make_sock and 'no listening sockets' errors
69c36bbae91de0e99a682aaae9d6fa61fceb2771bnicholes but exits with 0 status. Even with -DONE_PROCESS -DNO_DETACH.
69c36bbae91de0e99a682aaae9d6fa61fceb2771bnicholes Justin says: "Is this really a showstopper? Most people won't
69c36bbae91de0e99a682aaae9d6fa61fceb2771bnicholes be using -DNO_DETACH and will look at the error
69c36bbae91de0e99a682aaae9d6fa61fceb2771bnicholes logs anyway."
69c36bbae91de0e99a682aaae9d6fa61fceb2771bnicholes * mod_rewrite's locks (win32) were broken by recent changes to the
69c36bbae91de0e99a682aaae9d6fa61fceb2771bnicholes apr_lock_t api (which we were deprecating in the first place, so
69c36bbae91de0e99a682aaae9d6fa61fceb2771bnicholes I'm far from certain why it changed.) It has to do with creating
69c36bbae91de0e99a682aaae9d6fa61fceb2771bnicholes the rewrite logs mutex. The solution is to find the brokenness in
a1a615ca49b162d71d88089210395c9a9cfeb539rpluem apr_lock_t, but if it goes away, why waste the time?
69c36bbae91de0e99a682aaae9d6fa61fceb2771bnicholes * ap_directory_walk skips some per-dir config merge functions
69c36bbae91de0e99a682aaae9d6fa61fceb2771bnicholes if there is no "<Directory />" block in the configuration
69c36bbae91de0e99a682aaae9d6fa61fceb2771bnicholes Message-ID: <m3itbdiijq.fsf@rdu163-40-092.nc.rr.com>
69c36bbae91de0e99a682aaae9d6fa61fceb2771bnicholes * That is very unlikely, merges are additive. Much more likely,
69c36bbae91de0e99a682aaae9d6fa61fceb2771bnicholes the default SetOutputFilter default or merge is borked.
69c36bbae91de0e99a682aaae9d6fa61fceb2771bnicholes Unless it's the code that permits 'null' merges, per module.
69c36bbae91de0e99a682aaae9d6fa61fceb2771bnicholes Still, it's probably in SetOutputFilter's behavior.
69c36bbae91de0e99a682aaae9d6fa61fceb2771bnicholes * Test suite failures:
69c36bbae91de0e99a682aaae9d6fa61fceb2771bnicholes o worker is also failing some of the 'cgi' subtests
69c36bbae91de0e99a682aaae9d6fa61fceb2771bnicholes (see <URL:http://Source-Zone.Org/Apache/regression/>):
69c36bbae91de0e99a682aaae9d6fa61fceb2771bnicholes Justin says: "Worker should be fine and passes httpd-test here."
1b0dce86d7fc8a5aa4c89b05255be26e508c615crpluem * If any request gets to the core handler, without a flag that this
1b0dce86d7fc8a5aa4c89b05255be26e508c615crpluem r->filename was tested by dir/file_walk, we need to 500 at the very
1b0dce86d7fc8a5aa4c89b05255be26e508c615crpluem end of the ap_process_request_internal() processing. This provides
1b0dce86d7fc8a5aa4c89b05255be26e508c615crpluem authors of older modules better compatibility, while still improving
b608bdb688ec8fbe57c760766aaaeb1ff5c11f5fbrianp the security and robustness of 2.0.
b608bdb688ec8fbe57c760766aaaeb1ff5c11f5fbrianp Status: still need to decide where this goes, OtherBill comments...
b608bdb688ec8fbe57c760766aaaeb1ff5c11f5fbrianp Message-ID: <065701c14526$495203b0$96c0b0d0@roweclan.net>
4d8cb6462304414eae9acb78ca95d56ef1afa0edjerenkrantz we need to look at halting this in the 'default handler' case,
4d8cb6462304414eae9acb78ca95d56ef1afa0edjerenkrantz and that implies pushing the 'handler election' into the request
4d8cb6462304414eae9acb78ca95d56ef1afa0edjerenkrantz internal processing phase from the run request phase.
269457ff24416ce42a2b810fd0b8c6e62a04cbf3rpluem * There is a bug in how we sort some hooks, at least the pre-config
269457ff24416ce42a2b810fd0b8c6e62a04cbf3rpluem hook. The first time we call the hooks, they are in the correct
c1e9dd034870fa337890e93d3dba8de37ceed122jerenkrantz order, but the second time, we don't sort them correctly. Currently,
edc5389f50ce4153e6192740f3c7a188c8cf8d67niq the modules/http/config.m4 file has been renamed to
edc5389f50ce4153e6192740f3c7a188c8cf8d67niq modules/http/config2.m4 to work around this problem, it should moved
6c05afd314b4ddd545d63b4ff5de822cc30eec79trawick back when this is fixed. rbb
6c05afd314b4ddd545d63b4ff5de822cc30eec79trawick Justin says: "Is this really a showstopper? This has been here
6c05afd314b4ddd545d63b4ff5de822cc30eec79trawick forever. What's wrong? Does this have to do with
13cd67e9c1dacbd6b9f040bda337c725cedd98f3brianp autoconf or m4?"
a623efbff95aab78da9e030524b0fa69b054f6d0brianp * The Add...Filter and Set...Filter directives do not allow the
a623efbff95aab78da9e030524b0fa69b054f6d0brianp administrator to order filters, beyond the order of filename (mime)
a623efbff95aab78da9e030524b0fa69b054f6d0brianp extensions. It isn't clear if Set...Filter(s) should be inserted
a623efbff95aab78da9e030524b0fa69b054f6d0brianp before or after the Add...Filter(s) which are ordered by sequence of
a623efbff95aab78da9e030524b0fa69b054f6d0brianp filename extensions. At minimum, some sort of +-[0-10] syntax seems
a623efbff95aab78da9e030524b0fa69b054f6d0brianp like the quickest fix for a 2.0 gold release.
0b4b04d8621478ba59f0a6ba2950ddc02ab92b58colm Justin says: "Could we delay this for a point release or 2.1?"
0b4b04d8621478ba59f0a6ba2950ddc02ab92b58colm * Convert all instances of the old apr_lock_t type to the new
2f1bb5376c5c4022383bb729679ca751dd75a2eabrianp types (once they are fully supported in APR).
2f1bb5376c5c4022383bb729679ca751dd75a2eabrianp Status: Aaron is working on converting INTRAPROCESS
ad862ab5716726a2d72a292ba1dfb29566c86153brianp to apr_thread_mutex_t types. Full replacements for
ad862ab5716726a2d72a292ba1dfb29566c86153brianp LOCKALL and CROSS_PROCESS are not yet complete on all
ad862ab5716726a2d72a292ba1dfb29566c86153brianp platforms, and should only be used in MPMs like worker
7461431ba407b0e1eac3d6a81440a4184e652e9fniq with limited OS exposure.
17d53ea32c4968e47733f1c2c063ae07d280efd6jerenkrantz * A binbuild installation picks up the right libraries when
17d53ea32c4968e47733f1c2c063ae07d280efd6jerenkrantz running apachectl because we set the appropriate environment
17d53ea32c4968e47733f1c2c063ae07d280efd6jerenkrantz variable, but ab, htpasswd, etc. don't know how to pick up apr,
2d5532b13110a8d85653da92e97795b09cc25cc2trawick apr-util, etc.
b38565306421ff53e9f7499bc728d6df5cec294dpquerna Message-ID: <20020116000226.GA15991@ebuilt.com>
b38565306421ff53e9f7499bc728d6df5cec294dpquerna Justin says: "You could always build all of the binaries
b38565306421ff53e9f7499bc728d6df5cec294dpquerna statically if you don't want to let libtool
b38565306421ff53e9f7499bc728d6df5cec294dpquerna handle it. At most, I don't consider this a
cbb903af5066589fe0e73f3ecf06abdc71e38effrpluem a release showstopper as it is a packaging issue."
cbb903af5066589fe0e73f3ecf06abdc71e38effrpluem Status: Aaron volunteers: I'll commit a patch to allow all
412da000890e343cb9cc01e9958f9b9f85cf2fb0rpluem the support binaries to be built statically against
412da000890e343cb9cc01e9958f9b9f85cf2fb0rpluem libapr/libaprutil/etc and then enable it in binbuild.
412da000890e343cb9cc01e9958f9b9f85cf2fb0rpluemRELEASE NON-SHOWSTOPPERS BUT WOULD BE REAL NICE TO WRAP THESE UP:
412da000890e343cb9cc01e9958f9b9f85cf2fb0rpluem * If the parent process dies, should the remaining child processes
412da000890e343cb9cc01e9958f9b9f85cf2fb0rpluem "gracefully" self-terminate.
cbb903af5066589fe0e73f3ecf06abdc71e38effrpluem Message-ID: <3C58232C.FE91F19F@Golux.Com>
cbb903af5066589fe0e73f3ecf06abdc71e38effrpluem +1: Jim, Justin
cbb903af5066589fe0e73f3ecf06abdc71e38effrpluem * Make some modifications to the scoreboard creation routines,
dc681b5cee49eb24c3ae8d6f50c4c9c9c05faa52rpluem prefer anonymous shared memory, and allow a configuration
dc681b5cee49eb24c3ae8d6f50c4c9c9c05faa52rpluem directive to override the defaults.
dc681b5cee49eb24c3ae8d6f50c4c9c9c05faa52rpluem Message-ID: <20020130080804.C16977@clove.org>
2c3721611cc8d1ebd87b152e7d933a6dec23b3dcwrowe Status: Aaron volunteers
2c3721611cc8d1ebd87b152e7d933a6dec23b3dcwrowe * Get perchild to work on platforms other than Linux. This
2c3721611cc8d1ebd87b152e7d933a6dec23b3dcwrowe will require a portable mechanism to pass data and file/socket
2c3721611cc8d1ebd87b152e7d933a6dec23b3dcwrowe descriptors between vhost child groups. An API was proposed
2c3721611cc8d1ebd87b152e7d933a6dec23b3dcwrowe on dev@apr:
2c3721611cc8d1ebd87b152e7d933a6dec23b3dcwrowe Message-ID: <20020111115006.K1529@clove.org>
2c3721611cc8d1ebd87b152e7d933a6dec23b3dcwrowe Vote: Is a non-portable perchild going to hold up a GA release?
cbb903af5066589fe0e73f3ecf06abdc71e38effrpluem * Modify the worker MPM so that it doesn't need to create and
cbb903af5066589fe0e73f3ecf06abdc71e38effrpluem destroy a pool for each request--possibly by adopting a
cbb903af5066589fe0e73f3ecf06abdc71e38effrpluem leader/follower model in which each worker owns a persistent
cbb903af5066589fe0e73f3ecf06abdc71e38effrpluem ptrans pool (like the prefork MPM) and the workers take
cbb903af5066589fe0e73f3ecf06abdc71e38effrpluem turns acting as listeners...this approach might also help
cbb903af5066589fe0e73f3ecf06abdc71e38effrpluem reduce context-switching
cbb903af5066589fe0e73f3ecf06abdc71e38effrpluem * CGI single-byte reads
cbb903af5066589fe0e73f3ecf06abdc71e38effrpluem BrianP suggests that this is caused by the ap_scan_script_header_err()
cbb903af5066589fe0e73f3ecf06abdc71e38effrpluem routine, which will do single-byte reads until it finds the end
cbb903af5066589fe0e73f3ecf06abdc71e38effrpluem of the header, at which point it constructs a pipe-bucket (buffered)
cbb903af5066589fe0e73f3ecf06abdc71e38effrpluem to read from.
c0bcd91f17456a06ce290005aab0d1e360482ea5rpluem Proposed solution in:
c0bcd91f17456a06ce290005aab0d1e360482ea5rpluem Message-ID: <3C36ADAF.60601@cnet.com>
c0bcd91f17456a06ce290005aab0d1e360482ea5rpluem * Try to get libtool inter-library dependency code working on AIX.
941fcca87a4607a388e88cff3fd0cdefc29bb81cjerenkrantz Message-ID: <cm3n10lx555.fsf@rdu163-40-092.nc.rr.com>
d05d20b9ae48c6768f40277a76ed198d30aed06ajim Justin says: If we get it working on AIX, we can enable this
d05d20b9ae48c6768f40277a76ed198d30aed06ajim on all platforms and clean up our build system
f6316e220101e5d4d99d82aab75ca7d668870942rpluem * Handling of %2f in URIs. Currently both 1.3 and 2.0
f6316e220101e5d4d99d82aab75ca7d668870942rpluem completely disallow %2f in the request URI path (see
f6316e220101e5d4d99d82aab75ca7d668870942rpluem ap_unescape_url() in util.c). It's permitted and passed
f6316e220101e5d4d99d82aab75ca7d668870942rpluem through in the query string, however. Roy says the
c83edb2f6bafd5f7013cd6bad78753a1e0c5fbe6jim original reason for disallowing it, from five years ago,
c83edb2f6bafd5f7013cd6bad78753a1e0c5fbe6jim was to protect CGI scripts that applied PATH_INFO to
c83edb2f6bafd5f7013cd6bad78753a1e0c5fbe6jim a filesystem location and which might be tricked by
c83edb2f6bafd5f7013cd6bad78753a1e0c5fbe6jim ..%2f..%2f(...). We *should* allow path-info of the
c83edb2f6bafd5f7013cd6bad78753a1e0c5fbe6jim Since we've revamped a lot of our processing of path
c83edb2f6bafd5f7013cd6bad78753a1e0c5fbe6jim segments, it would be nice to allow this, or at least
3aeb30211790fef38a8297f990b7ad3b2c46ece9colm allow it conditionally with a directive.
b7ad88322f8ea46b7ddac349956eed358b6980f1rpluem * FreeBSD, threads, and worker MPM. All seems to work fine
b7ad88322f8ea46b7ddac349956eed358b6980f1rpluem if you only have one worker process with many threads. Add
b7ad88322f8ea46b7ddac349956eed358b6980f1rpluem a second worker process and the accept lock seems to be
b7ad88322f8ea46b7ddac349956eed358b6980f1rpluem lost. This might be an APR issue with how it deals with
b7ad88322f8ea46b7ddac349956eed358b6980f1rpluem the child_init hook (i.e. the fcntl lock needs to be resynced).
b7ad88322f8ea46b7ddac349956eed358b6980f1rpluem More examination and analysis is required.
b7ad88322f8ea46b7ddac349956eed358b6980f1rpluem Status: This has also been reported on Cygwin.
06eb85bfb9e99bcf9c38c07c2434850543354d94rpluem Message-ID: <3C2CC514.8EF3BED1@wapme-systems.de> (cygnus)
79d1d5666b3ceb59c959b09600aa2bce32324677rpluem Justin says: So, FreeBSD-CURRENT and Cywin have the same
79d1d5666b3ceb59c959b09600aa2bce32324677rpluem problem. Yum. If another platform has this
79d1d5666b3ceb59c959b09600aa2bce32324677rpluem with worker, this becomes a showstopper.
a6ea86151dd968120a12b48867d45947ef2bb9darpluem Aaron says: I spent some time disecting this and have come to
a6ea86151dd968120a12b48867d45947ef2bb9darpluem the conclusion that it is not a problem in the worker MPM
a6ea86151dd968120a12b48867d45947ef2bb9darpluem (or at least, it is not isolated to a problem in worker).
a6ea86151dd968120a12b48867d45947ef2bb9darpluem I'll list some of the problems I'm seeing in case someone
a17ca1093e7dc28c1a54cfd9741f65653f5b2b19jim else wants to pick up where I've left off:
a17ca1093e7dc28c1a54cfd9741f65653f5b2b19jim - Delivery of just about any signal to one of the child
a17ca1093e7dc28c1a54cfd9741f65653f5b2b19jim processes will send it into an infinite loop as well.
fa735cac4e86858f11c0de4f7cea50fa63eab87ecolm - Even though the parent is spinning out of control,
fa735cac4e86858f11c0de4f7cea50fa63eab87ecolm at first the child or children will appear to work
fa735cac4e86858f11c0de4f7cea50fa63eab87ecolm properly. At times it is possible to get it into a state,
dbb3b82abaa9c0ad199a0a3d6a7a505136137c61colm however, where a request will hang until another concurrent
dbb3b82abaa9c0ad199a0a3d6a7a505136137c61colm request "kicks" the first, at which point the second will
dbb3b82abaa9c0ad199a0a3d6a7a505136137c61colm hang. My theory is that this has to do with the
3d76cdd5855071bbfb95a27de7d77a8855d1e893niq pthread_cond_*() implementation in FreeBSD, but it's still
3d76cdd5855071bbfb95a27de7d77a8855d1e893niq possible that it is in APR.
3d76cdd5855071bbfb95a27de7d77a8855d1e893niq Justin adds: Oh, FreeBSD threads are implemented entirely with
3d76cdd5855071bbfb95a27de7d77a8855d1e893niq select()/poll()/longjmp(). Welcome to the nightmare.
3d76cdd5855071bbfb95a27de7d77a8855d1e893niq So, that means a ktrace output also has the thread
3d76cdd5855071bbfb95a27de7d77a8855d1e893niq scheduling internals in it (since it is all the same to
3d76cdd5855071bbfb95a27de7d77a8855d1e893niq the kernel). Which makes it hard to distinguish between
bf0dec31a67271a4a67a6d3a3a9b3cdb41278390rpluem our select() calls and their select() calls.
bf0dec31a67271a4a67a6d3a3a9b3cdb41278390rpluem *bangs head on wall repeatedly* But, some of the libc_r
bf0dec31a67271a4a67a6d3a3a9b3cdb41278390rpluem files have a DBG_MSG #define. This is moderately helpful
bf0dec31a67271a4a67a6d3a3a9b3cdb41278390rpluem when used with -DNO_DETACH. The kernel scheduler isn't
7612604db51e61f6ec4a6593c188e6d0640646abwrowe waking up the threads on a select(). Yum. And, I bet
7612604db51e61f6ec4a6593c188e6d0640646abwrowe those decrementing select calls have to do with the
7612604db51e61f6ec4a6593c188e6d0640646abwrowe scheduler. Time to brush up on our OS fundamentals.
7612604db51e61f6ec4a6593c188e6d0640646abwrowe * There is increasing demand from module writers for an API
7612604db51e61f6ec4a6593c188e6d0640646abwrowe that will allow them to control the server � la apachectl.
0a2c6381b34b977101327f412a4b1ddc8a361ae7rpluem Reasons include sole-function servers that need to die if
0a2c6381b34b977101327f412a4b1ddc8a361ae7rpluem an external dependency (e.g., a database) fails, et cetera.
0a2c6381b34b977101327f412a4b1ddc8a361ae7rpluem Perhaps something in the (ever more abused) scoreboard?
0a2c6381b34b977101327f412a4b1ddc8a361ae7rpluem rbb: I don't believe the scoreboard is the correct mechanism
8828f8c859fabac8c0080440392bc279877f9802rpluem for this. We already have a pipe that goes between parent
8828f8c859fabac8c0080440392bc279877f9802rpluem and child for graceful shutdown events, along with an API that
8828f8c859fabac8c0080440392bc279877f9802rpluem can be used to send a message down that pipe. In threaded MPMs,
0830215c2542165c027f092c18924f2885c98b49trawick it is easy enough to make that one pipe be used for graceful
0830215c2542165c027f092c18924f2885c98b49trawick and graceless events, and it is also easy to open that pipe
0830215c2542165c027f092c18924f2885c98b49trawick to both parent and child for writing. Then we just need to figure
0830215c2542165c027f092c18924f2885c98b49trawick out how to do graceless on non-threaded MPMs.
8f1c5a2aca731a7676bf74cdce67ec28557ec299trawick * revamp the input filter behavior, per discussions since
bc886ee8559b4feb30d32dc8a50ba2c013c3cc02trawick February (and especially at the hackathon last
8f1c5a2aca731a7676bf74cdce67ec28557ec299trawick April). Specifically, ap_get_brigade will return a brigade with
9505e1a9012d92bfd2b6228e675156f44a27aa04colm *up to* a specific number of bytes, or a "line" of data. The
9505e1a9012d92bfd2b6228e675156f44a27aa04colm read may be blocking or nonblocking. ap_getline() will be
3aeb30211790fef38a8297f990b7ad3b2c46ece9colm refactored into apr_brigade_getline(), and then DECHUNK can use
3aeb30211790fef38a8297f990b7ad3b2c46ece9colm f->next (ap_getline will always read "top of input stack"). Also
3aeb30211790fef38a8297f990b7ad3b2c46ece9colm fix the bug where request body content will end up closing the
3aeb30211790fef38a8297f990b7ad3b2c46ece9colm connection (buggering up persistent conns).
6eb3a057b5e4f046faba79104d7941552de755dacolm Status: Justin is working on this as fast as he can.
6eb3a057b5e4f046faba79104d7941552de755dacolm The core input filters, HTTP-related filters, mod_ssl, and
2e02c854b4531480fe3f7ab353154c99c5bd97e8jorton mod_proxy are switched to the new logic.
2e02c854b4531480fe3f7ab353154c99c5bd97e8jorton However, ap_getline() still needs to be refactored out. But,
2e02c854b4531480fe3f7ab353154c99c5bd97e8jorton there's a problem there: ap_getline() peeks ahead for MIME
2e02c854b4531480fe3f7ab353154c99c5bd97e8jorton continuation (first character on line is space or \t) and
2e02c854b4531480fe3f7ab353154c99c5bd97e8jorton stores unused data in core_request_config which violates the
c07be90520f577cdffc0ee59a81e7f2be4ab24c6martin abstraction. That's cheating. So, we may not be able to
2e02c854b4531480fe3f7ab353154c99c5bd97e8jorton implement this without setting some data aside (yuck!).
2e02c854b4531480fe3f7ab353154c99c5bd97e8jorton I believe this is OtherBill's main complaint with the current
2e02c854b4531480fe3f7ab353154c99c5bd97e8jorton AIUI (correct me if I'm wrong!), OtherBill believes we
2e02c854b4531480fe3f7ab353154c99c5bd97e8jorton should have a pushback option so that we can return unread
6a9a6a63bae82591ea75fc602d5260109b8fa02ccolm data - this would solve this case. However, my question to
70706a4d1ead0d72a443eac1cc446a0f1a4de069colm him is how do we handle stuff like mod_ssl - we can't "unread"
70706a4d1ead0d72a443eac1cc446a0f1a4de069colm data. So, do we have two brigades for each filter? An in
70706a4d1ead0d72a443eac1cc446a0f1a4de069colm brigade and a returned brigade? That seems messy. To
70706a4d1ead0d72a443eac1cc446a0f1a4de069colm everyone else, can we refactor ap_getline() without pushback
9cfe09f7e922f7bdf61bfe10a47e8f25145cdf92colm - socket bucket and core input filter changes. see end of
9cfe09f7e922f7bdf61bfe10a47e8f25145cdf92colm message ID (Feb 27): <20010227075326.S2297@lyra.org>
be2abc811ae62b872703a7b43235598d641f8d8dmjc - fix up ap_get_brigade() semantics, fix bug in DECHUNK /
2e02c854b4531480fe3f7ab353154c99c5bd97e8jorton ap_getline. many messages (plus their threads) (Apr/May):
2e02c854b4531480fe3f7ab353154c99c5bd97e8jorton Message-ID: <20010402101207.J27539@lyra.org>
2e02c854b4531480fe3f7ab353154c99c5bd97e8jorton Message-ID: <3AF7F921.D2EEC41A@algroup.co.uk>
2e02c854b4531480fe3f7ab353154c99c5bd97e8jorton Message-ID: <20010508190029.E18404@lyra.org>
2e02c854b4531480fe3f7ab353154c99c5bd97e8jorton - further work with combining/tweaking the builtin filters:
2e02c854b4531480fe3f7ab353154c99c5bd97e8jorton Message-ID: <20010509115445.D1374@lyra.org>
a09378e32e823d5619f766394fcdde1cfa16c5dajim - thoughts on filter modes:
a09378e32e823d5619f766394fcdde1cfa16c5dajim Message-ID: <021b01c14dee$09782af0$93c0b0d0@roweclan.net>
a09378e32e823d5619f766394fcdde1cfa16c5dajim - proposal for getline rewrite:
a09378e32e823d5619f766394fcdde1cfa16c5dajim Message-ID: <20011231104019.GD3904@ebuilt.com>
ce8e09ffeb8c696f7fc6c5dfa8814c52caaac59bcolm - patch for getline rewrite:
ce8e09ffeb8c696f7fc6c5dfa8814c52caaac59bcolm Message-ID: <20020122072605.GF28051@ebuilt.com>
ce8e09ffeb8c696f7fc6c5dfa8814c52caaac59bcolm - Change ap_get_brigade prototype:
5340bb8547e62ec8ede44b1b086b7da37ee64eefcolm Message-ID: <20020120115136.GI17601@ebuilt.com>
5340bb8547e62ec8ede44b1b086b7da37ee64eefcolm * Allow the DocumentRoot directive within <Location > scopes? This
f6e71f622ffde40c654288851ff65b8505cf7596colm allows the beloved (crusty) Alias /foo/ /somepath/foo/ followed
f6e71f622ffde40c654288851ff65b8505cf7596colm by a <Directory /somepath/foo> to become simply
f6e71f622ffde40c654288851ff65b8505cf7596colm <Location /foo/> DocumentRoot /somefile/foo (IMHO a bit more legible
f8ae06c7d0615d8b9441099794fe80a3f2915469colm and in-your-face.) DocumentRoot unset would be accepted [and would
f8ae06c7d0615d8b9441099794fe80a3f2915469colm not permit content to be served, only virtual resources such as
f8ae06c7d0615d8b9441099794fe80a3f2915469colm server-info or server-status.
6eb3a057b5e4f046faba79104d7941552de755dacolm This proposed change would _not_ depricate Alias.
6eb3a057b5e4f046faba79104d7941552de755dacolm * Win32: Rotatelogs sometimes is not terminated when Apache
6aa2272cc4af77e605ba2c4a4781f8567408b7e3pquerna goes down hard. FirstBill was looking at possibly tracking the
6eb3a057b5e4f046faba79104d7941552de755dacolm child's-child processes in the parent process.
be2abc811ae62b872703a7b43235598d641f8d8dmjc OtherBill asks, wasn't this fixed?
dface6840439ab84a36749f40e7773ebfdc2066ejorton OtherBill asks again, what is the status?
dface6840439ab84a36749f40e7773ebfdc2066ejorton * Win32: Add a simple hold console open patch (wait for close or
dface6840439ab84a36749f40e7773ebfdc2066ejorton the ESC key, with a nice message) if the server died a bad
1febae173a82bc2a71c3c0ba4105cf674000791bjim death (non-zero exit code) in console mode.
1febae173a82bc2a71c3c0ba4105cf674000791bjim Resolution: bring forward same ugly hacks from 1.3.13-.20
1febae173a82bc2a71c3c0ba4105cf674000791bjim * Port of mod_ssl to Apache 2.0:
2b409b0583ef844d6f85002160cbc6df95b9afa8jerenkrantz The current porting state is summarized in modules/ssl/README. The
2b409b0583ef844d6f85002160cbc6df95b9afa8jerenkrantz remaining work includes:
2b409b0583ef844d6f85002160cbc6df95b9afa8jerenkrantz (1) stablizing/optimizing the SSL filter logic
2b409b0583ef844d6f85002160cbc6df95b9afa8jerenkrantz (2) Enabling the various SSL caching mechanisms (shmcb, shmht)
454d060a0d7b13606978c536f632be71e311959djerenkrantz (3) Enabling SSL extentions
454d060a0d7b13606978c536f632be71e311959djerenkrantz (4) Trying to seperate the https filter logic from mod_ssl -
454d060a0d7b13606978c536f632be71e311959djerenkrantz This is to facilitate other modules that wish to use the https
454d060a0d7b13606978c536f632be71e311959djerenkrantz filter or the mod_ssl logic or both as required.
1392b9997000dceb6aca39da0a1e6405ccb6f0a6jerenkrantz Justin: mod_ssl filter logic is redone, so that should be fine.
1392b9997000dceb6aca39da0a1e6405ccb6f0a6jerenkrantz Madhu has submitted a patch for SSL caching - however, I
1392b9997000dceb6aca39da0a1e6405ccb6f0a6jerenkrantz am -0 on that patch as I *think* we could implement the
4b0a0ae3fd77f8fc8fd2a693be33ffe4d99cb94ajerenkrantz shared memory another way that is much cleaner (i.e.
4b0a0ae3fd77f8fc8fd2a693be33ffe4d99cb94ajerenkrantz treat shmem directly as a dbm via APR routines). Justin
4b0a0ae3fd77f8fc8fd2a693be33ffe4d99cb94ajerenkrantz also thinks that the https filter logic may be sufficiently
4b0a0ae3fd77f8fc8fd2a693be33ffe4d99cb94ajerenkrantz decoupled now, but isn't really sure.
4b0a0ae3fd77f8fc8fd2a693be33ffe4d99cb94ajerenkrantz * Performance & Debug: Eliminate most (and perhaps all) of the
ee508128c414648982d1cca7801f63b01a0a4f8aminfrin malloc/free calls in the bucket brigade code. Need some
ee508128c414648982d1cca7801f63b01a0a4f8aminfrin light weight memory management functions that allow freeing
ee508128c414648982d1cca7801f63b01a0a4f8aminfrin memory (putting it back into a memory pool) when it is no
ee508128c414648982d1cca7801f63b01a0a4f8aminfrin longer needed. Enabling simple debugging features like guard
f07ae092933846055c4ca046c5a65de83d79c9c6pquerna bands, double free detection, etc. would be cool but certainly
f07ae092933846055c4ca046c5a65de83d79c9c6pquerna not a hard requirement.
9b88ec397b4155f3ff79bf7c08de5f15d21765b3pquerna Status: Cliff started to implement this using SMS as has
9b88ec397b4155f3ff79bf7c08de5f15d21765b3pquerna been discussed at length for months, but since
9b88ec397b4155f3ff79bf7c08de5f15d21765b3pquerna SMS is not being used anywhere else in the server,
a42c13195a5c68d55f6b2bbf5b5e1e66ab45d5d4pquerna several people expressed the opinion that we should
a42c13195a5c68d55f6b2bbf5b5e1e66ab45d5d4pquerna get rid of it entirely, meaning that the buckets
a42c13195a5c68d55f6b2bbf5b5e1e66ab45d5d4pquerna need their own memory management (free list) functions.
2a90918318af50d5712ae7159e5625476332c184pquerna Cliff will implement that this weekend so we at least
2a90918318af50d5712ae7159e5625476332c184pquerna have something to look at/compare with.
ede0584e7ec53cc1ea832c5c50281431d9ecc576pquerna * Eliminate unnecessary creation of pipes in mod_cgid
ede0584e7ec53cc1ea832c5c50281431d9ecc576pquerna * the autoconf setup should be fixed to default to using the
ede0584e7ec53cc1ea832c5c50281431d9ecc576pquerna "Apache" layout from config.layout, and each variable settable
82d2a5debc5a6ed2118ac5916d9ba36ad0b5d78btrawick in a layout should be overridable on the command line. Plus,
42167da203d969a1402cf7ce09c14586c04af1dfjim what we do right now just doesn't seem to fully fit into how autoconf
42167da203d969a1402cf7ce09c14586c04af1dfjim works, eg. AC_PREFIX_DEFAULT issues.
42167da203d969a1402cf7ce09c14586c04af1dfjim Message-ID: <Pine.BSF.4.20.0104031557420.20876-100000@alive.znep.com>
e62b34b26eb8e82bf6f06ff99a08e304b48792ecpquerna * Combine log_child and piped_log_spawn. Clean up http_log.c.
e62b34b26eb8e82bf6f06ff99a08e304b48792ecpquerna Common logging API.
27420b7ba57bf78f7a85b19fb0155053a3aff80dwrowe * Document mod_file_cache.
27420b7ba57bf78f7a85b19fb0155053a3aff80dwrowe * Platforms that do not support fork (primarily Win32 and AS/400)
ceeb4b77f172b9d0125f1b4b323675d700b2c43bminfrin Architect start-up code that avoids initializing all the modules
ceeb4b77f172b9d0125f1b4b323675d700b2c43bminfrin in the parent process on platforms that do not support fork.
9152b0c4366af633c4b7982af01bd9a821f410d5pquerna * Win32: Migrate the MPM over to use APR thread/process calls. This
9152b0c4366af633c4b7982af01bd9a821f410d5pquerna would eliminate some code in the Win32 branch that essentially
9152b0c4366af633c4b7982af01bd9a821f410d5pquerna duplicates what is in APR.
87d944bf70927764edf8ef69e46d3b4b8fa09131pquerna * There are still a number of places in the code where we are
87d944bf70927764edf8ef69e46d3b4b8fa09131pquerna losing error status (i.e. throwing away the error returned by a
424c1a743525e9c5008e29b39a99363723409a1dtrawick system call and replacing it with a generic error code)
424c1a743525e9c5008e29b39a99363723409a1dtrawick * Mass vhosting version of suEXEC.
4375cabac8fc22b3717687ffdbce9bbdf095f255trawick * All DBMs suffer from confusion in support/dbmmanage (perl script) since
4375cabac8fc22b3717687ffdbce9bbdf095f255trawick the dbmmanage employs the first-matched dbm format. This is not
39dbd3f60b93f5e0fbf46d9ae237f6742e113442pquerna necessarily the library that Apache was built with. Aught to
39dbd3f60b93f5e0fbf46d9ae237f6742e113442pquerna rewrite dbmmanage upon installation to bin/ with the proper library
89211a3153be8b03353c3bfbca45fed67cb80f0bpquerna for predictable mod_auth_dbm administration.
89211a3153be8b03353c3bfbca45fed67cb80f0bpquerna Questions; htdbm exists, time to kill dbmmanage, or does it remain
89211a3153be8b03353c3bfbca45fed67cb80f0bpquerna useful as a perl dbm management example? If we keep it,
89211a3153be8b03353c3bfbca45fed67cb80f0bpquerna do we address the issue above?
89211a3153be8b03353c3bfbca45fed67cb80f0bpquerna * use apu_dbm in mod_auth_dbm
89211a3153be8b03353c3bfbca45fed67cb80f0bpquerna Status: Greg +1 (low-priority volunteer)
89211a3153be8b03353c3bfbca45fed67cb80f0bpquerna Justin says: "Seems like this is already there, so should we just
a14ccf0f7e9b44c6848334823542a1799577f669pquerna remove the other DBM code in that file? If you want
a14ccf0f7e9b44c6848334823542a1799577f669pquerna to use gdbm, or dbm, etc, you should tell apr-util."
a14ccf0f7e9b44c6848334823542a1799577f669pquerna Will says: "bs - I may choose the fastest - most efficient native
43c24b7301a9df7014ce54c0bc55ac4754cf5b0dpquerna dbm implementation, for shared proc caches, ssl session
43c24b7301a9df7014ce54c0bc55ac4754cf5b0dpquerna caching, etc, but that has nothing to do with maintaining
a2e37e48efb113e8e4f9d9ca9c8286aaac6c936cpquerna a userlist via dbm, which has to remain readable between
a2e37e48efb113e8e4f9d9ca9c8286aaac6c936cpquerna builds/machines, etc. The use-multiple database schema
a2e37e48efb113e8e4f9d9ca9c8286aaac6c936cpquerna for apr-util would let us do this with just apr, though."
007ba11dc68651df7f872f35947ae82a595d9e02pquerna Ian says: "multi-dbm is in, but it still has ndbm support hardcoded
007ba11dc68651df7f872f35947ae82a595d9e02pquerna is this still required? isn't ndbm supported via gdbm?"
44ca834b970b454b844efb96f219bdf49fee71e5trawick * Integrate mod_dav.
44ca834b970b454b844efb96f219bdf49fee71e5trawick Some additional items remaining:
eb8430fd3bc941c0b3ba8bba3884b7fc464bf458pquerna - case_preserved_filename stuff
eb8430fd3bc941c0b3ba8bba3884b7fc464bf458pquerna (use the new canonical name stuff?)
eb8430fd3bc941c0b3ba8bba3884b7fc464bf458pquerna - find a new home for ap_text(_header)
fde88bb682426885c679198ee130d2d5a29e8c0fbnicholes - is it possible to remove the DAV: namespace stuff from util_xml?
fde88bb682426885c679198ee130d2d5a29e8c0fbnicholes * ap_core_translate() and its use by mod_mmap_static and mod_file_cache
fde88bb682426885c679198ee130d2d5a29e8c0fbnicholes are a bit wonky. The function should probably be exposed as a utility
fde88bb682426885c679198ee130d2d5a29e8c0fbnicholes function (such as ap_translate_url2fs() or ap_validate_fs_url() or
fde88bb682426885c679198ee130d2d5a29e8c0fbnicholes something). Another approach would be a new hook phase after
fde88bb682426885c679198ee130d2d5a29e8c0fbnicholes "translate" which would allow the module to munge what the
fde88bb682426885c679198ee130d2d5a29e8c0fbnicholes translation has decided to do.
60e385aa992e11a6cb0504e8d4fc35186e8d848bpquerna Status: Greg +1 (volunteers), Ryan +1
60e385aa992e11a6cb0504e8d4fc35186e8d848bpquerna * Explore use of a post-config hook for the code in http_main.c which
d66ac514cc15e99228d72c56c6c3daf25da8d360niq calls ap_fixup_virutal_hosts(), ap_fini_vhost_config(), and
d66ac514cc15e99228d72c56c6c3daf25da8d360niq ap_sort_hooks() [to reduce the logic in main()]
d66ac514cc15e99228d72c56c6c3daf25da8d360niq * read the config tree just once, and process N times (as necessary)
e23b77006a8b079c0ad52e42ba2029e759455b8fjorton * (possibly) use UUIDs in mod_unique_id and/or mod_usertrack
e23b77006a8b079c0ad52e42ba2029e759455b8fjorton * (possibly) port the bug fix for PR 6942 (segv when LoadModule is put
22c347c08cd77d5e2c32536b467b389fd33d631fpquerna into a VirtualHost container) to 2.0.
b6e2b36d6972a0e243368eb567e91d511fb40b1bbnicholes * shift stuff to mod_core.h
b6e2b36d6972a0e243368eb567e91d511fb40b1bbnicholes * callers of ap_run_create_request() should check the return value
b6e2b36d6972a0e243368eb567e91d511fb40b1bbnicholes for failure (Doug volunteers)
ad4a5834dfdf9488af4ac3238ea4d8b18dce49acpquerna * Win32: Get Apache working on Windows 95/98. The following work
e5882a36d7756850cc829f5f2286120b877458b1pquerna (at least) needs to be done:
e5882a36d7756850cc829f5f2286120b877458b1pquerna - winnt MPM: Fix 95/98 code paths in the winnt MPM. There is some NT
8b7594c66e764f5cd3506b6f2459497ab65a8b03pquerna specific code that is still not in NT only code paths
8b7594c66e764f5cd3506b6f2459497ab65a8b03pquerna - IOL binds to APR sendfile, implemented with TransmitFile, which
8b7594c66e764f5cd3506b6f2459497ab65a8b03pquerna is not available on 95/98.
8c46f9f81ae6357dc97560d6d85154f19cf251aejim - Document warning that OSR2 is required (for Crypt functions, in
8c46f9f81ae6357dc97560d6d85154f19cf251aejim rand.c, at least.) This could be resolved with an SSL library, or
20f1b1a67eef5ab0f3295608c89964a7dca4fdd1pquerna randomization in APR itself.
20f1b1a67eef5ab0f3295608c89964a7dca4fdd1pquerna - Bring the Win9xConHook.dll from 1.3 into 2.0 (no sense till it
20f1b1a67eef5ab0f3295608c89964a7dca4fdd1pquerna actually works) and add in a splash of Win9x service code.
599c5abcc7fec60611956f56b732eca033c287d2pquerna * In order to use a DSO version of mod_ssl we have to link with
599c5abcc7fec60611956f56b732eca033c287d2pquerna -lssl and -lcrypto. A workaround is in place right now where the
599c5abcc7fec60611956f56b732eca033c287d2pquerna entire EXTRA_LIBS macro is being appended to the objects list, but
e76fdcdfb8994ad70776526f50fa013b3e9a6033bnicholes this is a hack. We should either revamp the APACHE_CHECK_SSL_TOOLKIT
e76fdcdfb8994ad70776526f50fa013b3e9a6033bnicholes autoconf function or come up with some other autoconf checks to
e76fdcdfb8994ad70776526f50fa013b3e9a6033bnicholes search for libssl and libcrypto and properly add them to mod_ssl's
eb9b491d7b262dad572ec2f1f75eea592283f81apquerna link flags.
1ef90d1763e489969ec7a01c77bcb158ece5ae83jorton * Make the worker MPM the default MPM for threaded Unix boxes.
c5cb4c9e7c8696907ddebab2a0037717c710b0f6jorton +1: Justin, Jeff, Ian
6d00a5e2c08ddbff9614ecc17a1c085462811f89slive -0: Aaron (premature decision, needs more discussion), Lars
6d00a5e2c08ddbff9614ecc17a1c085462811f89slive -0: Cliff (I think the default config should be the safest possible)
6d00a5e2c08ddbff9614ecc17a1c085462811f89slive * Fix the worker MPM to use POD to kill child processes instead
6d00a5e2c08ddbff9614ecc17a1c085462811f89slive of ap_os_killpg, regardless of how they should die. (Ryan Bloom)
6d00a5e2c08ddbff9614ecc17a1c085462811f89slivePRs that have been suspended forever waiting for someone to
6d00a5e2c08ddbff9614ecc17a1c085462811f89sliveput them into 'the next release':
6d00a5e2c08ddbff9614ecc17a1c085462811f89slive * PR#76: general
6d00a5e2c08ddbff9614ecc17a1c085462811f89slive missing call to "setlocale();"
6d00a5e2c08ddbff9614ecc17a1c085462811f89slive * PR#78: mod_include
6d00a5e2c08ddbff9614ecc17a1c085462811f89slive Additional status for XBitHack directive
4f6e78091b3f45a5782389ae25b62516a7a0c842bnicholes * PR#362: mod_proxy
4f6e78091b3f45a5782389ae25b62516a7a0c842bnicholes Mod_proxy doesn't allow change of error pages
7d15331eeb5429d7148d13d6fd914a641bf1c000pquerna * PR#370: mod_env
7d15331eeb5429d7148d13d6fd914a641bf1c000pquerna Modified PATH environemnt variable is not passed, instead
7d15331eeb5429d7148d13d6fd914a641bf1c000pquerna system's is used
e9be3aacfd6c0a1208e6c91a133be92ed0f94fe1bnicholes * PR#440: mod_proxy
e9be3aacfd6c0a1208e6c91a133be92ed0f94fe1bnicholes Proxy doesn't deliver documents if not connected
6e5f5644328bf50c3aa295d0ab20903369010829gregames * PR#534: mod_proxy
6e5f5644328bf50c3aa295d0ab20903369010829gregames proxy converts ~name to %7Ename when name starts with a dot (.)
e978097e0bf2ae161b6f9dde40eaf089bf046c89ake * PR#537: mod_access
e978097e0bf2ae161b6f9dde40eaf089bf046c89ake mod_access syntax allows hosts that should be restricted
275419d6395e6f072962fb701b89accaff1f3690jerenkrantz * PR#557: mod_auth-any
e777da9fa7ff3138fead7860b53ef00e67a40e26jerenkrantz ~UserHome directories are not honored in absolute pathname
e777da9fa7ff3138fead7860b53ef00e67a40e26jerenkrantz requests (.htaccess)
8bdea88407c848c1c2693655e2f8b23abde12307bnicholes * PR#612: mod_proxy
8bdea88407c848c1c2693655e2f8b23abde12307bnicholes Proxy FTP Authentication Fails
8e7fb6968047a527d1ccde25ad2aed20da5150ddjerenkrantz * PR#623: mod_include
8e7fb6968047a527d1ccde25ad2aed20da5150ddjerenkrantz A smarter "Last Modified" value for SSI documents (see PR number 600)
5a9fa4d75086e942f6e850e1a2e96c4c27a845d0jerenkrantz * PR#628: config
5a9fa4d75086e942f6e850e1a2e96c4c27a845d0jerenkrantz Request of "Options SymLinksIfGroupMatch"
5a9fa4d75086e942f6e850e1a2e96c4c27a845d0jerenkrantz * PR#700: mod_proxy
36c8049de63c446926139936c3d195330a0539cetrawick Proxy doesn't do links right for OpenVMS files through ftp:
36c8049de63c446926139936c3d195330a0539cetrawick * PR#759: mod_imap
36c8049de63c446926139936c3d195330a0539cetrawick imap should read <MAP><AREA>*</MAP> too!
9174efb969475801d0dc88eee35aae40c748d450nd * PR#793: general
0206c121a68a63559b2e843288e81bcf16093e46jerenkrantz RLimitCPU and RLimitMEM don't apply to all children like they should
628cf3411e8a6d09e27b3666312e43832dda93f3jorton * PR#921: suexec
628cf3411e8a6d09e27b3666312e43832dda93f3jorton Uses cwd before filling it in, doesn't use syslog
f0d89a5a23222e031db8113478645f28688fa748jorton * PR#922: config
f0d89a5a23222e031db8113478645f28688fa748jorton it is useful to allow specifiction that root-owned symlinks
f0d89a5a23222e031db8113478645f28688fa748jorton should always be followed
f0d89a5a23222e031db8113478645f28688fa748jorton * PR#980: mod_proxy
f0d89a5a23222e031db8113478645f28688fa748jorton Controlling Access to Remote Proxies would be nice...
65d743d7fbb53143636ee2dec8fe8d8a1a581a6bjerenkrantz * PR#994: mod_proxy
65d743d7fbb53143636ee2dec8fe8d8a1a581a6bjerenkrantz Adding authentication "on the fly" through the proxy module
a0fd132e01ab69f1c48e3d6a6791447cb6d65e51jerenkrantz * PR#1004: apache-api
a0fd132e01ab69f1c48e3d6a6791447cb6d65e51jerenkrantz request_config field in request_rec is moderately bogus
a6bb6f2fb734e488a9b6335fabea3431f9dcf253jerenkrantz * PR#1028: other
3ca80e703a960eca0760c23636b7fe502a8f0342bnicholes DoS attacks involving memory consumption
3ca80e703a960eca0760c23636b7fe502a8f0342bnicholes * PR#1050: mod_log-any
3ca80e703a960eca0760c23636b7fe502a8f0342bnicholes Logging of virtual server to error_log as well
5300d4a4fabe3b594da950e4b9ab0f90e076546ejerenkrantz * PR#1085: mod_proxy
b1306729566b49fb30aed5c46adaf07a637115afjerenkrantz ProxyRemote make a dead cycle.
aa50e6503e0c7fbdc7d9bfab7d3a5fa640c2c8c1wrowe * PR#1117: mod_auth-any
aa50e6503e0c7fbdc7d9bfab7d3a5fa640c2c8c1wrowe Using NIS passwd.byname dbm files with AuthDBMUserFile
5c6cb72bfeee541644cea8177aefce1157571d3bjerenkrantz * PR#1120: suexec
5c6cb72bfeee541644cea8177aefce1157571d3bjerenkrantz suexec does not parse arguments to #exec cmd
db8551deeb08fa799e7f27e8b748a9397f747bdcjorton * PR#1145: mod_include
db8551deeb08fa799e7f27e8b748a9397f747bdcjorton Allow for Last-Modified: without resorting to XBitHack
9e3209bc06ddf32f23e4b254faa45914bc323cc9jim * PR#1158: apache-api
45ed846f4ed90f05c084f6a33d688e642be4e623jerenkrantz improvements to child spawning API
54d22ed1c429b903b029bbd62621f11a9e286137minfrin * PR#1166: mod_proxy
54d22ed1c429b903b029bbd62621f11a9e286137minfrin ``nph-'' not honored (no buffering) for ProxyRemote mapping
54d22ed1c429b903b029bbd62621f11a9e286137minfrin * PR#1176: mod_cgi
54d22ed1c429b903b029bbd62621f11a9e286137minfrin Apache cannot handle continuation line in headers
92e8e44864d94866eefcbfde0a53fa3d12855149jerenkrantz * PR#1191: general
92e8e44864d94866eefcbfde0a53fa3d12855149jerenkrantz setlogin() is not called, causing problems with e.g. identd
06106b6b63b112a09de1b66fa29596035c0176ffthommay * PR#1204: general
06106b6b63b112a09de1b66fa29596035c0176ffthommay regerror() exists, use it
e335319a08e12eb7daff9afa80e985dc53f652b8jorton * PR#1233: apache-api
e335319a08e12eb7daff9afa80e985dc53f652b8jorton there is no way to keep per-connection per-module state
c8794ec1e4c474ae101ce3835080f638136e7860erikabele * PR#1263: mod_autoexec
c8794ec1e4c474ae101ce3835080f638136e7860erikabele Add frame-safe anchor attribute to mod_autoindex links
bb07ee33bce1a448bcc60ca43720b1ab1c413f87minfrin * PR#1268: suexec
bb07ee33bce1a448bcc60ca43720b1ab1c413f87minfrin CGI scripts running as Apache user: security (suexec etc.)
7b6ba9c468f26bdb3492d5e8cb79628a3b04e8c8wrowe * PR#1285: suexec
7b6ba9c468f26bdb3492d5e8cb79628a3b04e8c8wrowe Error messages could be easier to spot in cgi.log file for suexec.c
77edcaaccc089335938f3844b752e1044f7eb278trawick * PR#1287: mod_access
77edcaaccc089335938f3844b752e1044f7eb278trawick add allow,deny/deny,allow warning to mod_access
59b1b6c3fd51c83c3bb9f02a8f08751335f9fb1dminfrin * PR#1290: mod_proxy
59b1b6c3fd51c83c3bb9f02a8f08751335f9fb1dminfrin Need to know "hit-rate" on proxy cache
9f1a88897168c3f1e5009acb585daf01e38a0299jim * PR#1358: mod_log-any
9f1a88897168c3f1e5009acb585daf01e38a0299jim Selective url-encode of log fields (or maybe a pseudo
9f1a88897168c3f1e5009acb585daf01e38a0299jim log_rewrite module?)
616a448c1fca1648622707df5a1aae7316bb3d5djim * PR#1383: mod_headers
616a448c1fca1648622707df5a1aae7316bb3d5djim I make mod_headers to modify request headers as well as
616a448c1fca1648622707df5a1aae7316bb3d5djim response ones.
5a8bb5948d2a258145174320587706de3219d8a3pquerna * PR#1532: mod_proxy
5a8bb5948d2a258145174320587706de3219d8a3pquerna Proxy transfer logging
0cba3a63e59bfa77f67955cb4e034264ed6c5523jerenkrantz * PR#1547: mod_proxy
e5abee85fe751fc27c5d4fc9a8ebe3b80f3d6603trawick No HTTP_X_FORWARDED_FOR set...
ca3a25d12c5fe0926aa97550be39f0d88d0decb1pquerna * PR#1567: mod_proxy
ca3a25d12c5fe0926aa97550be39f0d88d0decb1pquerna ProxyRemote proxy requests fail authentication by firewall
74def8815c725f8128a4e76ab1f5704df80b024ajerenkrantz * PR#1582: mod_rewrite
74def8815c725f8128a4e76ab1f5704df80b024ajerenkrantz mod_rewrite forms REQUEST_URI different than mod_cgi does
8f868f15482c7406db01216b6e4778ddabb26898trawick * PR#1677: mod_headers
8f868f15482c7406db01216b6e4778ddabb26898trawick mod_headers should allow mod_log_config-style formats in
8f868f15482c7406db01216b6e4778ddabb26898trawick header values
f902601ea431a9b56106e0f5f641dd5fd7efbc30jorton * PR#1702: mod_proxy
f902601ea431a9b56106e0f5f641dd5fd7efbc30jorton mod_proxy to support persistent conns?
1e1e4f4f810b99732f06fc05141f42ca1965a9edpquerna * PR#1803: mod_include
1e1e4f4f810b99732f06fc05141f42ca1965a9edpquerna patches to mod_include to allow for file tests
5f80956ca9d98d5482f38eef0c465df0923d7781jerenkrantz * PR#1809: mod_auth-any
78fcc425fc9fc58202a1693fe40dd0ce75c031ffjorton Suggestion for improving authentication modules and core source
78fcc425fc9fc58202a1693fe40dd0ce75c031ffjorton code, problem with 401 and ErrorDocument
d2ffb32434f79782ff7a364ffa31064698c5c645jorton * PR#1878: mod_proxy
d2ffb32434f79782ff7a364ffa31064698c5c645jorton listing of proxy cache content
01847067cfc639c14e1aa77d3b3e98f239447a12jerenkrantz * PR#1905: suexec
01847067cfc639c14e1aa77d3b3e98f239447a12jerenkrantz Allow modules to set user:group for execution.
9fe74ffcdea85800f04a7222f716f78ae60cce51jerenkrantz * PR#2024: apache-api
9fe74ffcdea85800f04a7222f716f78ae60cce51jerenkrantz adding auth_why to conn_rec
dae3cb64cc6681b5f6b0fd12e7f8f6296ffaa19abnicholes * PR#2073: mod_log-any
d2ffb32434f79782ff7a364ffa31064698c5c645jorton pipelined connections are not logged correctly
d2ffb32434f79782ff7a364ffa31064698c5c645jorton * PR#2074: mod_rewrite
81540a0eb1da2c687e22de3367d8ded55e17e330pquerna mod_rewrite doesn't pass Proxy Throughput on internal subrequests
ce66c67eba79a20118d8664b65b0c9eeec1bebdabnicholes * PR#2113: config
ce66c67eba79a20118d8664b65b0c9eeec1bebdabnicholes HTTP Server Rebuild Line Needs Changing for the better
6cfc2ed5a8e633c5a40fec65775868d53952d992trawick * PR#2138: mod_status
6cfc2ed5a8e633c5a40fec65775868d53952d992trawick mod_status always displays 256 possible connection slots
6cfc2ed5a8e633c5a40fec65775868d53952d992trawick * PR#2221: documentation
6cfc2ed5a8e633c5a40fec65775868d53952d992trawick Make online documentation search link back to my installation
5159c40648868a58745aa11981f706948ff0f0d1pquerna * PR#2284: general
c178b761acd6bffa199c2fd28c4469492b989699nd Can not POST to ErrorDocument - Apache/1.3b6
9c6bbd67082b5a47fb17cfa5b61b8a7e1fb01875pquerna * PR#2314: mod_proxy
9c6bbd67082b5a47fb17cfa5b61b8a7e1fb01875pquerna patterns in ProxyRemote
572f5b8a84bb399e51b02a562776f4aec119aa95pquerna * PR#2343: mod_status
572f5b8a84bb399e51b02a562776f4aec119aa95pquerna Status module averages are for entire uptime
1b03ca18c41f51a25dcf1a623a8f558bd779e0a4jerenkrantz * PR#2360: suexec
1b03ca18c41f51a25dcf1a623a8f558bd779e0a4jerenkrantz suexec for general access of user content?
1b03ca18c41f51a25dcf1a623a8f558bd779e0a4jerenkrantz * PR#2396: general
1b03ca18c41f51a25dcf1a623a8f558bd779e0a4jerenkrantz Proposal for TimeZone directive
1b03ca18c41f51a25dcf1a623a8f558bd779e0a4jerenkrantz * PR#2415: mod_info
db443e0132f14dac789ab97ec23ce124360d74c2nd /server-info doesn't check for the virtual host to list the info
4ac7a7c09ac5732b09f8bf28873f7e9efcab34d5bnicholes * PR#2421: config
4ac7a7c09ac5732b09f8bf28873f7e9efcab34d5bnicholes problem specifying ndbm library for build ?with autoconfigure
4ac7a7c09ac5732b09f8bf28873f7e9efcab34d5bnicholes * PR#2431: general
4ac7a7c09ac5732b09f8bf28873f7e9efcab34d5bnicholes A small addition to rotatelogs.c to improve program functionality.
4ac7a7c09ac5732b09f8bf28873f7e9efcab34d5bnicholes * PR#2446: config
ce8490f3812311582d1deee96b012c377311b317minfrin AllowOverride FileInfo is too coarse
38f6ebaca968b7b23c25c0b30d0be1c7aad7412bjorton * PR#2460: mod_cgi
38f6ebaca968b7b23c25c0b30d0be1c7aad7412bjorton TimeOut applies to output of CGI scripts
cc7d8b55b16eee88be925a090473ca94b0a6e770jorton * PR#2512: mod_access
cc7d8b55b16eee88be925a090473ca94b0a6e770jorton <IfDenied> directive wanted
2555a6b5da21d61804f47084d8fcc98eb4acbc42wrowe * PR#2573: suexec
2555a6b5da21d61804f47084d8fcc98eb4acbc42wrowe CGI's for general use still have to be run as another user
2555a6b5da21d61804f47084d8fcc98eb4acbc42wrowe with suExec
2555a6b5da21d61804f47084d8fcc98eb4acbc42wrowe * PR#2648: general
2555a6b5da21d61804f47084d8fcc98eb4acbc42wrowe Cache file names in Proxy module
eb6e3a3e784021abc1e28247158db3a8b07186b9minfrin * PR#2760: config
eb6e3a3e784021abc1e28247158db3a8b07186b9minfrin [PATCH] User/Group for <Directory> and <Location> i.e. not only
eb6e3a3e784021abc1e28247158db3a8b07186b9minfrin in global and <Virtual>.
f5098e2eb6ed9dfd7052f25761bb5e97779b8460minfrin * PR#2763: general
f5098e2eb6ed9dfd7052f25761bb5e97779b8460minfrin mailto tags and bundling bug report script
c79b5caf7c5462ef707363c69cc63ea2606310abtrawick * PR#2785: os-aix
c79b5caf7c5462ef707363c69cc63ea2606310abtrawick Support for System Resource Controller
1551c2366551d7a361bedbf7ae54f1abd74ef822minfrin * PR#2793: protocol
1551c2366551d7a361bedbf7ae54f1abd74ef822minfrin When will Apache support P3P? Any Plans?
f1826b4cf1fbeb049c569761cc95b04e3496efd6bnicholes * PR#2873: config
103a93c625bcde1a6a7a5155b64dcda36f612180pquerna * PR#2889: general
103a93c625bcde1a6a7a5155b64dcda36f612180pquerna Inclusion of RPM spec file in CVS/distributions
87a528a7622973988232079ba02763748c9c7071pquerna * PR#2906: general
1a986bbad9314beb8739401cac822e87bb04bbfepquerna Propose that Apache recommend $UNIQUE_ID for all "session id"
1a986bbad9314beb8739401cac822e87bb04bbfepquerna * PR#2907: config
752ec04fdfa6e936047811b9e506cf829d7a69f7pquerna suggestion: power up your Include directive :)
752ec04fdfa6e936047811b9e506cf829d7a69f7pquerna * PR#3018: general
cf82d17373559b6163dbd6c844c1af82cab56883niq cannot limit some HTTP methods
cf82d17373559b6163dbd6c844c1af82cab56883niq * PR#3143: apache-api
c4e6f5320f87eff25838c21c107a24b688d539d0niq No module specific data hook for per-connection data
c4e6f5320f87eff25838c21c107a24b688d539d0niq * PR#3191: mod_negotiation
c4e6f5320f87eff25838c21c107a24b688d539d0niq no way to set global quality-of-source (qs) coneg values
c4e6f5320f87eff25838c21c107a24b688d539d0niq with multiviews
174a8f105b1c11aa50a05993ea8ba2d117d3df6end * PR#3568: mod_proxy
174a8f105b1c11aa50a05993ea8ba2d117d3df6end Accessing URL through proxy server corrupts data.
3ad0f071604de33c0b3733b649cd0445ba00f54dake * PR#3605: mod_proxy
3ad0f071604de33c0b3733b649cd0445ba00f54dake Some anonymous FTP URLs ask for authentication
9174efb969475801d0dc88eee35aae40c748d450nd * PR#3677: general
c2549f0b237ac86f3623a601a766969d805dbc2and New ErrorDocumentMatch directive
f84d3d83a741c21154d42e0ebdec9b9b37efeedcjorton * PR#4241: config
f84d3d83a741c21154d42e0ebdec9b9b37efeedcjorton Need to be able to override shebang line to make CGI scripts
7efe7de73c89c26518714a504359244d03cfbbc5jorton more portable.
d60027cde3289d3ef35bc9d815882975eb60e6abjorton * PR#4244: config
d60027cde3289d3ef35bc9d815882975eb60e6abjorton "Files" and "FilesMatch" regexp does not recognize bang as
504498fd20729dcc29dfbfb78e64a8521c6a767ajerenkrantz negation operator
d406be0d6a4f9e61bbdfa6c8cde2e09338dd650atrawick * PR#4448: mod_log-any
d406be0d6a4f9e61bbdfa6c8cde2e09338dd650atrawick Please allow CGI env variables (QUERY_STRING, ...) to be logged
9174efb969475801d0dc88eee35aae40c748d450nd * PR#4459: mod_include
c1ba3b0ac00c4fee2f4b950dfdb167b39611b661nd Suggestion for better handling of Last-modified headers
b95a84193578c904426cef6dda84f7118a400a16jim * PR#4490: mod_cgi
b95a84193578c904426cef6dda84f7118a400a16jim mod_cgi prevents handling of OPTIONS requests
03a3ed87983471816561562f957390ed935d7b3bnd * PR#5713: os-windows
9174efb969475801d0dc88eee35aae40c748d450nd [PATCH] install as win32 service with domain account
c533ecac2227dc228070e686fb14dc6860f497f8nd Status: Cannot accept password-as-arg, we should prompt the
56cefde0af6b8db6fda0f1d95d8cdca54f397cd0nd user when -k install/-k config with a user argument.
56cefde0af6b8db6fda0f1d95d8cdca54f397cd0nd * PR#5993: general
41369ed0bc7f2db6272278c27025f6aabf97fe63nd AllowOverride should have a 'CheckNone' and 'AllowNone' argument
9174efb969475801d0dc88eee35aae40c748d450nd instead of only 'None'
496f8f3966319d43455675630a849bae019d2a32ndOther bugs that need fixing:
9e8c2603790f490398a0fabf97866b6815748a54ianh * MaxRequestsPerChild measures connections, not requests.
9e8c2603790f490398a0fabf97866b6815748a54ianh Until someone has a better way, we'll probably just rename it
9e8c2603790f490398a0fabf97866b6815748a54ianh "MaxConnectionsPerChild".
9174efb969475801d0dc88eee35aae40c748d450nd * Regex containers don't work in an intutive way
a21b3b9d8ebb12fd51fa1d17e44d5644a35a9a5fnd Status: No one has come up with an efficient way to fix this
1fbcf7a8e7b18be1b0e77b7bc38b7b71f8dfb052nd behavior. Dean has suggested getting rid of regex containers
9174efb969475801d0dc88eee35aae40c748d450nd completely.
1fbcf7a8e7b18be1b0e77b7bc38b7b71f8dfb052nd OtherBill suggests: We at least seem to agree on eliminating
f769c33501f474aed3e0f7c769477c8c4f478783geoff the <Container ~ foo> forms, and using only
f769c33501f474aed3e0f7c769477c8c4f478783geoff <ContainerMatch foo> semantics.
f769c33501f474aed3e0f7c769477c8c4f478783geoff * SIGSEGV on Linux (glibc 2.1.2) isn't caught properly by a
d7c7669331357296719d67d1963d40d713ed455atrawick sigwaiting thread. We need to work around this, perhaps unless
d7c7669331357296719d67d1963d40d713ed455atrawick there is hope soon for a fixed glibc.
d7c7669331357296719d67d1963d40d713ed455atrawick * orig_ct in the byterange/multipart handling may not be
92d95be777d4365eb79444a7a558355a7a92081ajorton needed. Apache 1.3 just never stashed "multipart" into
92d95be777d4365eb79444a7a558355a7a92081ajorton r->content_type. We should probably follow suit since the
92d95be777d4365eb79444a7a558355a7a92081ajorton byterange stuff doesn't want the rest of the code to see the
92d95be777d4365eb79444a7a558355a7a92081ajorton multipart content-type; the other code should still think it is
af8dee354a287249dd9f3f77bbe850108e5afe43trawick dealing with the <orig_ct> stuff.
af8dee354a287249dd9f3f77bbe850108e5afe43trawick Status: Greg volunteers to investigate (esp. since he was most
af8dee354a287249dd9f3f77bbe850108e5afe43trawick likely the one to break it :-)
af8dee354a287249dd9f3f77bbe850108e5afe43trawickOther features that need writing:
f78505c7d260473bf11002f5028186f27d0ed8a0geoff * Finish infrastructure in core for async MPMs
f78505c7d260473bf11002f5028186f27d0ed8a0geoff Status: post 2.0
f78505c7d260473bf11002f5028186f27d0ed8a0geoff * TODO in source -- just do an egrep on "TODO" and see what's there
f78505c7d260473bf11002f5028186f27d0ed8a0geoffAvailable Patches:
ea981941b1e5ecfb1a27b9e9aff32b9c98640ed9nd * Jon Travis's <jtravis@covalent.net> patch to deal with thread-safe
ea981941b1e5ecfb1a27b9e9aff32b9c98640ed9nd issues with inet_ntoa. See message <20001201163220.A12827@covalent.net>
ea981941b1e5ecfb1a27b9e9aff32b9c98640ed9nd Status: This is being set aside until the IPv6 work is finished
d57b793ddd42a997c2cb3bd389bbcf1161226fd8trawick so that we know exactly what is required.
d57b793ddd42a997c2cb3bd389bbcf1161226fd8trawick * Martin Sojka <msojka@gmx.de>'s patch to add error reporting for failed
d11bf78168abbb07d8ed9f54d7ea0953d46aad83nd htpasswd actions due to a full /tmp volume (other programs may have
d11bf78168abbb07d8ed9f54d7ea0953d46aad83nd similar problems?)
bf28a00d8218aa76a56eddabca308e8d0e691626nd * Mike Abbott's <mja@trudge.engr.sgi.com> patches to improve
bf28a00d8218aa76a56eddabca308e8d0e691626nd performance
9174efb969475801d0dc88eee35aae40c748d450nd Status: These were written for 1.3, and are awaiting a port to
9174efb969475801d0dc88eee35aae40c748d450nd * Jim Winstead's <jimw@trainedmonkey.com> patch to add CookieDomain and
70f665a8945f3bdfe8bbeaa063de2aa722fc9b29nd other small mod_usertrack features
c08f5a0cf2b8903255186ba998a4481dfac9c796nd * Dan Rench's <drench@xnet.com> patch to add allow the errmsg and timefmt
9174efb969475801d0dc88eee35aae40c748d450nd of SSI's to be modified in the config file. Patch is available in
9174efb969475801d0dc88eee35aae40c748d450ndOpen issues:
02b0660d7f44ff8106bd5095653e9c19e39ec37dnd * Which MPMs will be included with Apache 2.0?