STATUS revision b6de8bd60732d9dec6511a8ee4191bb85cdc8f1b
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark AndrewsAPACHE 2.0 STATUS: -*-text-*-
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark AndrewsLast modified at [$Date: 2002/01/11 08:07:07 $]
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews 2.0.31 : In development
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews 2.0.30 : tagged January 8, 2002.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews 2.0.29 : tagged November 27, 2001. not rolled.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews 2.0.28 : released November 13, 2001
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews 2.0.27 : rolled November 6, 2001
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews 2.0.26 : tagged October 16, 2001. not rolled.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews 2.0.25 : rolled August 29, 2001
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews 2.0.24 : rolled August 18, 2001
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews 2.0.23 : rolled August 9, 2001
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews 2.0.22 : rolled July 29, 2001
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews 2.0.21 : rolled July 20, 2001
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews 2.0.20 : rolled July 8, 2001
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews 2.0.19 : rolled June 27, 2001
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews 2.0.18 : rolled May 18, 2001
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews 2.0.17 : rolled April 17, 2001
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews 2.0.16 : rolled April 4, 2001
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews 2.0.15 : rolled March 21, 2001
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews 2.0.14 : rolled March 7, 2001
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews 2.0a9 : released December 12, 2000
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews 2.0a8 : released November 20, 2000
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews 2.0a7 : released October 8, 2000
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews 2.0a6 : released August 18, 2000
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews 2.0a5 : released August 4, 2000
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews 2.0a4 : released June 7, 2000
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews 2.0a3 : released April 28, 2000
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews 2.0a2 : released March 31, 2000
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews 2.0a1 : released March 10, 2000
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark AndrewsPlease consult the following STATUS files for information
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrewson related projects:
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark AndrewsRELEASE SHOWSTOPPERS:
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * ap_directory_walk skips some per-dir config merge functions
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews if there is no "<Directory />" block in the configuration
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Message-ID: <m3itbdiijq.fsf@rdu163-40-092.nc.rr.com>
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * Test suite failures:
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews o perchild doesn't even build
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews o worker is also failing some of the 'cgi' subtests
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews (see <URL:http://Source-Zone.Org/Apache/regression/>):
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * If any request gets to the core handler, without a flag that this
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews r->filename was tested by dir/file_walk, we need to 500 at the very
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews end of the ap_process_request_internal() processing. This provides
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews authors of older modules better compatibility, while still improving
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews the security and robustness of 2.0.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Status: still need to decide where this goes, OtherBill comments...
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Message-ID: <065701c14526$495203b0$96c0b0d0@roweclan.net>
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews we need to look at halting this in the 'default handler' case,
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews and that implies pushing the 'handler election' into the request
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews internal processing phase from the run request phase.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * There is a bug in how we sort some hooks, at least the pre-config
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews hook. The first time we call the hooks, they are in the correct
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews order, but the second time, we don't sort them correctly. Currently,
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews the modules/http/config.m4 file has been renamed to
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews modules/http/config2.m4 to work around this problem, it should moved
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews back when this is fixed. rbb
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * The Add...Filter and Set...Filter directives do not allow the
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews administrator to order filters, beyond the order of filename (mime)
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews extensions. It isn't clear if Set...Filter(s) should be inserted
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews before or after the Add...Filter(s) which are ordered by sequence of
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews filename extensions. At minimum, some sort of +-[0-10] syntax seems
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews like the quickest fix for a 2.0 gold release.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * mod_negotiation needs a new option or directive, something like
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews ForceLanguagePriority, to fall back to the LanguagePriority
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews directive instead of returning a "no acceptable variant" error.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Status: Bill has some code in his tree that accomplishes
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews this, and will commit it Friday after it's tested.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * Fold mod_auth_db features back into mod_auth_dbm, and depricate it.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews This can't wait until we have a 2.0-gold release, if folks need
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews to move over to auth_dbm, we can't do that to them after 2.0 gold.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Status: Ian says.. auth_dbm can now handle multiple DBM types,
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews is this still an issue?
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Vote: Remove mod_auth_db
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * Convert all instances of the old apr_lock_t type to the new
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews types (once they are fully supported in APR).
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Status: Aaron is working on converting INTRAPROCESS
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews to apr_thread_mutex_t types. Full replacements for
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews LOCKALL and CROSS_PROCESS are not yet complete on all
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews platforms, and should only be used in MPMs like worker
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews with limited OS exposure.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * ap_create_scoreboard() can exit the process, leaving stuff like
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews mod_cgid's daemon process stranded. Either ap_create_scoreboard()
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews needs to be called at a different time or the pre-mpm hook needs
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews to be able to return an error code.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark AndrewsRELEASE NON-SHOWSTOPPERS BUT WOULD BE REAL NICE TO WRAP THESE UP:
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * Add a Logging Directive to exclude logging based on mime type.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Message-ID: <ELEDJONBOPPAEGANDEEIEENCCAAA.joshua@slive.ca>
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews -1: Joshua Silve, WRowe, Lars
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews +1: Ian (surprise)
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews But is matching on mime-types really a required feature?
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews IMHO most people are happy with the current conditional logging
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews feature, e.g., matching on REQUEST_URI.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews I tend to be -1 on your patch Ian, because the directive is too
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews specific. We either need a more generic one or just stick with
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews the current (or an enhanced) conditional logging functionality.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Large Sites/Sites with other heavy CPU loads need a 'cheap'
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews way to exclude 90% of all log lines (eg images)
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Regex Matching takes ~5-10 times longer than a simple hash get.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * Try to get libtool inter-library dependency code working on AIX.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Message-ID: <cm3n10lx555.fsf@rdu163-40-092.nc.rr.com>
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Justin says: If we get it working on AIX, we can enable this
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews on all platforms and clean up our build system
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * Handling of %2f in URIs. Currently both 1.3 and 2.0
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews completely disallow %2f in the request URI path (see
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews ap_unescape_url() in util.c). It's permitted and passed
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews through in the query string, however. Roy says the
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews original reason for disallowing it, from five years ago,
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews was to protect CGI scripts that applied PATH_INFO to
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews a filesystem location and which might be tricked by
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews ..%2f..%2f(...). We *should* allow path-info of the
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews form 'http://foo.com/index.cgi/path/to/path%2finfo'.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Since we've revamped a lot of our processing of path
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews segments, it would be nice to allow this, or at least
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews allow it conditionally with a directive.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * FreeBSD, threads, and worker MPM. All seems to work fine
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews if you only have one worker process with many threads. Add
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews a second worker process and the accept lock seems to be
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews lost. This might be an APR issue with how it deals with
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews the child_init hook (i.e. the fcntl lock needs to be resynced).
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews More examination and analysis is required.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Status: This has also been reported on Cygwin.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Message-ID: <3C2CC514.8EF3BED1@wapme-systems.de> (cygnus)
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Justin says: So, FreeBSD-CURRENT and Cywin have the same
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews problem. Yum. If another platform has this
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews with worker, this becomes a showstopper.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Aaron says: I spent some time disecting this and have come to
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews the conclusion that it is not a problem in the worker MPM
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews (or at least, it is not isolated to a problem in worker).
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews I'll list some of the problems I'm seeing in case someone
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews else wants to pick up where I've left off:
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews - Delivery of just about any signal to one of the child
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews processes will send it into an infinite loop as well.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews - Even though the parent is spinning out of control,
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews at first the child or children will appear to work
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews properly. At times it is possible to get it into a state,
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews however, where a request will hang until another concurrent
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews request "kicks" the first, at which point the second will
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews hang. My theory is that this has to do with the
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews pthread_cond_*() implementation in FreeBSD, but it's still
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews possible that it is in APR.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Justin adds: Oh, FreeBSD threads are implemented entirely with
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews select()/poll()/longjmp(). Welcome to the nightmare.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews So, that means a ktrace output also has the thread
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews scheduling internals in it (since it is all the same to
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews the kernel). Which makes it hard to distinguish between
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews our select() calls and their select() calls.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews *bangs head on wall repeatedly* But, some of the libc_r
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews files have a DBG_MSG #define. This is moderately helpful
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews when used with -DNO_DETACH. The kernel scheduler isn't
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews waking up the threads on a select(). Yum. And, I bet
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews those decrementing select calls have to do with the
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews scheduler. Time to brush up on our OS fundamentals.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * There is increasing demand from module writers for an API
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews that will allow them to control the server � la apachectl.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Reasons include sole-function servers that need to die if
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews an external dependency (e.g., a database) fails, et cetera.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Perhaps something in the (ever more abused) scoreboard?
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews rbb: I don't believe the scoreboard is the correct mechanism
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews for this. We already have a pipe that goes between parent
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews and child for graceful shutdown events, along with an API that
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews can be used to send a message down that pipe. In threaded MPMs,
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews it is easy enough to make that one pipe be used for graceful
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews and graceless events, and it is also easy to open that pipe
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews to both parent and child for writing. Then we just need to figure
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews out how to do graceless on non-threaded MPMs.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * revamp the input filter behavior, per discussions since
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews February (and especially at the hackathon last
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews April). Specifically, ap_get_brigade will return a brigade with
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews *up to* a specific number of bytes, or a "line" of data. The
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews read may be blocking or nonblocking. ap_getline() will be
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews refactored into apr_brigade_getline(), and then DECHUNK can use
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews f->next (ap_getline will always read "top of input stack"). Also
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews fix the bug where request body content will end up closing the
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews connection (buggering up persistent conns).
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Status: Justin is working on this as fast as he can.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews The core input filters, HTTP-related filters, mod_ssl, and
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews mod_proxy are switched to the new logic.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews However, ap_getline() still needs to be refactored out. But,
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews there's a problem there: ap_getline() peeks ahead for MIME
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews continuation (first character on line is space or \t) and
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews stores unused data in core_request_config which violates the
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews abstraction. That's cheating. So, we may not be able to
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews implement this without setting some data aside (yuck!).
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews I believe this is OtherBill's main complaint with the current
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews AIUI (correct me if I'm wrong!), OtherBill believes we
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews should have a pushback option so that we can return unread
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews data - this would solve this case. However, my question to
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews him is how do we handle stuff like mod_ssl - we can't "unread"
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews data. So, do we have two brigades for each filter? An in
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews brigade and a returned brigade? That seems messy. To
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews everyone else, can we refactor ap_getline() without pushback
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews - socket bucket and core input filter changes. see end of
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews message ID (Feb 27): <20010227075326.S2297@lyra.org>
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews - fix up ap_get_brigade() semantics, fix bug in DECHUNK /
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews ap_getline. many messages (plus their threads) (Apr/May):
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Message-ID: <20010402101207.J27539@lyra.org>
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Message-ID: <3AF7F921.D2EEC41A@algroup.co.uk>
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Message-ID: <20010508190029.E18404@lyra.org>
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews - further work with combining/tweaking the builtin filters:
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Message-ID: <20010509115445.D1374@lyra.org>
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews - thoughts on filter modes:
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Message-ID: <021b01c14dee$09782af0$93c0b0d0@roweclan.net>
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews - proposal for getline rewrite:
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Message-ID: <20011231104019.GD3904@ebuilt.com>
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * Allow the DocumentRoot directive within <Location > scopes? This
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews allows the beloved (crusty) Alias /foo/ /somepath/foo/ followed
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews by a <Directory /somepath/foo> to become simply
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews <Location /foo/> DocumentRoot /somefile/foo (IMHO a bit more legible
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews and in-your-face.) DocumentRoot unset would be accepted [and would
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews not permit content to be served, only virtual resources such as
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews server-info or server-status.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews This proposed change would _not_ depricate Alias.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * Win32: Rotatelogs sometimes is not terminated when Apache
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews goes down hard. FirstBill was looking at possibly tracking the
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews child's-child processes in the parent process.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews OtherBill asks, wasn't this fixed?
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * Win32: Add a simple hold console open patch (wait for close or
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews the ESC key, with a nice message) if the server died a bad
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews death (non-zero exit code) in console mode.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Resolution: bring forward same ugly hacks from 1.3.13-.20
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * Port of mod_ssl to Apache 2.0:
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews The current porting state is summarized in modules/ssl/README. The
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews remaining work includes:
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews (1) stablizing/optimizing the SSL filter logic
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews (2) Enabling the various SSL caching mechanisms (shmcb, shmht)
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews (3) Enabling SSL extentions
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews (4) Trying to seperate the https filter logic from mod_ssl -
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews This is to facilitate other modules that wish to use the https
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews filter or the mod_ssl logic or both as required.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Justin: mod_ssl filter logic is redone, so that should be fine.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Madhu has submitted a patch for SSL caching - however, I
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews am -0 on that patch as I *think* we could implement the
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews shared memory another way that is much cleaner (i.e.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews treat shmem directly as a dbm via APR routines). Justin
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews also thinks that the https filter logic may be sufficiently
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews decoupled now, but isn't really sure.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * Performance: Get the SINGLE_LISTEN_UNSERIALIZED_ACCEPT
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews optimization working in worker. prefork's new design for how
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews to notice data on the pod should be sufficient.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * Performance & Debug: Eliminate most (and perhaps all) of the
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews malloc/free calls in the bucket brigade code. Need some
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews light weight memory management functions that allow freeing
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews memory (putting it back into a memory pool) when it is no
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews longer needed. Enabling simple debugging features like guard
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews bands, double free detection, etc. would be cool but certainly
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews not a hard requirement.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Status: Cliff started to implement this using SMS as has
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews been discussed at length for months, but since
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews SMS is not being used anywhere else in the server,
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews several people expressed the opinion that we should
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews get rid of it entirely, meaning that the buckets
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews need their own memory management (free list) functions.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Cliff will implement that this weekend so we at least
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews have something to look at/compare with.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * Eliminate unnecessary creation of pipes in mod_cgid
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * the autoconf setup should be fixed to default to using the
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews "Apache" layout from config.layout, and each variable settable
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews in a layout should be overridable on the command line. Plus,
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews what we do right now just doesn't seem to fully fit into how autoconf
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews works, eg. AC_PREFIX_DEFAULT issues.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Message-ID: <Pine.BSF.4.20.0104031557420.20876-100000@alive.znep.com>
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * Combine log_child and piped_log_spawn. Clean up http_log.c.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Common logging API.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * Document mod_file_cache.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * OS/2: Make mod_status work for spmt_os2 MPM.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * Platforms that do not support fork (primarily Win32 and AS/400)
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Architect start-up code that avoids initializing all the modules
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews in the parent process on platforms that do not support fork.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * Win32: Migrate the MPM over to use APR thread/process calls. This
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews would eliminate some code in the Win32 branch that essentially
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews duplicates what is in APR.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * There are still a number of places in the code where we are
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews losing error status (i.e. throwing away the error returned by a
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews system call and replacing it with a generic error code)
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * Mass vhosting version of suEXEC.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * All DBMs suffer from confusion in support/dbmmanage (perl script) since
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews the dbmmanage employs the first-matched dbm format. This is not
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews necessarily the library that Apache was built with. Aught to
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews rewrite dbmmanage upon installation to bin/ with the proper library
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews for predictable mod_auth_db/dbm administration.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Status: Mladen Turk has posted several patches and ideas.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Key question, part of htpasswd, or a seperate utility?
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews prefer htpasswd:
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews prefer seperate: OtherBill
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * use apu_dbm in mod_auth_dbm
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Status: Greg +1 (low-priority volunteer)
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Justin says: "Seems like this is already there, so should we just
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews remove the other DBM code in that file? If you want
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews to use gdbm, or dbm, etc, you should tell apr-util."
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Will says: "bs - I may choose the fastest - most efficient native
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews dbm implementation, for shared proc caches, ssl session
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews caching, etc, but that has nothing to do with maintaining
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews a userlist via dbm, which has to remain readable between
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews builds/machines, etc. The use-multiple database schema
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews for apr-util would let us do this with just apr, though."
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Ian says: "multi-dbm is in, but it still has ndbm support hardcoded
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews is this still required? isn't ndbm supported via gdbm?"
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * Integrate mod_dav.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Some additional items remaining:
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews - case_preserved_filename stuff
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews (use the new canonical name stuff?)
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews - find a new home for ap_text(_header)
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews - is it possible to remove the DAV: namespace stuff from util_xml?
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * ap_core_translate() and its use by mod_mmap_static and mod_file_cache
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews are a bit wonky. The function should probably be exposed as a utility
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews function (such as ap_translate_url2fs() or ap_validate_fs_url() or
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews something). Another approach would be a new hook phase after
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews "translate" which would allow the module to munge what the
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews translation has decided to do.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Status: Greg +1 (volunteers), Ryan +1
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * Explore use of a post-config hook for the code in http_main.c which
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews calls ap_fixup_virutal_hosts(), ap_fini_vhost_config(), and
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews ap_sort_hooks() [to reduce the logic in main()]
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * read the config tree just once, and process N times (as necessary)
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * (possibly) use UUIDs in mod_unique_id and/or mod_usertrack
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * (possibly) port the bug fix for PR 6942 (segv when LoadModule is put
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews into a VirtualHost container) to 2.0.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * shift stuff to mod_core.h
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * callers of ap_run_create_request() should check the return value
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews for failure (Doug volunteers)
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * Win32: Get Apache working on Windows 95/98. The following work
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews (at least) needs to be done:
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews - winnt MPM: Fix 95/98 code paths in the winnt MPM. There is some NT
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews specific code that is still not in NT only code paths
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews - IOL binds to APR sendfile, implemented with TransmitFile, which
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews is not available on 95/98.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews - Document warning that OSR2 is required (for Crypt functions, in
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews rand.c, at least.) This could be resolved with an SSL library, or
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews randomization in APR itself.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews - Bring the Win9xConHook.dll from 1.3 into 2.0 (no sense till it
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews actually works) and add in a splash of Win9x service code.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * In order to use a DSO version of mod_ssl we have to link with
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews -lssl and -lcrypto. A workaround is in place right now where the
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews entire EXTRA_LIBS macro is being appended to the objects list, but
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews this is a hack. We should either revamp the APACHE_CHECK_SSL_TOOLKIT
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews autoconf function or come up with some other autoconf checks to
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews search for libssl and libcrypto and properly add them to mod_ssl's
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * Make the worker MPM the default MPM for threaded Unix boxes.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews +1: Justin, Jeff, Ian
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews -0: Aaron (premature decision, needs more discussion)
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews -0: Cliff (I think the default config should be the safest possible)
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * Fix the worker MPM to use POD to kill child processes instead
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews of ap_os_killpg, regardless of how they should die. (Ryan Bloom)
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark AndrewsPRs that have been suspended forever waiting for someone to
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrewsput them into 'the next release':
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#76: general
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews missing call to "setlocale();"
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#78: mod_include
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Additional status for XBitHack directive
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#362: mod_proxy
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Mod_proxy doesn't allow change of error pages
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#370: mod_env
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Modified PATH environemnt variable is not passed, instead
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews system's is used
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#440: mod_proxy
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Proxy doesn't deliver documents if not connected
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#534: mod_proxy
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews proxy converts ~name to %7Ename when name starts with a dot (.)
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#537: mod_access
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews mod_access syntax allows hosts that should be restricted
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#557: mod_auth-any
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews ~UserHome directories are not honored in absolute pathname
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews requests (.htaccess)
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#612: mod_proxy
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Proxy FTP Authentication Fails
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#623: mod_include
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews A smarter "Last Modified" value for SSI documents (see PR number 600)
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#628: config
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Request of "Options SymLinksIfGroupMatch"
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#700: mod_proxy
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Proxy doesn't do links right for OpenVMS files through ftp:
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#759: mod_imap
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews imap should read <MAP><AREA>*</MAP> too!
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#793: general
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews RLimitCPU and RLimitMEM don't apply to all children like they should
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#921: suexec
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Uses cwd before filling it in, doesn't use syslog
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#922: config
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews it is useful to allow specifiction that root-owned symlinks
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews should always be followed
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#980: mod_proxy
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Controlling Access to Remote Proxies would be nice...
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#994: mod_proxy
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Adding authentication "on the fly" through the proxy module
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#1004: apache-api
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews request_config field in request_rec is moderately bogus
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#1028: other
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews DoS attacks involving memory consumption
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#1050: mod_log-any
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Logging of virtual server to error_log as well
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#1085: mod_proxy
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews ProxyRemote make a dead cycle.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#1117: mod_auth-any
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Using NIS passwd.byname dbm files with AuthDBMUserFile
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#1120: suexec
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews suexec does not parse arguments to #exec cmd
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#1145: mod_include
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Allow for Last-Modified: without resorting to XBitHack
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#1158: apache-api
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews improvements to child spawning API
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#1166: mod_proxy
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews ``nph-'' not honored (no buffering) for ProxyRemote mapping
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#1176: mod_cgi
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Apache cannot handle continuation line in headers
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#1191: general
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews setlogin() is not called, causing problems with e.g. identd
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#1204: general
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews regerror() exists, use it
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#1233: apache-api
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews there is no way to keep per-connection per-module state
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#1263: mod_autoexec
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Add frame-safe anchor attribute to mod_autoindex links
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#1268: suexec
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews CGI scripts running as Apache user: security (suexec etc.)
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#1285: suexec
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Error messages could be easier to spot in cgi.log file for suexec.c
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#1287: mod_access
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews add allow,deny/deny,allow warning to mod_access
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#1290: mod_proxy
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Need to know "hit-rate" on proxy cache
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#1358: mod_log-any
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Selective url-encode of log fields (or maybe a pseudo
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews log_rewrite module?)
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#1383: mod_headers
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews I make mod_headers to modify request headers as well as
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews response ones.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#1532: mod_proxy
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Proxy transfer logging
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#1547: mod_proxy
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews No HTTP_X_FORWARDED_FOR set...
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#1567: mod_proxy
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews ProxyRemote proxy requests fail authentication by firewall
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#1582: mod_rewrite
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews mod_rewrite forms REQUEST_URI different than mod_cgi does
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#1677: mod_headers
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews mod_headers should allow mod_log_config-style formats in
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews header values
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#1702: mod_proxy
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews mod_proxy to support persistent conns?
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#1803: mod_include
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews patches to mod_include to allow for file tests
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#1809: mod_auth-any
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Suggestion for improving authentication modules and core source
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews code, problem with 401 and ErrorDocument
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#1878: mod_proxy
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews listing of proxy cache content
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#1905: suexec
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Allow modules to set user:group for execution.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#2024: apache-api
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews adding auth_why to conn_rec
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#2073: mod_log-any
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews pipelined connections are not logged correctly
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#2074: mod_rewrite
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews mod_rewrite doesn't pass Proxy Throughput on internal subrequests
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#2113: config
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews HTTP Server Rebuild Line Needs Changing for the better
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#2138: mod_status
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews mod_status always displays 256 possible connection slots
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#2221: documentation
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Make online documentation search link back to my installation
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#2284: general
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Can not POST to ErrorDocument - Apache/1.3b6
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#2314: mod_proxy
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews patterns in ProxyRemote
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#2343: mod_status
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Status module averages are for entire uptime
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#2360: suexec
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews suexec for general access of user content?
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#2396: general
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Proposal for TimeZone directive
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#2415: mod_info
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews /server-info doesn't check for the virtual host to list the info
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#2421: config
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews problem specifying ndbm library for build ?with autoconfigure
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#2431: general
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews A small addition to rotatelogs.c to improve program functionality.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#2446: config
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews AllowOverride FileInfo is too coarse
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#2460: mod_cgi
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews TimeOut applies to output of CGI scripts
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#2512: mod_access
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews <IfDenied> directive wanted
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#2573: suexec
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews CGI's for general use still have to be run as another user
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#2648: general
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Cache file names in Proxy module
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#2760: config
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews [PATCH] User/Group for <Directory> and <Location> i.e. not only
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews in global and <Virtual>.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#2763: general
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews mailto tags and bundling bug report script
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#2785: os-aix
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Support for System Resource Controller
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#2793: protocol
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews When will Apache support P3P? Any Plans?
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#2873: config
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#2889: general
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Inclusion of RPM spec file in CVS/distributions
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#2906: general
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Propose that Apache recommend $UNIQUE_ID for all "session id"
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#2907: config
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews suggestion: power up your Include directive :)
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#3018: general
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews cannot limit some HTTP methods
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#3143: apache-api
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews No module specific data hook for per-connection data
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#3191: mod_negotiation
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews no way to set global quality-of-source (qs) coneg values
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews with multiviews
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#3568: mod_proxy
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Accessing URL through proxy server corrupts data.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#3605: mod_proxy
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Some anonymous FTP URLs ask for authentication
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#3677: general
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews New ErrorDocumentMatch directive
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#4241: config
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Need to be able to override shebang line to make CGI scripts
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews more portable.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#4244: config
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews "Files" and "FilesMatch" regexp does not recognize bang as
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews negation operator
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#4448: mod_log-any
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Please allow CGI env variables (QUERY_STRING, ...) to be logged
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#4459: mod_include
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Suggestion for better handling of Last-modified headers
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#4490: mod_cgi
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews mod_cgi prevents handling of OPTIONS requests
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#5713: os-windows
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews [PATCH] install as win32 service with domain account
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * PR#5993: general
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews AllowOverride should have a 'CheckNone' and 'AllowNone' argument
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews instead of only 'None'
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark AndrewsOther bugs that need fixing:
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * MaxRequestsPerChild measures connections, not requests.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Until someone has a better way, we'll probably just rename it
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews "MaxConnectionsPerChild".
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * Regex containers don't work in an intutive way
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Status: No one has come up with an efficient way to fix this
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews behavior. Dean has suggested getting rid of regex containers
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * SIGSEGV on Linux (glibc 2.1.2) isn't caught properly by a
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews sigwaiting thread. We need to work around this, perhaps unless
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews there is hope soon for a fixed glibc.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * orig_ct in the byterange/multipart handling may not be
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews needed. Apache 1.3 just never stashed "multipart" into
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews r->content_type. We should probably follow suit since the
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews byterange stuff doesn't want the rest of the code to see the
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews multipart content-type; the other code should still think it is
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews dealing with the <orig_ct> stuff.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Status: Greg volunteers to investigate (esp. since he was most
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews likely the one to break it :-)
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark AndrewsOther features that need writing:
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * Finish infrastructure in core for async MPMs
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Status: post 2.0
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * TODO in source -- just do an egrep on "TODO" and see what's there
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark AndrewsAvailable Patches:
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * Jon Travis's <jtravis@covalent.net> patch to deal with thread-safe
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews issues with inet_ntoa. See message <20001201163220.A12827@covalent.net>
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Status: This is being set aside until the IPv6 work is finished
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews so that we know exactly what is required.
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * Martin Sojka <msojka@gmx.de>'s patch to add error reporting for failed
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews htpasswd actions due to a full /tmp volume (other programs may have
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews similar problems?)
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * Mike Abbott's <mja@trudge.engr.sgi.com> patches to improve
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews Status: These were written for 1.3, and are awaiting a port to
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * Jim Winstead's <jimw@trainedmonkey.com> patch to add CookieDomain and
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews other small mod_usertrack features
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * Dan Rench's <drench@xnet.com> patch to add allow the errmsg and timefmt
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews of SSI's to be modified in the config file. Patch is available in
29747dfe5e073a299b3681e01f5c55540f8bfed7Mark Andrews * Which MPMs will be included with Apache 2.0?