STATUS revision 45740d5cd30d22f067c05b95f04f866abd7ebed8
355b4669e025ff377602b6fc7caaf30dbc218371jacobsAPACHE 2.0 STATUS: -*-text-*-
355b4669e025ff377602b6fc7caaf30dbc218371jacobsLast modified at [$Date: 2002/06/11 05:33:07 $]
355b4669e025ff377602b6fc7caaf30dbc218371jacobs 2.0.38 : in development.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs 2.0.37 : tagged June 11, 2002.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs 2.0.36 : released May 6, 2002 as GA.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs 2.0.35 : released April 5, 2002 as GA.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs 2.0.34 : tagged March 26, 2002.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs 2.0.33 : tagged March 6, 2002. not released.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs 2.0.32 : released Feburary 16, 2002 as beta.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs 2.0.31 : rolled Feburary 1, 2002. not released.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs 2.0.30 : tagged January 8, 2002. not rolled.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs 2.0.29 : tagged November 27, 2001. not rolled.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs 2.0.28 : released November 13, 2001 as beta.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs 2.0.27 : rolled November 6, 2001
355b4669e025ff377602b6fc7caaf30dbc218371jacobs 2.0.26 : tagged October 16, 2001. not rolled.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs 2.0.25 : rolled August 29, 2001
355b4669e025ff377602b6fc7caaf30dbc218371jacobs 2.0.24 : rolled August 18, 2001
355b4669e025ff377602b6fc7caaf30dbc218371jacobs 2.0.23 : rolled August 9, 2001
355b4669e025ff377602b6fc7caaf30dbc218371jacobs 2.0.22 : rolled July 29, 2001
355b4669e025ff377602b6fc7caaf30dbc218371jacobs 2.0.21 : rolled July 20, 2001
355b4669e025ff377602b6fc7caaf30dbc218371jacobs 2.0.20 : rolled July 8, 2001
355b4669e025ff377602b6fc7caaf30dbc218371jacobs 2.0.19 : rolled June 27, 2001
355b4669e025ff377602b6fc7caaf30dbc218371jacobs 2.0.18 : rolled May 18, 2001
355b4669e025ff377602b6fc7caaf30dbc218371jacobs 2.0.17 : rolled April 17, 2001
355b4669e025ff377602b6fc7caaf30dbc218371jacobs 2.0.16 : rolled April 4, 2001
355b4669e025ff377602b6fc7caaf30dbc218371jacobs 2.0.15 : rolled March 21, 2001
355b4669e025ff377602b6fc7caaf30dbc218371jacobs 2.0.14 : rolled March 7, 2001
355b4669e025ff377602b6fc7caaf30dbc218371jacobs 2.0a9 : released December 12, 2000
355b4669e025ff377602b6fc7caaf30dbc218371jacobs 2.0a8 : released November 20, 2000
355b4669e025ff377602b6fc7caaf30dbc218371jacobs 2.0a7 : released October 8, 2000
355b4669e025ff377602b6fc7caaf30dbc218371jacobs 2.0a6 : released August 18, 2000
355b4669e025ff377602b6fc7caaf30dbc218371jacobs 2.0a5 : released August 4, 2000
355b4669e025ff377602b6fc7caaf30dbc218371jacobs 2.0a4 : released June 7, 2000
355b4669e025ff377602b6fc7caaf30dbc218371jacobs 2.0a3 : released April 28, 2000
355b4669e025ff377602b6fc7caaf30dbc218371jacobs 2.0a2 : released March 31, 2000
355b4669e025ff377602b6fc7caaf30dbc218371jacobs 2.0a1 : released March 10, 2000
355b4669e025ff377602b6fc7caaf30dbc218371jacobsPlease consult the following STATUS files for information
355b4669e025ff377602b6fc7caaf30dbc218371jacobson related projects:
355b4669e025ff377602b6fc7caaf30dbc218371jacobsCURRENT RELEASE NOTES:
355b4669e025ff377602b6fc7caaf30dbc218371jacobsRELEASE SHOWSTOPPERS:
355b4669e025ff377602b6fc7caaf30dbc218371jacobsCURRENT VOTES:
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * apachectl should revert to just being an init script and
355b4669e025ff377602b6fc7caaf30dbc218371jacobs httpd.sh should be the wrapper for httpd which sources envvars
355b4669e025ff377602b6fc7caaf30dbc218371jacobs and allows any options to be passed through
355b4669e025ff377602b6fc7caaf30dbc218371jacobs +1: trawick
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * Should we always build [support*] binaries statically unless otherwise
355b4669e025ff377602b6fc7caaf30dbc218371jacobs Message-ID: <20020129210006.B23512@Lithium.MeepZor.Com>
355b4669e025ff377602b6fc7caaf30dbc218371jacobs +1: Ken, *wrowe [they are PITAs on OSX]
355b4669e025ff377602b6fc7caaf30dbc218371jacobs -1: Justin, Ian
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * If the parent process dies, should the remaining child processes
355b4669e025ff377602b6fc7caaf30dbc218371jacobs "gracefully" self-terminate. Or maybe we should make it a runtime
355b4669e025ff377602b6fc7caaf30dbc218371jacobs option, or have a concept of 2 parent processes (one being a
355b4669e025ff377602b6fc7caaf30dbc218371jacobs "hot spare").
355b4669e025ff377602b6fc7caaf30dbc218371jacobs See: Message-ID: <3C58232C.FE91F19F@Golux.Com>
355b4669e025ff377602b6fc7caaf30dbc218371jacobs Self-destruct: Ken, Martin
355b4669e025ff377602b6fc7caaf30dbc218371jacobs Not self-destruct: BrianP, Ian, Cliff, BillS
355b4669e025ff377602b6fc7caaf30dbc218371jacobs Make it runtime configurable: Aaron, Jim, Justin
355b4669e025ff377602b6fc7caaf30dbc218371jacobs Have 2 parents: +1: Jim
355b4669e025ff377602b6fc7caaf30dbc218371jacobs -1: Justin, wrowe [for 2.0]
355b4669e025ff377602b6fc7caaf30dbc218371jacobs +0: Martin (while standing by, could it do
355b4669e025ff377602b6fc7caaf30dbc218371jacobs something useful?)
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * Make the worker MPM the default MPM for threaded Unix boxes.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs +1: Justin, Ian, Cliff, BillS
355b4669e025ff377602b6fc7caaf30dbc218371jacobs +0: BrianP, Aaron (mutex contention is looking better with the
355b4669e025ff377602b6fc7caaf30dbc218371jacobs latest code, let's continue tuning and testing)
355b4669e025ff377602b6fc7caaf30dbc218371jacobsRELEASE NON-SHOWSTOPPERS BUT WOULD BE REAL NICE TO WRAP THESE UP:
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * Returning invalid 304 errors when filters are present. [PR 9673]
355b4669e025ff377602b6fc7caaf30dbc218371jacobs Message-ID: <Pine.GSO.4.44.0206061713320.3848-100000@lab26.lacfas.hec.ca>
355b4669e025ff377602b6fc7caaf30dbc218371jacobs Status: We should set r->no_local_copy, but where can we set it?
355b4669e025ff377602b6fc7caaf30dbc218371jacobs We must set it before the handlers are run so that the
355b4669e025ff377602b6fc7caaf30dbc218371jacobs appropriate handler can call ap_meets_conditions() before
355b4669e025ff377602b6fc7caaf30dbc218371jacobs generating the data.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * server pushed CGI's not working. (Is this a showstopper??)
355b4669e025ff377602b6fc7caaf30dbc218371jacobs Message-ID: <3CE15B85.2FF45121@apache.org>
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * HP/UX 10.20: compile breakage in APR. Looks like it should be easy
355b4669e025ff377602b6fc7caaf30dbc218371jacobs to fix, probably just some extraneous #include's that are fouling
355b4669e025ff377602b6fc7caaf30dbc218371jacobs Jeff: See my reply and patch in the PR (and previous commit to
355b4669e025ff377602b6fc7caaf30dbc218371jacobs stop using "pipe" as a field name). If patch is committed, we
355b4669e025ff377602b6fc7caaf30dbc218371jacobs should be okay. I'll wait to see if the user tests the patch.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * exec cmd and suexec arg-passing enhancements
355b4669e025ff377602b6fc7caaf30dbc218371jacobs Status: Patches proposed
355b4669e025ff377602b6fc7caaf30dbc218371jacobs Message-ID: <20020526041748.A29148@prodigy.Redbrick.DCU.IE>
355b4669e025ff377602b6fc7caaf30dbc218371jacobs (see the "proc.patch" and "suexec-shell.patch" links in this message)
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * Get mod_cache/mod_mem_cache out of experimental (still some
355b4669e025ff377602b6fc7caaf30dbc218371jacobs work items left to complete)
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * The 2.0.36 worker MPM graceless shutdown changes work but are
355b4669e025ff377602b6fc7caaf30dbc218371jacobs a bit clunky on some platforms; eg, on Linux, the loop to
355b4669e025ff377602b6fc7caaf30dbc218371jacobs join each worker thread seems to hang, and the parent ends up
355b4669e025ff377602b6fc7caaf30dbc218371jacobs killing off the child with SIGKILL. But at least it shuts down.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * --enable-mods-shared="foo1 foo2" is busted on Darwin. Pier
355b4669e025ff377602b6fc7caaf30dbc218371jacobs posted a patch (Message-ID: <B8DBBE8D.575A%pier@betaversion.org>).
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * We do not properly substitute the prefix-variables in the configuration
355b4669e025ff377602b6fc7caaf30dbc218371jacobs scripts or generated-configs. (i.e. if sysconfdir is etc,
355b4669e025ff377602b6fc7caaf30dbc218371jacobs httpd-std.conf points to conf.)
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * If any request gets through ap_process_request_internal() and is
355b4669e025ff377602b6fc7caaf30dbc218371jacobs scheduled to be served by the core handler, without a flag that this
355b4669e025ff377602b6fc7caaf30dbc218371jacobs r->filename was tested by dir/file_walk, we need to 500 at the very
355b4669e025ff377602b6fc7caaf30dbc218371jacobs end of the ap_process_request_internal() processing so sub_req-esters
355b4669e025ff377602b6fc7caaf30dbc218371jacobs know this request cannot be run. This provides authors of older
355b4669e025ff377602b6fc7caaf30dbc218371jacobs modules better compatibility, while still improving the security and
355b4669e025ff377602b6fc7caaf30dbc218371jacobs robustness of 2.0.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs Status: still need to decide where this goes, OtherBill comments...
355b4669e025ff377602b6fc7caaf30dbc218371jacobs Message-ID: <065701c14526$495203b0$96c0b0d0@roweclan.net>
355b4669e025ff377602b6fc7caaf30dbc218371jacobs [Deleted comments regarding the ap_run_handler phase, as irrelevant
355b4669e025ff377602b6fc7caaf30dbc218371jacobs as BillS points out that "common case will be caught in
355b4669e025ff377602b6fc7caaf30dbc218371jacobs default_handler already (with the r->finfo.filetype == 0 check)"
355b4669e025ff377602b6fc7caaf30dbc218371jacobs and the issue is detecting this -before- we try to run the req.]
355b4669e025ff377602b6fc7caaf30dbc218371jacobs gregames says: can this happen somehow without a broken module
355b4669e025ff377602b6fc7caaf30dbc218371jacobs being involved? If not, why waste cycles trying to defend against
355b4669e025ff377602b6fc7caaf30dbc218371jacobs potential broken modules? It seems futile.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs wrowe counters: no, it shouldn't happen unless the module is broken.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs But the right answer is to fail the request up-front in dir/file
355b4669e025ff377602b6fc7caaf30dbc218371jacobs walk if the path was entirely invalid; and we can't do that either
355b4669e025ff377602b6fc7caaf30dbc218371jacobs or we break modules that are unwilling to hook map_to_storage.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * Rewrite core_output_filter. It is nearly impossible to support
355b4669e025ff377602b6fc7caaf30dbc218371jacobs it with predictable results as it is implemented now.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * With AP_MODE_EXHAUSTIVE in the core, it is finally clear to me
355b4669e025ff377602b6fc7caaf30dbc218371jacobs how the Perchild MPM should be re-written. It hasn't worked
355b4669e025ff377602b6fc7caaf30dbc218371jacobs correctly since filters were added because it wasn't possible to
355b4669e025ff377602b6fc7caaf30dbc218371jacobs get the content that had already been written and the socket at
355b4669e025ff377602b6fc7caaf30dbc218371jacobs the same time. This mode lets us do that, so the MPM can be
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * htpasswd blindly processes the file you give it, and does no
355b4669e025ff377602b6fc7caaf30dbc218371jacobs sanity checking before totally corrupting whatever file it was
355b4669e025ff377602b6fc7caaf30dbc218371jacobs you thought you had. It should check the input file and bail
355b4669e025ff377602b6fc7caaf30dbc218371jacobs if it finds non-comment lines that do not contain exactly 1
355b4669e025ff377602b6fc7caaf30dbc218371jacobs ':' character.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs Message-ID: <20020217150457.A31632@clove.org>
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * Can a static httpd be built reliably?
355b4669e025ff377602b6fc7caaf30dbc218371jacobs Message-ID: <20020207142751.T31582@clove.org>
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * [Ken] Test suite failures:
355b4669e025ff377602b6fc7caaf30dbc218371jacobs o worker is also failing some of the 'cgi' subtests
355b4669e025ff377602b6fc7caaf30dbc218371jacobs (see <URL:http://Source-Zone.Org/Apache/regression/>):
355b4669e025ff377602b6fc7caaf30dbc218371jacobs Justin says: "Worker should be fine and passes httpd-test here.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs If you can provide evidence that it can be reproduced
355b4669e025ff377602b6fc7caaf30dbc218371jacobs outside of httpd-test, then it's a showstopper. I
355b4669e025ff377602b6fc7caaf30dbc218371jacobs think it's a perl or a httpd-test problem."
355b4669e025ff377602b6fc7caaf30dbc218371jacobs Not a showstopper: Justin
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * Usage of APR_BRIGADE_NORMALIZE in core_input_filter should be
355b4669e025ff377602b6fc7caaf30dbc218371jacobs removed if possible.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs Message-ID: <Pine.LNX.4.33.0201202232430.318-100000@deepthought.cs.virginia.edu>
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * There is a bug in how we sort some hooks, at least the pre-config
355b4669e025ff377602b6fc7caaf30dbc218371jacobs hook. The first time we call the hooks, they are in the correct
355b4669e025ff377602b6fc7caaf30dbc218371jacobs order, but the second time, we don't sort them correctly. Currently,
355b4669e025ff377602b6fc7caaf30dbc218371jacobs the modules/http/config.m4 file has been renamed to
355b4669e025ff377602b6fc7caaf30dbc218371jacobs modules/http/config2.m4 to work around this problem, it should moved
355b4669e025ff377602b6fc7caaf30dbc218371jacobs back when this is fixed. rbb
355b4669e025ff377602b6fc7caaf30dbc218371jacobs Justin says: "Is this really a showstopper? This has been here
355b4669e025ff377602b6fc7caaf30dbc218371jacobs forever. What's wrong? Does this have to do with
355b4669e025ff377602b6fc7caaf30dbc218371jacobs autoconf or m4?"
355b4669e025ff377602b6fc7caaf30dbc218371jacobs Not a showstopper: Justin, BrianP, trawick, gregames
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * The Add...Filter and Set...Filter directives do not allow the
355b4669e025ff377602b6fc7caaf30dbc218371jacobs administrator to order filters, beyond the order of filename (mime)
355b4669e025ff377602b6fc7caaf30dbc218371jacobs extensions. It isn't clear if Set...Filter(s) should be inserted
355b4669e025ff377602b6fc7caaf30dbc218371jacobs before or after the Add...Filter(s) which are ordered by sequence of
355b4669e025ff377602b6fc7caaf30dbc218371jacobs filename extensions. At minimum, some sort of +-[0-10] syntax seems
355b4669e025ff377602b6fc7caaf30dbc218371jacobs like the quickest fix for a 2.0 gold release.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs Justin says: "Could we delay this for a point release or 2.1?"
355b4669e025ff377602b6fc7caaf30dbc218371jacobs Not a showstopper: justin, wrowe, trawick, stoddard, Jim, Ian, Aaron,
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * Get perchild to work on platforms other than Linux. This
355b4669e025ff377602b6fc7caaf30dbc218371jacobs will require a portable mechanism to pass data and file/socket
355b4669e025ff377602b6fc7caaf30dbc218371jacobs descriptors between vhost child groups. An API was proposed
355b4669e025ff377602b6fc7caaf30dbc218371jacobs on dev@apr:
355b4669e025ff377602b6fc7caaf30dbc218371jacobs Message-ID: <20020111115006.K1529@clove.org>
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * Try to get libtool inter-library dependency code working on AIX.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs Message-ID: <cm3n10lx555.fsf@rdu163-40-092.nc.rr.com>
355b4669e025ff377602b6fc7caaf30dbc218371jacobs Justin says: If we get it working on AIX, we can enable this
355b4669e025ff377602b6fc7caaf30dbc218371jacobs on all platforms and clean up our build system
355b4669e025ff377602b6fc7caaf30dbc218371jacobs Jeff says: I thought I tested a patch for you sometime in
355b4669e025ff377602b6fc7caaf30dbc218371jacobs January that you were going to commit within a few
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * Handling of %2f in URIs. Currently both 1.3 and 2.0
355b4669e025ff377602b6fc7caaf30dbc218371jacobs completely disallow %2f in the request URI path (see
355b4669e025ff377602b6fc7caaf30dbc218371jacobs ap_unescape_url() in util.c). It's permitted and passed
355b4669e025ff377602b6fc7caaf30dbc218371jacobs through in the query string, however. Roy says the
355b4669e025ff377602b6fc7caaf30dbc218371jacobs original reason for disallowing it, from five years ago,
355b4669e025ff377602b6fc7caaf30dbc218371jacobs was to protect CGI scripts that applied PATH_INFO to
355b4669e025ff377602b6fc7caaf30dbc218371jacobs a filesystem location and which might be tricked by
355b4669e025ff377602b6fc7caaf30dbc218371jacobs ..%2f..%2f(...). We *should* allow path-info of the
355b4669e025ff377602b6fc7caaf30dbc218371jacobs Since we've revamped a lot of our processing of path
355b4669e025ff377602b6fc7caaf30dbc218371jacobs segments, it would be nice to allow this, or at least
355b4669e025ff377602b6fc7caaf30dbc218371jacobs allow it conditionally with a directive.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * FreeBSD, threads, and worker MPM. All seems to work fine
355b4669e025ff377602b6fc7caaf30dbc218371jacobs if you only have one worker process with many threads. Add
355b4669e025ff377602b6fc7caaf30dbc218371jacobs a second worker process and the accept lock seems to be
355b4669e025ff377602b6fc7caaf30dbc218371jacobs lost. This might be an APR issue with how it deals with
355b4669e025ff377602b6fc7caaf30dbc218371jacobs the child_init hook (i.e. the fcntl lock needs to be resynced).
355b4669e025ff377602b6fc7caaf30dbc218371jacobs More examination and analysis is required.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs Status: This has also been reported on Cygwin.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs Message-ID: <3C2CC514.8EF3BED1@wapme-systems.de> (cygnus)
355b4669e025ff377602b6fc7caaf30dbc218371jacobs Justin says: So, FreeBSD-CURRENT and Cywin have the same
355b4669e025ff377602b6fc7caaf30dbc218371jacobs problem. Yum. If another platform has this
355b4669e025ff377602b6fc7caaf30dbc218371jacobs with worker, this becomes a showstopper.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs Aaron says: I spent some time disecting this and have come to
355b4669e025ff377602b6fc7caaf30dbc218371jacobs the conclusion that it is not a problem in the worker MPM
355b4669e025ff377602b6fc7caaf30dbc218371jacobs (or at least, it is not isolated to a problem in worker).
355b4669e025ff377602b6fc7caaf30dbc218371jacobs I'll list some of the problems I'm seeing in case someone
355b4669e025ff377602b6fc7caaf30dbc218371jacobs else wants to pick up where I've left off:
355b4669e025ff377602b6fc7caaf30dbc218371jacobs - Delivery of just about any signal to one of the child
355b4669e025ff377602b6fc7caaf30dbc218371jacobs processes will send it into an infinite loop as well.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs - Even though the parent is spinning out of control,
355b4669e025ff377602b6fc7caaf30dbc218371jacobs at first the child or children will appear to work
355b4669e025ff377602b6fc7caaf30dbc218371jacobs properly. At times it is possible to get it into a state,
355b4669e025ff377602b6fc7caaf30dbc218371jacobs however, where a request will hang until another concurrent
355b4669e025ff377602b6fc7caaf30dbc218371jacobs request "kicks" the first, at which point the second will
355b4669e025ff377602b6fc7caaf30dbc218371jacobs hang. My theory is that this has to do with the
355b4669e025ff377602b6fc7caaf30dbc218371jacobs pthread_cond_*() implementation in FreeBSD, but it's still
355b4669e025ff377602b6fc7caaf30dbc218371jacobs possible that it is in APR.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs Justin adds: Oh, FreeBSD threads are implemented entirely with
355b4669e025ff377602b6fc7caaf30dbc218371jacobs select()/poll()/longjmp(). Welcome to the nightmare.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs So, that means a ktrace output also has the thread
355b4669e025ff377602b6fc7caaf30dbc218371jacobs scheduling internals in it (since it is all the same to
355b4669e025ff377602b6fc7caaf30dbc218371jacobs the kernel). Which makes it hard to distinguish between
355b4669e025ff377602b6fc7caaf30dbc218371jacobs our select() calls and their select() calls.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs *bangs head on wall repeatedly* But, some of the libc_r
355b4669e025ff377602b6fc7caaf30dbc218371jacobs files have a DBG_MSG #define. This is moderately helpful
355b4669e025ff377602b6fc7caaf30dbc218371jacobs when used with -DNO_DETACH. The kernel scheduler isn't
355b4669e025ff377602b6fc7caaf30dbc218371jacobs waking up the threads on a select(). Yum. And, I bet
355b4669e025ff377602b6fc7caaf30dbc218371jacobs those decrementing select calls have to do with the
355b4669e025ff377602b6fc7caaf30dbc218371jacobs scheduler. Time to brush up on our OS fundamentals.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * There is increasing demand from module writers for an API
355b4669e025ff377602b6fc7caaf30dbc218371jacobs that will allow them to control the server � la apachectl.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs Reasons include sole-function servers that need to die if
355b4669e025ff377602b6fc7caaf30dbc218371jacobs an external dependency (e.g., a database) fails, et cetera.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs Perhaps something in the (ever more abused) scoreboard?
355b4669e025ff377602b6fc7caaf30dbc218371jacobs rbb: I don't believe the scoreboard is the correct mechanism
355b4669e025ff377602b6fc7caaf30dbc218371jacobs for this. We already have a pipe that goes between parent
355b4669e025ff377602b6fc7caaf30dbc218371jacobs and child for graceful shutdown events, along with an API that
355b4669e025ff377602b6fc7caaf30dbc218371jacobs can be used to send a message down that pipe. In threaded MPMs,
355b4669e025ff377602b6fc7caaf30dbc218371jacobs it is easy enough to make that one pipe be used for graceful
355b4669e025ff377602b6fc7caaf30dbc218371jacobs and graceless events, and it is also easy to open that pipe
355b4669e025ff377602b6fc7caaf30dbc218371jacobs to both parent and child for writing. Then we just need to
355b4669e025ff377602b6fc7caaf30dbc218371jacobs figure out how to do graceless on non-threaded MPMs.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * Allow the DocumentRoot directive within <Location > scopes? This
355b4669e025ff377602b6fc7caaf30dbc218371jacobs allows the beloved (crusty) Alias /foo/ /somepath/foo/ followed
355b4669e025ff377602b6fc7caaf30dbc218371jacobs by a <Directory /somepath/foo> to become simply
355b4669e025ff377602b6fc7caaf30dbc218371jacobs <Location /foo/> DocumentRoot /somefile/foo (IMHO a bit more legible
355b4669e025ff377602b6fc7caaf30dbc218371jacobs and in-your-face.) DocumentRoot unset would be accepted [and would
355b4669e025ff377602b6fc7caaf30dbc218371jacobs not permit content to be served, only virtual resources such as
355b4669e025ff377602b6fc7caaf30dbc218371jacobs server-info or server-status.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs This proposed change would _not_ depricate Alias.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * Win32: Rotatelogs sometimes is not terminated when Apache
355b4669e025ff377602b6fc7caaf30dbc218371jacobs goes down hard. FirstBill was looking at possibly tracking the
355b4669e025ff377602b6fc7caaf30dbc218371jacobs child's-child processes in the parent process.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs OtherBill asks, wasn't this fixed?
355b4669e025ff377602b6fc7caaf30dbc218371jacobs stoddard: Not fixed. Shared scoreboard might offer a good
355b4669e025ff377602b6fc7caaf30dbc218371jacobs way for the parent to keep track of 'other child' processes
355b4669e025ff377602b6fc7caaf30dbc218371jacobs and whack them if the child goes down.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs Other thoughts on walking the process chain using the NT kernel
355b4669e025ff377602b6fc7caaf30dbc218371jacobs have also been proposed on APR.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * Win32: Add a simple hold console open patch (wait for close or
355b4669e025ff377602b6fc7caaf30dbc218371jacobs the ESC key, with a nice message) if the server died a bad
355b4669e025ff377602b6fc7caaf30dbc218371jacobs death (non-zero exit code) in console mode.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs Resolution: bring forward same ugly hacks from 1.3.13-.20
355b4669e025ff377602b6fc7caaf30dbc218371jacobs This is not so simple. Any exit() from APR or other libraries
355b4669e025ff377602b6fc7caaf30dbc218371jacobs can't be caught unless we add some sort of apr_exit(rv) with
355b4669e025ff377602b6fc7caaf30dbc218371jacobs registered apr_atexit() fn's that have the return code as an
355b4669e025ff377602b6fc7caaf30dbc218371jacobs argument to the registered fn.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * Eliminate unnecessary creation of pipes in mod_cgid
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * Combine log_child and piped_log_spawn. Clean up http_log.c.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs Common logging API.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * Document mod_file_cache.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * Platforms that do not support fork (primarily Win32 and AS/400)
355b4669e025ff377602b6fc7caaf30dbc218371jacobs Architect start-up code that avoids initializing all the modules
355b4669e025ff377602b6fc7caaf30dbc218371jacobs in the parent process on platforms that do not support fork.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * Win32: Migrate the MPM over to use APR thread/process calls. This
355b4669e025ff377602b6fc7caaf30dbc218371jacobs would eliminate some code in the Win32 branch that essentially
355b4669e025ff377602b6fc7caaf30dbc218371jacobs duplicates what is in APR.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * There are still a number of places in the code where we are
355b4669e025ff377602b6fc7caaf30dbc218371jacobs losing error status (i.e. throwing away the error returned by a
355b4669e025ff377602b6fc7caaf30dbc218371jacobs system call and replacing it with a generic error code)
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * Mass vhosting version of suEXEC.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * All DBMs suffer from confusion in support/dbmmanage (perl script) since
355b4669e025ff377602b6fc7caaf30dbc218371jacobs the dbmmanage employs the first-matched dbm format. This is not
355b4669e025ff377602b6fc7caaf30dbc218371jacobs necessarily the library that Apache was built with. Aught to
355b4669e025ff377602b6fc7caaf30dbc218371jacobs rewrite dbmmanage upon installation to bin/ with the proper library
355b4669e025ff377602b6fc7caaf30dbc218371jacobs for predictable mod_auth_dbm administration.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs Questions; htdbm exists, time to kill dbmmanage, or does it remain
355b4669e025ff377602b6fc7caaf30dbc218371jacobs useful as a perl dbm management example? If we keep it,
355b4669e025ff377602b6fc7caaf30dbc218371jacobs do we address the issue above?
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * Integrate mod_dav.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs Some additional items remaining:
355b4669e025ff377602b6fc7caaf30dbc218371jacobs - case_preserved_filename stuff
355b4669e025ff377602b6fc7caaf30dbc218371jacobs (use the new canonical name stuff?)
355b4669e025ff377602b6fc7caaf30dbc218371jacobs - find a new home for ap_text(_header)
355b4669e025ff377602b6fc7caaf30dbc218371jacobs - is it possible to remove the DAV: namespace stuff from util_xml?
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * ap_core_translate() and its use by mod_mmap_static and mod_file_cache
355b4669e025ff377602b6fc7caaf30dbc218371jacobs are a bit wonky. The function should probably be exposed as a utility
355b4669e025ff377602b6fc7caaf30dbc218371jacobs function (such as ap_translate_url2fs() or ap_validate_fs_url() or
355b4669e025ff377602b6fc7caaf30dbc218371jacobs something). Another approach would be a new hook phase after
355b4669e025ff377602b6fc7caaf30dbc218371jacobs "translate" which would allow the module to munge what the
355b4669e025ff377602b6fc7caaf30dbc218371jacobs translation has decided to do.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs Status: Greg +1 (volunteers), Ryan +1
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * Explore use of a post-config hook for the code in http_main.c which
355b4669e025ff377602b6fc7caaf30dbc218371jacobs calls ap_fixup_virutal_hosts(), ap_fini_vhost_config(), and
355b4669e025ff377602b6fc7caaf30dbc218371jacobs ap_sort_hooks() [to reduce the logic in main()]
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * read the config tree just once, and process N times (as necessary)
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * (possibly) use UUIDs in mod_unique_id and/or mod_usertrack
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * (possibly) port the bug fix for PR 6942 (segv when LoadModule is put
355b4669e025ff377602b6fc7caaf30dbc218371jacobs into a VirtualHost container) to 2.0.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * shift stuff to mod_core.h
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * callers of ap_run_create_request() should check the return value
355b4669e025ff377602b6fc7caaf30dbc218371jacobs for failure (Doug volunteers)
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * Win32: Get Apache working on Windows 95/98. The following work
355b4669e025ff377602b6fc7caaf30dbc218371jacobs (at least) needs to be done:
355b4669e025ff377602b6fc7caaf30dbc218371jacobs - Document warning that OSR2 is required (for Crypt functions, in
355b4669e025ff377602b6fc7caaf30dbc218371jacobs rand.c, at least.) This could be resolved with an SSL library, or
355b4669e025ff377602b6fc7caaf30dbc218371jacobs randomization in APR itself.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs - Bring the Win9xConHook.dll from 1.3 into 2.0 (no sense till it
355b4669e025ff377602b6fc7caaf30dbc218371jacobs actually works) and add in a splash of Win9x service code.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * Fix the worker MPM to use POD to kill child processes instead
355b4669e025ff377602b6fc7caaf30dbc218371jacobs of ap_os_killpg, regardless of how they should die. (Ryan Bloom)
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * Scoreboard structures could be changed in the future such that
355b4669e025ff377602b6fc7caaf30dbc218371jacobs proper alignment is not maintained, leading to segfaults on
355b4669e025ff377602b6fc7caaf30dbc218371jacobs some systems. Cliff posted a patch to deal with this issue but
355b4669e025ff377602b6fc7caaf30dbc218371jacobs later recanted. See this message to dev@apr.apache.org:
355b4669e025ff377602b6fc7caaf30dbc218371jacobs Message-ID: <Pine.LNX.4.44.0203011354090.16457-200000@deepthought
355b4669e025ff377602b6fc7caaf30dbc218371jacobsTODO ISSUES REMAINING IN MOD_SSL:
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * In order to use a DSO version of mod_ssl we have to link with
355b4669e025ff377602b6fc7caaf30dbc218371jacobs -lssl and -lcrypto. A workaround is in place right now where the
355b4669e025ff377602b6fc7caaf30dbc218371jacobs entire EXTRA_LIBS macro is being appended to the objects list, but
355b4669e025ff377602b6fc7caaf30dbc218371jacobs this is a hack. We should either revamp the APACHE_CHECK_SSL_TOOLKIT
355b4669e025ff377602b6fc7caaf30dbc218371jacobs autoconf function or come up with some other autoconf checks to
355b4669e025ff377602b6fc7caaf30dbc218371jacobs search for libssl and libcrypto and properly add them to mod_ssl's
355b4669e025ff377602b6fc7caaf30dbc218371jacobs link flags.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * SSL renegotiations in combination with POST request
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * Port or dispose all code inside #if 0...#endif blocks that remain
355b4669e025ff377602b6fc7caaf30dbc218371jacobs from the porting effort.
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * Do we need SSL_set_read_ahead()?
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * the ssl_expr api is NOT THREAD SAFE. race conditions exist:
355b4669e025ff377602b6fc7caaf30dbc218371jacobs -in ssl_expr_comp() if SSLRequire is used in .htaccess
355b4669e025ff377602b6fc7caaf30dbc218371jacobs (ssl_expr_info is global)
355b4669e025ff377602b6fc7caaf30dbc218371jacobs -is ssl_expr_eval() if there is an error
355b4669e025ff377602b6fc7caaf30dbc218371jacobs (ssl_expr_error is global)
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * SSLRequire directive (parsing of) leaks memory
0a44ef6d9afbfe052a7e975f55ea0d2954b62a82jacobs * Diffie-Hellman-Parameters for temporary keys are hardcoded in
0a44ef6d9afbfe052a7e975f55ea0d2954b62a82jacobs ssl_engine_dh.c, while the comment in ssl_engine_kernel.c says:
355b4669e025ff377602b6fc7caaf30dbc218371jacobs "it is suggested that keys be changed daily or every 500
355b4669e025ff377602b6fc7caaf30dbc218371jacobs transactions, and more often if possible."
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * ssl_var_lookup could be rewritten to be MUCH faster
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * CRL callback should be pluggable
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * session cache store should be pluggable
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * init functions should return status code rather than ssl_die()
355b4669e025ff377602b6fc7caaf30dbc218371jacobs * ssl_engine_pphrase.c needs to be reworked so it is generic enough
355b4669e025ff377602b6fc7caaf30dbc218371jacobs to also decrypt proxy keys
* mod_cache: CacheEnable/CacheDisable should accept regular expressions.
* mod_mem_cache/mod_disk_cache: Need to be able to query cache
* mod_mem_cache/mod_disk_cache: Complete implementing config
* Sample config for mod_cache/mod_mem_cache/mod_disk_cache for
inclusion into httpd.conf.
* mod_cache/mod_mem_cache/mod_disk_cache: Documentation.
PR#1191: setlogin() is not called, causing problems with e.g. identd
PR#1287: add allow,deny/deny,allow warning to mod_access
PR#1117: Using NIS passwd.byname dbm files with AuthDBMUserFile
PR#2873: Feedback/Comment on APACI
PR#2431: A small addition to rotatelogs.c to improve program functionality.
PR#2889: Inclusion of RPM spec file in CVS/distributions
* orig_ct in the byterange/multipart handling may not be
Solaris 8.x/sparc yes Jim Jagielski