STATUS revision c942e80120ec33154f4089ea19c010239da34d82
843e19887f64dde75055cf8842fc4db2171eff45johnlevAPACHE 2.0 STATUS: -*-text-*-
843e19887f64dde75055cf8842fc4db2171eff45johnlevLast modified at [$Date: 2002/01/23 19:13:15 $]
843e19887f64dde75055cf8842fc4db2171eff45johnlev 2.0.31 : In development
843e19887f64dde75055cf8842fc4db2171eff45johnlev 2.0.30 : tagged January 8, 2002.
843e19887f64dde75055cf8842fc4db2171eff45johnlev 2.0.29 : tagged November 27, 2001. not rolled.
843e19887f64dde75055cf8842fc4db2171eff45johnlev 2.0.28 : released November 13, 2001
843e19887f64dde75055cf8842fc4db2171eff45johnlev 2.0.27 : rolled November 6, 2001
843e19887f64dde75055cf8842fc4db2171eff45johnlev 2.0.26 : tagged October 16, 2001. not rolled.
843e19887f64dde75055cf8842fc4db2171eff45johnlev 2.0.25 : rolled August 29, 2001
843e19887f64dde75055cf8842fc4db2171eff45johnlev 2.0.24 : rolled August 18, 2001
843e19887f64dde75055cf8842fc4db2171eff45johnlev 2.0.23 : rolled August 9, 2001
843e19887f64dde75055cf8842fc4db2171eff45johnlev 2.0.22 : rolled July 29, 2001
843e19887f64dde75055cf8842fc4db2171eff45johnlev 2.0.21 : rolled July 20, 2001
843e19887f64dde75055cf8842fc4db2171eff45johnlev 2.0.20 : rolled July 8, 2001
843e19887f64dde75055cf8842fc4db2171eff45johnlev 2.0.19 : rolled June 27, 2001
843e19887f64dde75055cf8842fc4db2171eff45johnlev 2.0.18 : rolled May 18, 2001
843e19887f64dde75055cf8842fc4db2171eff45johnlev 2.0.17 : rolled April 17, 2001
843e19887f64dde75055cf8842fc4db2171eff45johnlev 2.0.16 : rolled April 4, 2001
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab 2.0.15 : rolled March 21, 2001
843e19887f64dde75055cf8842fc4db2171eff45johnlev 2.0.14 : rolled March 7, 2001
843e19887f64dde75055cf8842fc4db2171eff45johnlev 2.0a9 : released December 12, 2000
843e19887f64dde75055cf8842fc4db2171eff45johnlev 2.0a8 : released November 20, 2000
843e19887f64dde75055cf8842fc4db2171eff45johnlev 2.0a7 : released October 8, 2000
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab 2.0a6 : released August 18, 2000
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab 2.0a5 : released August 4, 2000
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab 2.0a4 : released June 7, 2000
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab 2.0a3 : released April 28, 2000
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab 2.0a2 : released March 31, 2000
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab 2.0a1 : released March 10, 2000
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rabPlease consult the following STATUS files for information
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rabon related projects:
843e19887f64dde75055cf8842fc4db2171eff45johnlevRELEASE SHOWSTOPPERS:
843e19887f64dde75055cf8842fc4db2171eff45johnlev * win32 scoreboard should _not_ use anon with no scoreboard name, it
843e19887f64dde75055cf8842fc4db2171eff45johnlev should revert to the old method if scoreboard name is null (use the
843e19887f64dde75055cf8842fc4db2171eff45johnlev NOT_SHARED bit that we do in -X). At this moment Win32 doesn't
843e19887f64dde75055cf8842fc4db2171eff45johnlev -need- to have a shared score, until we create multiple processes,
843e19887f64dde75055cf8842fc4db2171eff45johnlev or unless a module in the parent wants scoreboard data from the child.
843e19887f64dde75055cf8842fc4db2171eff45johnlev If such a module is installed [for the moment] they aught to configure
843e19887f64dde75055cf8842fc4db2171eff45johnlev the ScoreboardFile directive.
843e19887f64dde75055cf8842fc4db2171eff45johnlev * mod_rewrite's locks (win32) were broken by recent changes to the
843e19887f64dde75055cf8842fc4db2171eff45johnlev apr_lock_t api (which we were deprecating in the first place, so
843e19887f64dde75055cf8842fc4db2171eff45johnlev I'm far from certain why it changed.) It has to do with creating
843e19887f64dde75055cf8842fc4db2171eff45johnlev the rewrite logs mutex. The solution is to find the brokenness in
843e19887f64dde75055cf8842fc4db2171eff45johnlev apr_lock_t, but if it goes away, why waste the time?
843e19887f64dde75055cf8842fc4db2171eff45johnlev * ap_directory_walk skips some per-dir config merge functions
843e19887f64dde75055cf8842fc4db2171eff45johnlev if there is no "<Directory />" block in the configuration
843e19887f64dde75055cf8842fc4db2171eff45johnlev Message-ID: <m3itbdiijq.fsf@rdu163-40-092.nc.rr.com>
843e19887f64dde75055cf8842fc4db2171eff45johnlev * Test suite failures:
843e19887f64dde75055cf8842fc4db2171eff45johnlev o perchild doesn't even build
843e19887f64dde75055cf8842fc4db2171eff45johnlev o worker is also failing some of the 'cgi' subtests
843e19887f64dde75055cf8842fc4db2171eff45johnlev (see <URL:http://Source-Zone.Org/Apache/regression/>):
843e19887f64dde75055cf8842fc4db2171eff45johnlev * If any request gets to the core handler, without a flag that this
843e19887f64dde75055cf8842fc4db2171eff45johnlev r->filename was tested by dir/file_walk, we need to 500 at the very
843e19887f64dde75055cf8842fc4db2171eff45johnlev end of the ap_process_request_internal() processing. This provides
843e19887f64dde75055cf8842fc4db2171eff45johnlev authors of older modules better compatibility, while still improving
843e19887f64dde75055cf8842fc4db2171eff45johnlev the security and robustness of 2.0.
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab Status: still need to decide where this goes, OtherBill comments...
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab Message-ID: <065701c14526$495203b0$96c0b0d0@roweclan.net>
843e19887f64dde75055cf8842fc4db2171eff45johnlev we need to look at halting this in the 'default handler' case,
843e19887f64dde75055cf8842fc4db2171eff45johnlev and that implies pushing the 'handler election' into the request
843e19887f64dde75055cf8842fc4db2171eff45johnlev internal processing phase from the run request phase.
843e19887f64dde75055cf8842fc4db2171eff45johnlev * There is a bug in how we sort some hooks, at least the pre-config
843e19887f64dde75055cf8842fc4db2171eff45johnlev hook. The first time we call the hooks, they are in the correct
843e19887f64dde75055cf8842fc4db2171eff45johnlev order, but the second time, we don't sort them correctly. Currently,
843e19887f64dde75055cf8842fc4db2171eff45johnlev the modules/http/config.m4 file has been renamed to
843e19887f64dde75055cf8842fc4db2171eff45johnlev modules/http/config2.m4 to work around this problem, it should moved
843e19887f64dde75055cf8842fc4db2171eff45johnlev back when this is fixed. rbb
843e19887f64dde75055cf8842fc4db2171eff45johnlev * The Add...Filter and Set...Filter directives do not allow the
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab administrator to order filters, beyond the order of filename (mime)
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab extensions. It isn't clear if Set...Filter(s) should be inserted
5d2eda970e48f8985448151c73e699614ce9f357John Levon before or after the Add...Filter(s) which are ordered by sequence of
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab filename extensions. At minimum, some sort of +-[0-10] syntax seems
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab like the quickest fix for a 2.0 gold release.
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab * mod_negotiation needs a new option or directive, something like
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab ForceLanguagePriority, to fall back to the LanguagePriority
843e19887f64dde75055cf8842fc4db2171eff45johnlev directive instead of returning a "no acceptable variant" error.
843e19887f64dde75055cf8842fc4db2171eff45johnlev Status: Bill has some code in his tree that accomplishes
843e19887f64dde75055cf8842fc4db2171eff45johnlev this, and will commit it Friday after it's tested.
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab * Fold mod_auth_db features back into mod_auth_dbm, and depricate it.
5d2eda970e48f8985448151c73e699614ce9f357John Levon This can't wait until we have a 2.0-gold release, if folks need
843e19887f64dde75055cf8842fc4db2171eff45johnlev to move over to auth_dbm, we can't do that to them after 2.0 gold.
843e19887f64dde75055cf8842fc4db2171eff45johnlev Status: Ian says.. auth_dbm can now handle multiple DBM types,
5d2eda970e48f8985448151c73e699614ce9f357John Levon is this still an issue?
5d2eda970e48f8985448151c73e699614ce9f357John Levon Vote: Remove mod_auth_db
843e19887f64dde75055cf8842fc4db2171eff45johnlev +1: Justin, Ian, Lars
843e19887f64dde75055cf8842fc4db2171eff45johnlev * Convert all instances of the old apr_lock_t type to the new
843e19887f64dde75055cf8842fc4db2171eff45johnlev types (once they are fully supported in APR).
843e19887f64dde75055cf8842fc4db2171eff45johnlev Status: Aaron is working on converting INTRAPROCESS
843e19887f64dde75055cf8842fc4db2171eff45johnlev to apr_thread_mutex_t types. Full replacements for
843e19887f64dde75055cf8842fc4db2171eff45johnlev LOCKALL and CROSS_PROCESS are not yet complete on all
843e19887f64dde75055cf8842fc4db2171eff45johnlev platforms, and should only be used in MPMs like worker
843e19887f64dde75055cf8842fc4db2171eff45johnlev with limited OS exposure.
843e19887f64dde75055cf8842fc4db2171eff45johnlev * ap_create_scoreboard() can exit the process, leaving stuff like
843e19887f64dde75055cf8842fc4db2171eff45johnlev mod_cgid's daemon process stranded. Either ap_create_scoreboard()
843e19887f64dde75055cf8842fc4db2171eff45johnlev needs to be called at a different time or the pre-mpm hook needs
843e19887f64dde75055cf8842fc4db2171eff45johnlev to be able to return an error code.
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab * A binbuild installation picks up the right libraries when
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab running apachectl because we set the appropriate environment
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab variable, but ab, htpasswd, etc. don't know how to pick up apr,
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab apr-util, etc.
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab Message-ID: <20020116000226.GA15991@ebuilt.com>
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab * mod_cgid leaks pipe descriptors, apparently for every request.
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab I would guess that this is due to recent apr_file_dup[2]
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab I was unable to recreate the problem with mod_cgi (yea!).
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab Message-ID: <20020123070827.GV28051@ebuilt.com> (dev@apr)
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab * Aaron is working on an update to the scoreboard creation
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab routines that will be more "correct" in terms of using
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab anonymous shared memory vs. name-based shared memory on
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab the platforms that support either.
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab * mod_negotiation seems to have out-of-control recursive subreqs.
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab Message-ID: <3C4903A2.68A7B4C5@remulak.net>
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rabRELEASE NON-SHOWSTOPPERS BUT WOULD BE REAL NICE TO WRAP THESE UP:
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab * Try to get libtool inter-library dependency code working on AIX.
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab Message-ID: <cm3n10lx555.fsf@rdu163-40-092.nc.rr.com>
843e19887f64dde75055cf8842fc4db2171eff45johnlev Justin says: If we get it working on AIX, we can enable this
843e19887f64dde75055cf8842fc4db2171eff45johnlev on all platforms and clean up our build system
843e19887f64dde75055cf8842fc4db2171eff45johnlev * Handling of %2f in URIs. Currently both 1.3 and 2.0
843e19887f64dde75055cf8842fc4db2171eff45johnlev completely disallow %2f in the request URI path (see
843e19887f64dde75055cf8842fc4db2171eff45johnlev ap_unescape_url() in util.c). It's permitted and passed
843e19887f64dde75055cf8842fc4db2171eff45johnlev through in the query string, however. Roy says the
843e19887f64dde75055cf8842fc4db2171eff45johnlev original reason for disallowing it, from five years ago,
843e19887f64dde75055cf8842fc4db2171eff45johnlev was to protect CGI scripts that applied PATH_INFO to
843e19887f64dde75055cf8842fc4db2171eff45johnlev a filesystem location and which might be tricked by
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab ..%2f..%2f(...). We *should* allow path-info of the
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab Since we've revamped a lot of our processing of path
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab segments, it would be nice to allow this, or at least
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab allow it conditionally with a directive.
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab * FreeBSD, threads, and worker MPM. All seems to work fine
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab if you only have one worker process with many threads. Add
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab a second worker process and the accept lock seems to be
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab lost. This might be an APR issue with how it deals with
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab the child_init hook (i.e. the fcntl lock needs to be resynced).
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab More examination and analysis is required.
843e19887f64dde75055cf8842fc4db2171eff45johnlev Status: This has also been reported on Cygwin.
843e19887f64dde75055cf8842fc4db2171eff45johnlev Message-ID: <3C2CC514.8EF3BED1@wapme-systems.de> (cygnus)
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab Justin says: So, FreeBSD-CURRENT and Cywin have the same
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab problem. Yum. If another platform has this
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab with worker, this becomes a showstopper.
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab Aaron says: I spent some time disecting this and have come to
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab the conclusion that it is not a problem in the worker MPM
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab (or at least, it is not isolated to a problem in worker).
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab I'll list some of the problems I'm seeing in case someone
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab else wants to pick up where I've left off:
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab - Delivery of just about any signal to one of the child
843e19887f64dde75055cf8842fc4db2171eff45johnlev processes will send it into an infinite loop as well.
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab - Even though the parent is spinning out of control,
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab at first the child or children will appear to work
843e19887f64dde75055cf8842fc4db2171eff45johnlev properly. At times it is possible to get it into a state,
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab however, where a request will hang until another concurrent
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab request "kicks" the first, at which point the second will
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab hang. My theory is that this has to do with the
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab pthread_cond_*() implementation in FreeBSD, but it's still
843e19887f64dde75055cf8842fc4db2171eff45johnlev possible that it is in APR.
843e19887f64dde75055cf8842fc4db2171eff45johnlev Justin adds: Oh, FreeBSD threads are implemented entirely with
843e19887f64dde75055cf8842fc4db2171eff45johnlev select()/poll()/longjmp(). Welcome to the nightmare.
843e19887f64dde75055cf8842fc4db2171eff45johnlev So, that means a ktrace output also has the thread
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab scheduling internals in it (since it is all the same to
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab the kernel). Which makes it hard to distinguish between
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab our select() calls and their select() calls.
843e19887f64dde75055cf8842fc4db2171eff45johnlev *bangs head on wall repeatedly* But, some of the libc_r
843e19887f64dde75055cf8842fc4db2171eff45johnlev files have a DBG_MSG #define. This is moderately helpful
843e19887f64dde75055cf8842fc4db2171eff45johnlev when used with -DNO_DETACH. The kernel scheduler isn't
843e19887f64dde75055cf8842fc4db2171eff45johnlev waking up the threads on a select(). Yum. And, I bet
843e19887f64dde75055cf8842fc4db2171eff45johnlev those decrementing select calls have to do with the
843e19887f64dde75055cf8842fc4db2171eff45johnlev scheduler. Time to brush up on our OS fundamentals.
843e19887f64dde75055cf8842fc4db2171eff45johnlev * There is increasing demand from module writers for an API
843e19887f64dde75055cf8842fc4db2171eff45johnlev that will allow them to control the server � la apachectl.
843e19887f64dde75055cf8842fc4db2171eff45johnlev Reasons include sole-function servers that need to die if
843e19887f64dde75055cf8842fc4db2171eff45johnlev an external dependency (e.g., a database) fails, et cetera.
843e19887f64dde75055cf8842fc4db2171eff45johnlev Perhaps something in the (ever more abused) scoreboard?
843e19887f64dde75055cf8842fc4db2171eff45johnlev rbb: I don't believe the scoreboard is the correct mechanism
843e19887f64dde75055cf8842fc4db2171eff45johnlev for this. We already have a pipe that goes between parent
843e19887f64dde75055cf8842fc4db2171eff45johnlev and child for graceful shutdown events, along with an API that
843e19887f64dde75055cf8842fc4db2171eff45johnlev can be used to send a message down that pipe. In threaded MPMs,
843e19887f64dde75055cf8842fc4db2171eff45johnlev it is easy enough to make that one pipe be used for graceful
843e19887f64dde75055cf8842fc4db2171eff45johnlev and graceless events, and it is also easy to open that pipe
843e19887f64dde75055cf8842fc4db2171eff45johnlev to both parent and child for writing. Then we just need to figure
843e19887f64dde75055cf8842fc4db2171eff45johnlev out how to do graceless on non-threaded MPMs.
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab * revamp the input filter behavior, per discussions since
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab February (and especially at the hackathon last
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab April). Specifically, ap_get_brigade will return a brigade with
843e19887f64dde75055cf8842fc4db2171eff45johnlev *up to* a specific number of bytes, or a "line" of data. The
843e19887f64dde75055cf8842fc4db2171eff45johnlev read may be blocking or nonblocking. ap_getline() will be
843e19887f64dde75055cf8842fc4db2171eff45johnlev refactored into apr_brigade_getline(), and then DECHUNK can use
843e19887f64dde75055cf8842fc4db2171eff45johnlev f->next (ap_getline will always read "top of input stack"). Also
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab fix the bug where request body content will end up closing the
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab connection (buggering up persistent conns).
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab Status: Justin is working on this as fast as he can.
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab The core input filters, HTTP-related filters, mod_ssl, and
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab mod_proxy are switched to the new logic.
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab However, ap_getline() still needs to be refactored out. But,
843e19887f64dde75055cf8842fc4db2171eff45johnlev there's a problem there: ap_getline() peeks ahead for MIME
843e19887f64dde75055cf8842fc4db2171eff45johnlev continuation (first character on line is space or \t) and
843e19887f64dde75055cf8842fc4db2171eff45johnlev stores unused data in core_request_config which violates the
843e19887f64dde75055cf8842fc4db2171eff45johnlev abstraction. That's cheating. So, we may not be able to
843e19887f64dde75055cf8842fc4db2171eff45johnlev implement this without setting some data aside (yuck!).
843e19887f64dde75055cf8842fc4db2171eff45johnlev I believe this is OtherBill's main complaint with the current
843e19887f64dde75055cf8842fc4db2171eff45johnlev AIUI (correct me if I'm wrong!), OtherBill believes we
843e19887f64dde75055cf8842fc4db2171eff45johnlev should have a pushback option so that we can return unread
843e19887f64dde75055cf8842fc4db2171eff45johnlev data - this would solve this case. However, my question to
843e19887f64dde75055cf8842fc4db2171eff45johnlev him is how do we handle stuff like mod_ssl - we can't "unread"
843e19887f64dde75055cf8842fc4db2171eff45johnlev data. So, do we have two brigades for each filter? An in
843e19887f64dde75055cf8842fc4db2171eff45johnlev brigade and a returned brigade? That seems messy. To
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab everyone else, can we refactor ap_getline() without pushback
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab - socket bucket and core input filter changes. see end of
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab message ID (Feb 27): <20010227075326.S2297@lyra.org>
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab - fix up ap_get_brigade() semantics, fix bug in DECHUNK /
843e19887f64dde75055cf8842fc4db2171eff45johnlev ap_getline. many messages (plus their threads) (Apr/May):
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab Message-ID: <20010402101207.J27539@lyra.org>
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab Message-ID: <3AF7F921.D2EEC41A@algroup.co.uk>
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab Message-ID: <20010508190029.E18404@lyra.org>
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab - further work with combining/tweaking the builtin filters:
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab Message-ID: <20010509115445.D1374@lyra.org>
843e19887f64dde75055cf8842fc4db2171eff45johnlev - thoughts on filter modes:
843e19887f64dde75055cf8842fc4db2171eff45johnlev Message-ID: <021b01c14dee$09782af0$93c0b0d0@roweclan.net>
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab - proposal for getline rewrite:
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab Message-ID: <20011231104019.GD3904@ebuilt.com>
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab - patch for getline rewrite:
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab Message-ID: <20020122072605.GF28051@ebuilt.com>
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab - Change ap_get_brigade prototype:
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab Message-ID: <20020120115136.GI17601@ebuilt.com>
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab * Allow the DocumentRoot directive within <Location > scopes? This
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab allows the beloved (crusty) Alias /foo/ /somepath/foo/ followed
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab by a <Directory /somepath/foo> to become simply
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab <Location /foo/> DocumentRoot /somefile/foo (IMHO a bit more legible
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab and in-your-face.) DocumentRoot unset would be accepted [and would
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab not permit content to be served, only virtual resources such as
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab server-info or server-status.
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab This proposed change would _not_ depricate Alias.
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab * Win32: Rotatelogs sometimes is not terminated when Apache
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab goes down hard. FirstBill was looking at possibly tracking the
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab child's-child processes in the parent process.
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab OtherBill asks, wasn't this fixed?
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab * Win32: Add a simple hold console open patch (wait for close or
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab the ESC key, with a nice message) if the server died a bad
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab death (non-zero exit code) in console mode.
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab Resolution: bring forward same ugly hacks from 1.3.13-.20
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab * Port of mod_ssl to Apache 2.0:
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab The current porting state is summarized in modules/ssl/README. The
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab remaining work includes:
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab (1) stablizing/optimizing the SSL filter logic
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab (2) Enabling the various SSL caching mechanisms (shmcb, shmht)
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab (3) Enabling SSL extentions
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab (4) Trying to seperate the https filter logic from mod_ssl -
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab This is to facilitate other modules that wish to use the https
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab filter or the mod_ssl logic or both as required.
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab Justin: mod_ssl filter logic is redone, so that should be fine.
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab Madhu has submitted a patch for SSL caching - however, I
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab am -0 on that patch as I *think* we could implement the
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab shared memory another way that is much cleaner (i.e.
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab treat shmem directly as a dbm via APR routines). Justin
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab also thinks that the https filter logic may be sufficiently
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab decoupled now, but isn't really sure.
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab * Performance: Get the SINGLE_LISTEN_UNSERIALIZED_ACCEPT
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab optimization working in worker. prefork's new design for how
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab to notice data on the pod should be sufficient.
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab * Performance & Debug: Eliminate most (and perhaps all) of the
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab malloc/free calls in the bucket brigade code. Need some
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab light weight memory management functions that allow freeing
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab memory (putting it back into a memory pool) when it is no
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab longer needed. Enabling simple debugging features like guard
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab bands, double free detection, etc. would be cool but certainly
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab not a hard requirement.
843e19887f64dde75055cf8842fc4db2171eff45johnlev Status: Cliff started to implement this using SMS as has
843e19887f64dde75055cf8842fc4db2171eff45johnlev been discussed at length for months, but since
843e19887f64dde75055cf8842fc4db2171eff45johnlev SMS is not being used anywhere else in the server,
843e19887f64dde75055cf8842fc4db2171eff45johnlev several people expressed the opinion that we should
843e19887f64dde75055cf8842fc4db2171eff45johnlev get rid of it entirely, meaning that the buckets
843e19887f64dde75055cf8842fc4db2171eff45johnlev need their own memory management (free list) functions.
843e19887f64dde75055cf8842fc4db2171eff45johnlev Cliff will implement that this weekend so we at least
843e19887f64dde75055cf8842fc4db2171eff45johnlev have something to look at/compare with.
843e19887f64dde75055cf8842fc4db2171eff45johnlev * Eliminate unnecessary creation of pipes in mod_cgid
843e19887f64dde75055cf8842fc4db2171eff45johnlev * the autoconf setup should be fixed to default to using the
843e19887f64dde75055cf8842fc4db2171eff45johnlev "Apache" layout from config.layout, and each variable settable
843e19887f64dde75055cf8842fc4db2171eff45johnlev in a layout should be overridable on the command line. Plus,
843e19887f64dde75055cf8842fc4db2171eff45johnlev what we do right now just doesn't seem to fully fit into how autoconf
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab works, eg. AC_PREFIX_DEFAULT issues.
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab Message-ID: <Pine.BSF.4.20.0104031557420.20876-100000@alive.znep.com>
843e19887f64dde75055cf8842fc4db2171eff45johnlev * Combine log_child and piped_log_spawn. Clean up http_log.c.
843e19887f64dde75055cf8842fc4db2171eff45johnlev Common logging API.
843e19887f64dde75055cf8842fc4db2171eff45johnlev * Document mod_file_cache.
843e19887f64dde75055cf8842fc4db2171eff45johnlev * OS/2: Make mod_status work for spmt_os2 MPM.
843e19887f64dde75055cf8842fc4db2171eff45johnlev * Platforms that do not support fork (primarily Win32 and AS/400)
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab Architect start-up code that avoids initializing all the modules
843e19887f64dde75055cf8842fc4db2171eff45johnlev in the parent process on platforms that do not support fork.
843e19887f64dde75055cf8842fc4db2171eff45johnlev * Win32: Migrate the MPM over to use APR thread/process calls. This
843e19887f64dde75055cf8842fc4db2171eff45johnlev would eliminate some code in the Win32 branch that essentially
843e19887f64dde75055cf8842fc4db2171eff45johnlev duplicates what is in APR.
843e19887f64dde75055cf8842fc4db2171eff45johnlev * There are still a number of places in the code where we are
843e19887f64dde75055cf8842fc4db2171eff45johnlev losing error status (i.e. throwing away the error returned by a
843e19887f64dde75055cf8842fc4db2171eff45johnlev system call and replacing it with a generic error code)
843e19887f64dde75055cf8842fc4db2171eff45johnlev * Mass vhosting version of suEXEC.
843e19887f64dde75055cf8842fc4db2171eff45johnlev * All DBMs suffer from confusion in support/dbmmanage (perl script) since
843e19887f64dde75055cf8842fc4db2171eff45johnlev the dbmmanage employs the first-matched dbm format. This is not
843e19887f64dde75055cf8842fc4db2171eff45johnlev necessarily the library that Apache was built with. Aught to
843e19887f64dde75055cf8842fc4db2171eff45johnlev rewrite dbmmanage upon installation to bin/ with the proper library
843e19887f64dde75055cf8842fc4db2171eff45johnlev for predictable mod_auth_db/dbm administration.
843e19887f64dde75055cf8842fc4db2171eff45johnlev Status: Mladen Turk has posted several patches and ideas.
843e19887f64dde75055cf8842fc4db2171eff45johnlev Key question, part of htpasswd, or a seperate utility?
843e19887f64dde75055cf8842fc4db2171eff45johnlev prefer htpasswd: Lars
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab prefer seperate: OtherBill
843e19887f64dde75055cf8842fc4db2171eff45johnlev * use apu_dbm in mod_auth_dbm
843e19887f64dde75055cf8842fc4db2171eff45johnlev Status: Greg +1 (low-priority volunteer)
843e19887f64dde75055cf8842fc4db2171eff45johnlev Justin says: "Seems like this is already there, so should we just
843e19887f64dde75055cf8842fc4db2171eff45johnlev remove the other DBM code in that file? If you want
843e19887f64dde75055cf8842fc4db2171eff45johnlev to use gdbm, or dbm, etc, you should tell apr-util."
843e19887f64dde75055cf8842fc4db2171eff45johnlev Will says: "bs - I may choose the fastest - most efficient native
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab dbm implementation, for shared proc caches, ssl session
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab caching, etc, but that has nothing to do with maintaining
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab a userlist via dbm, which has to remain readable between
843e19887f64dde75055cf8842fc4db2171eff45johnlev builds/machines, etc. The use-multiple database schema
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab for apr-util would let us do this with just apr, though."
843e19887f64dde75055cf8842fc4db2171eff45johnlev Ian says: "multi-dbm is in, but it still has ndbm support hardcoded
843e19887f64dde75055cf8842fc4db2171eff45johnlev is this still required? isn't ndbm supported via gdbm?"
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab * Integrate mod_dav.
843e19887f64dde75055cf8842fc4db2171eff45johnlev Some additional items remaining:
843e19887f64dde75055cf8842fc4db2171eff45johnlev - case_preserved_filename stuff
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab (use the new canonical name stuff?)
843e19887f64dde75055cf8842fc4db2171eff45johnlev - find a new home for ap_text(_header)
843e19887f64dde75055cf8842fc4db2171eff45johnlev - is it possible to remove the DAV: namespace stuff from util_xml?
843e19887f64dde75055cf8842fc4db2171eff45johnlev * ap_core_translate() and its use by mod_mmap_static and mod_file_cache
843e19887f64dde75055cf8842fc4db2171eff45johnlev are a bit wonky. The function should probably be exposed as a utility
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab function (such as ap_translate_url2fs() or ap_validate_fs_url() or
843e19887f64dde75055cf8842fc4db2171eff45johnlev something). Another approach would be a new hook phase after
843e19887f64dde75055cf8842fc4db2171eff45johnlev "translate" which would allow the module to munge what the
843e19887f64dde75055cf8842fc4db2171eff45johnlev translation has decided to do.
843e19887f64dde75055cf8842fc4db2171eff45johnlev Status: Greg +1 (volunteers), Ryan +1
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab * Explore use of a post-config hook for the code in http_main.c which
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab calls ap_fixup_virutal_hosts(), ap_fini_vhost_config(), and
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab ap_sort_hooks() [to reduce the logic in main()]
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab * read the config tree just once, and process N times (as necessary)
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab * (possibly) use UUIDs in mod_unique_id and/or mod_usertrack
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab * (possibly) port the bug fix for PR 6942 (segv when LoadModule is put
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab into a VirtualHost container) to 2.0.
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab * shift stuff to mod_core.h
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab * callers of ap_run_create_request() should check the return value
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab for failure (Doug volunteers)
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab * Win32: Get Apache working on Windows 95/98. The following work
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab (at least) needs to be done:
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab - winnt MPM: Fix 95/98 code paths in the winnt MPM. There is some NT
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab specific code that is still not in NT only code paths
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab - IOL binds to APR sendfile, implemented with TransmitFile, which
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab is not available on 95/98.
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab - Document warning that OSR2 is required (for Crypt functions, in
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab rand.c, at least.) This could be resolved with an SSL library, or
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab randomization in APR itself.
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab - Bring the Win9xConHook.dll from 1.3 into 2.0 (no sense till it
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab actually works) and add in a splash of Win9x service code.
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab * In order to use a DSO version of mod_ssl we have to link with
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab -lssl and -lcrypto. A workaround is in place right now where the
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab entire EXTRA_LIBS macro is being appended to the objects list, but
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab this is a hack. We should either revamp the APACHE_CHECK_SSL_TOOLKIT
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab autoconf function or come up with some other autoconf checks to
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab search for libssl and libcrypto and properly add them to mod_ssl's
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab link flags.
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab * Make the worker MPM the default MPM for threaded Unix boxes.
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab +1: Justin, Jeff, Ian
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab -0: Aaron (premature decision, needs more discussion), Lars
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab -0: Cliff (I think the default config should be the safest possible)
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab * Fix the worker MPM to use POD to kill child processes instead
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab of ap_os_killpg, regardless of how they should die. (Ryan Bloom)
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rabPRs that have been suspended forever waiting for someone to
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rabput them into 'the next release':
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab * PR#76: general
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab missing call to "setlocale();"
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab * PR#78: mod_include
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab Additional status for XBitHack directive
5d2eda970e48f8985448151c73e699614ce9f357John Levon * PR#362: mod_proxy
5d2eda970e48f8985448151c73e699614ce9f357John Levon Mod_proxy doesn't allow change of error pages
5d2eda970e48f8985448151c73e699614ce9f357John Levon * PR#370: mod_env
5d2eda970e48f8985448151c73e699614ce9f357John Levon Modified PATH environemnt variable is not passed, instead
5d2eda970e48f8985448151c73e699614ce9f357John Levon system's is used
5d2eda970e48f8985448151c73e699614ce9f357John Levon * PR#440: mod_proxy
5d2eda970e48f8985448151c73e699614ce9f357John Levon Proxy doesn't deliver documents if not connected
5d2eda970e48f8985448151c73e699614ce9f357John Levon * PR#534: mod_proxy
5d2eda970e48f8985448151c73e699614ce9f357John Levon proxy converts ~name to %7Ename when name starts with a dot (.)
5d2eda970e48f8985448151c73e699614ce9f357John Levon * PR#537: mod_access
5d2eda970e48f8985448151c73e699614ce9f357John Levon mod_access syntax allows hosts that should be restricted
5d2eda970e48f8985448151c73e699614ce9f357John Levon * PR#557: mod_auth-any
5d2eda970e48f8985448151c73e699614ce9f357John Levon ~UserHome directories are not honored in absolute pathname
5d2eda970e48f8985448151c73e699614ce9f357John Levon requests (.htaccess)
5d2eda970e48f8985448151c73e699614ce9f357John Levon * PR#612: mod_proxy
5d2eda970e48f8985448151c73e699614ce9f357John Levon Proxy FTP Authentication Fails
5d2eda970e48f8985448151c73e699614ce9f357John Levon * PR#623: mod_include
5d2eda970e48f8985448151c73e699614ce9f357John Levon A smarter "Last Modified" value for SSI documents (see PR number 600)
5d2eda970e48f8985448151c73e699614ce9f357John Levon * PR#628: config
5d2eda970e48f8985448151c73e699614ce9f357John Levon Request of "Options SymLinksIfGroupMatch"
5d2eda970e48f8985448151c73e699614ce9f357John Levon * PR#700: mod_proxy
5d2eda970e48f8985448151c73e699614ce9f357John Levon Proxy doesn't do links right for OpenVMS files through ftp:
5d2eda970e48f8985448151c73e699614ce9f357John Levon * PR#759: mod_imap
5d2eda970e48f8985448151c73e699614ce9f357John Levon imap should read <MAP><AREA>*</MAP> too!
5d2eda970e48f8985448151c73e699614ce9f357John Levon * PR#793: general
5d2eda970e48f8985448151c73e699614ce9f357John Levon RLimitCPU and RLimitMEM don't apply to all children like they should
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#921: suexec
843e19887f64dde75055cf8842fc4db2171eff45johnlev Uses cwd before filling it in, doesn't use syslog
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#922: config
843e19887f64dde75055cf8842fc4db2171eff45johnlev it is useful to allow specifiction that root-owned symlinks
843e19887f64dde75055cf8842fc4db2171eff45johnlev should always be followed
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#980: mod_proxy
843e19887f64dde75055cf8842fc4db2171eff45johnlev Controlling Access to Remote Proxies would be nice...
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#994: mod_proxy
843e19887f64dde75055cf8842fc4db2171eff45johnlev Adding authentication "on the fly" through the proxy module
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#1004: apache-api
843e19887f64dde75055cf8842fc4db2171eff45johnlev request_config field in request_rec is moderately bogus
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#1028: other
843e19887f64dde75055cf8842fc4db2171eff45johnlev DoS attacks involving memory consumption
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#1050: mod_log-any
5d2eda970e48f8985448151c73e699614ce9f357John Levon Logging of virtual server to error_log as well
5d2eda970e48f8985448151c73e699614ce9f357John Levon * PR#1085: mod_proxy
5d2eda970e48f8985448151c73e699614ce9f357John Levon ProxyRemote make a dead cycle.
5d2eda970e48f8985448151c73e699614ce9f357John Levon * PR#1117: mod_auth-any
5d2eda970e48f8985448151c73e699614ce9f357John Levon Using NIS passwd.byname dbm files with AuthDBMUserFile
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#1120: suexec
843e19887f64dde75055cf8842fc4db2171eff45johnlev suexec does not parse arguments to #exec cmd
5d2eda970e48f8985448151c73e699614ce9f357John Levon * PR#1145: mod_include
843e19887f64dde75055cf8842fc4db2171eff45johnlev Allow for Last-Modified: without resorting to XBitHack
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#1158: apache-api
843e19887f64dde75055cf8842fc4db2171eff45johnlev improvements to child spawning API
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#1166: mod_proxy
843e19887f64dde75055cf8842fc4db2171eff45johnlev ``nph-'' not honored (no buffering) for ProxyRemote mapping
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#1176: mod_cgi
843e19887f64dde75055cf8842fc4db2171eff45johnlev Apache cannot handle continuation line in headers
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#1191: general
843e19887f64dde75055cf8842fc4db2171eff45johnlev setlogin() is not called, causing problems with e.g. identd
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#1204: general
843e19887f64dde75055cf8842fc4db2171eff45johnlev regerror() exists, use it
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#1233: apache-api
843e19887f64dde75055cf8842fc4db2171eff45johnlev there is no way to keep per-connection per-module state
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#1263: mod_autoexec
843e19887f64dde75055cf8842fc4db2171eff45johnlev Add frame-safe anchor attribute to mod_autoindex links
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#1268: suexec
843e19887f64dde75055cf8842fc4db2171eff45johnlev CGI scripts running as Apache user: security (suexec etc.)
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#1285: suexec
843e19887f64dde75055cf8842fc4db2171eff45johnlev Error messages could be easier to spot in cgi.log file for suexec.c
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#1287: mod_access
843e19887f64dde75055cf8842fc4db2171eff45johnlev add allow,deny/deny,allow warning to mod_access
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#1290: mod_proxy
843e19887f64dde75055cf8842fc4db2171eff45johnlev Need to know "hit-rate" on proxy cache
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#1358: mod_log-any
843e19887f64dde75055cf8842fc4db2171eff45johnlev Selective url-encode of log fields (or maybe a pseudo
843e19887f64dde75055cf8842fc4db2171eff45johnlev log_rewrite module?)
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#1383: mod_headers
843e19887f64dde75055cf8842fc4db2171eff45johnlev I make mod_headers to modify request headers as well as
843e19887f64dde75055cf8842fc4db2171eff45johnlev response ones.
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#1532: mod_proxy
843e19887f64dde75055cf8842fc4db2171eff45johnlev Proxy transfer logging
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#1547: mod_proxy
843e19887f64dde75055cf8842fc4db2171eff45johnlev No HTTP_X_FORWARDED_FOR set...
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#1567: mod_proxy
843e19887f64dde75055cf8842fc4db2171eff45johnlev ProxyRemote proxy requests fail authentication by firewall
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#1582: mod_rewrite
843e19887f64dde75055cf8842fc4db2171eff45johnlev mod_rewrite forms REQUEST_URI different than mod_cgi does
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#1677: mod_headers
843e19887f64dde75055cf8842fc4db2171eff45johnlev mod_headers should allow mod_log_config-style formats in
843e19887f64dde75055cf8842fc4db2171eff45johnlev header values
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#1702: mod_proxy
843e19887f64dde75055cf8842fc4db2171eff45johnlev mod_proxy to support persistent conns?
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#1803: mod_include
843e19887f64dde75055cf8842fc4db2171eff45johnlev patches to mod_include to allow for file tests
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#1809: mod_auth-any
843e19887f64dde75055cf8842fc4db2171eff45johnlev Suggestion for improving authentication modules and core source
843e19887f64dde75055cf8842fc4db2171eff45johnlev code, problem with 401 and ErrorDocument
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#1878: mod_proxy
843e19887f64dde75055cf8842fc4db2171eff45johnlev listing of proxy cache content
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#1905: suexec
843e19887f64dde75055cf8842fc4db2171eff45johnlev Allow modules to set user:group for execution.
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#2024: apache-api
843e19887f64dde75055cf8842fc4db2171eff45johnlev adding auth_why to conn_rec
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#2073: mod_log-any
843e19887f64dde75055cf8842fc4db2171eff45johnlev pipelined connections are not logged correctly
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab * PR#2074: mod_rewrite
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab mod_rewrite doesn't pass Proxy Throughput on internal subrequests
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab * PR#2113: config
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab HTTP Server Rebuild Line Needs Changing for the better
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab * PR#2138: mod_status
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab mod_status always displays 256 possible connection slots
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#2221: documentation
843e19887f64dde75055cf8842fc4db2171eff45johnlev Make online documentation search link back to my installation
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#2284: general
843e19887f64dde75055cf8842fc4db2171eff45johnlev Can not POST to ErrorDocument - Apache/1.3b6
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#2314: mod_proxy
843e19887f64dde75055cf8842fc4db2171eff45johnlev patterns in ProxyRemote
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#2343: mod_status
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab Status module averages are for entire uptime
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#2360: suexec
843e19887f64dde75055cf8842fc4db2171eff45johnlev suexec for general access of user content?
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#2396: general
843e19887f64dde75055cf8842fc4db2171eff45johnlev Proposal for TimeZone directive
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#2415: mod_info
843e19887f64dde75055cf8842fc4db2171eff45johnlev /server-info doesn't check for the virtual host to list the info
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#2421: config
843e19887f64dde75055cf8842fc4db2171eff45johnlev problem specifying ndbm library for build ?with autoconfigure
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#2431: general
843e19887f64dde75055cf8842fc4db2171eff45johnlev A small addition to rotatelogs.c to improve program functionality.
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#2446: config
843e19887f64dde75055cf8842fc4db2171eff45johnlev AllowOverride FileInfo is too coarse
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#2460: mod_cgi
843e19887f64dde75055cf8842fc4db2171eff45johnlev TimeOut applies to output of CGI scripts
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#2512: mod_access
843e19887f64dde75055cf8842fc4db2171eff45johnlev <IfDenied> directive wanted
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#2573: suexec
843e19887f64dde75055cf8842fc4db2171eff45johnlev CGI's for general use still have to be run as another user
843e19887f64dde75055cf8842fc4db2171eff45johnlev with suExec
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#2648: general
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab Cache file names in Proxy module
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#2760: config
843e19887f64dde75055cf8842fc4db2171eff45johnlev [PATCH] User/Group for <Directory> and <Location> i.e. not only
843e19887f64dde75055cf8842fc4db2171eff45johnlev in global and <Virtual>.
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#2763: general
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab mailto tags and bundling bug report script
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab * PR#2785: os-aix
a576ab5b6e08c47732b3dedca9eaa8a8cbb85720rab Support for System Resource Controller
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#2793: protocol
843e19887f64dde75055cf8842fc4db2171eff45johnlev When will Apache support P3P? Any Plans?
5d2eda970e48f8985448151c73e699614ce9f357John Levon * PR#2873: config
5d2eda970e48f8985448151c73e699614ce9f357John Levon * PR#2889: general
5d2eda970e48f8985448151c73e699614ce9f357John Levon Inclusion of RPM spec file in CVS/distributions
5d2eda970e48f8985448151c73e699614ce9f357John Levon * PR#2906: general
5d2eda970e48f8985448151c73e699614ce9f357John Levon Propose that Apache recommend $UNIQUE_ID for all "session id"
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#2907: config
843e19887f64dde75055cf8842fc4db2171eff45johnlev suggestion: power up your Include directive :)
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#3018: general
843e19887f64dde75055cf8842fc4db2171eff45johnlev cannot limit some HTTP methods
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#3143: apache-api
5d2eda970e48f8985448151c73e699614ce9f357John Levon No module specific data hook for per-connection data
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#3191: mod_negotiation
843e19887f64dde75055cf8842fc4db2171eff45johnlev no way to set global quality-of-source (qs) coneg values
843e19887f64dde75055cf8842fc4db2171eff45johnlev with multiviews
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#3568: mod_proxy
843e19887f64dde75055cf8842fc4db2171eff45johnlev Accessing URL through proxy server corrupts data.
5d2eda970e48f8985448151c73e699614ce9f357John Levon * PR#3605: mod_proxy
843e19887f64dde75055cf8842fc4db2171eff45johnlev Some anonymous FTP URLs ask for authentication
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#3677: general
843e19887f64dde75055cf8842fc4db2171eff45johnlev New ErrorDocumentMatch directive
5d2eda970e48f8985448151c73e699614ce9f357John Levon * PR#4241: config
5d2eda970e48f8985448151c73e699614ce9f357John Levon Need to be able to override shebang line to make CGI scripts
5d2eda970e48f8985448151c73e699614ce9f357John Levon more portable.
5d2eda970e48f8985448151c73e699614ce9f357John Levon * PR#4244: config
5d2eda970e48f8985448151c73e699614ce9f357John Levon "Files" and "FilesMatch" regexp does not recognize bang as
5d2eda970e48f8985448151c73e699614ce9f357John Levon negation operator
5d2eda970e48f8985448151c73e699614ce9f357John Levon * PR#4448: mod_log-any
843e19887f64dde75055cf8842fc4db2171eff45johnlev Please allow CGI env variables (QUERY_STRING, ...) to be logged
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#4459: mod_include
843e19887f64dde75055cf8842fc4db2171eff45johnlev Suggestion for better handling of Last-modified headers
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#4490: mod_cgi
843e19887f64dde75055cf8842fc4db2171eff45johnlev mod_cgi prevents handling of OPTIONS requests
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#5713: os-windows
843e19887f64dde75055cf8842fc4db2171eff45johnlev [PATCH] install as win32 service with domain account
843e19887f64dde75055cf8842fc4db2171eff45johnlev * PR#5993: general
843e19887f64dde75055cf8842fc4db2171eff45johnlev AllowOverride should have a 'CheckNone' and 'AllowNone' argument
843e19887f64dde75055cf8842fc4db2171eff45johnlev instead of only 'None'
843e19887f64dde75055cf8842fc4db2171eff45johnlevOther bugs that need fixing:
843e19887f64dde75055cf8842fc4db2171eff45johnlev * MaxRequestsPerChild measures connections, not requests.
843e19887f64dde75055cf8842fc4db2171eff45johnlev Until someone has a better way, we'll probably just rename it
843e19887f64dde75055cf8842fc4db2171eff45johnlev "MaxConnectionsPerChild".
843e19887f64dde75055cf8842fc4db2171eff45johnlev * Regex containers don't work in an intutive way
843e19887f64dde75055cf8842fc4db2171eff45johnlev Status: No one has come up with an efficient way to fix this
843e19887f64dde75055cf8842fc4db2171eff45johnlev behavior. Dean has suggested getting rid of regex containers
843e19887f64dde75055cf8842fc4db2171eff45johnlev completely.
843e19887f64dde75055cf8842fc4db2171eff45johnlev * SIGSEGV on Linux (glibc 2.1.2) isn't caught properly by a
843e19887f64dde75055cf8842fc4db2171eff45johnlev sigwaiting thread. We need to work around this, perhaps unless
843e19887f64dde75055cf8842fc4db2171eff45johnlev there is hope soon for a fixed glibc.
843e19887f64dde75055cf8842fc4db2171eff45johnlev * orig_ct in the byterange/multipart handling may not be
843e19887f64dde75055cf8842fc4db2171eff45johnlev needed. Apache 1.3 just never stashed "multipart" into
843e19887f64dde75055cf8842fc4db2171eff45johnlev r->content_type. We should probably follow suit since the
843e19887f64dde75055cf8842fc4db2171eff45johnlev byterange stuff doesn't want the rest of the code to see the
843e19887f64dde75055cf8842fc4db2171eff45johnlev multipart content-type; the other code should still think it is
843e19887f64dde75055cf8842fc4db2171eff45johnlev dealing with the <orig_ct> stuff.
843e19887f64dde75055cf8842fc4db2171eff45johnlev Status: Greg volunteers to investigate (esp. since he was most
843e19887f64dde75055cf8842fc4db2171eff45johnlev likely the one to break it :-)
843e19887f64dde75055cf8842fc4db2171eff45johnlevOther features that need writing:
843e19887f64dde75055cf8842fc4db2171eff45johnlev * Finish infrastructure in core for async MPMs
843e19887f64dde75055cf8842fc4db2171eff45johnlev Status: post 2.0
843e19887f64dde75055cf8842fc4db2171eff45johnlev * TODO in source -- just do an egrep on "TODO" and see what's there
843e19887f64dde75055cf8842fc4db2171eff45johnlevAvailable Patches:
843e19887f64dde75055cf8842fc4db2171eff45johnlev * Jon Travis's <jtravis@covalent.net> patch to deal with thread-safe
843e19887f64dde75055cf8842fc4db2171eff45johnlev issues with inet_ntoa. See message <20001201163220.A12827@covalent.net>
843e19887f64dde75055cf8842fc4db2171eff45johnlev Status: This is being set aside until the IPv6 work is finished
843e19887f64dde75055cf8842fc4db2171eff45johnlev so that we know exactly what is required.
843e19887f64dde75055cf8842fc4db2171eff45johnlev * Martin Sojka <msojka@gmx.de>'s patch to add error reporting for failed
843e19887f64dde75055cf8842fc4db2171eff45johnlev htpasswd actions due to a full /tmp volume (other programs may have
843e19887f64dde75055cf8842fc4db2171eff45johnlev similar problems?)
843e19887f64dde75055cf8842fc4db2171eff45johnlev * Mike Abbott's <mja@trudge.engr.sgi.com> patches to improve
843e19887f64dde75055cf8842fc4db2171eff45johnlev performance
843e19887f64dde75055cf8842fc4db2171eff45johnlev Status: These were written for 1.3, and are awaiting a port to
843e19887f64dde75055cf8842fc4db2171eff45johnlev * Jim Winstead's <jimw@trainedmonkey.com> patch to add CookieDomain and
843e19887f64dde75055cf8842fc4db2171eff45johnlev other small mod_usertrack features
843e19887f64dde75055cf8842fc4db2171eff45johnlev * Dan Rench's <drench@xnet.com> patch to add allow the errmsg and timefmt
843e19887f64dde75055cf8842fc4db2171eff45johnlev of SSI's to be modified in the config file. Patch is available in
843e19887f64dde75055cf8842fc4db2171eff45johnlevOpen issues:
843e19887f64dde75055cf8842fc4db2171eff45johnlev * Which MPMs will be included with Apache 2.0?