1 APACHE 2.0 STATUS: -*-text-*-
2 Last modified at [$Date: 2002/01/02 08:05:15 $]
6 2.0.30 : In development
7 2.0.29 : tagged November 27, 2001
8 2.0.28 : released November 13, 2001
9 2.0.27 : rolled November 6, 2001
10 2.0.26 : tagged October 16, 2001. not rolled.
11 2.0.25 : rolled August 29, 2001
12 2.0.24 : rolled August 18, 2001
13 2.0.23 : rolled August 9, 2001
14 2.0.22 : rolled July 29, 2001
15 2.0.21 : rolled July 20, 2001
16 2.0.20 : rolled July 8, 2001
17 2.0.19 : rolled June 27, 2001
18 2.0.18 : rolled May 18, 2001
19 2.0.17 : rolled April 17, 2001
20 2.0.16 : rolled April 4, 2001
21 2.0.15 : rolled March 21, 2001
22 2.0.14 : rolled March 7, 2001
23 2.0a9 : released December 12, 2000
24 2.0a8 : released November 20, 2000
25 2.0a7 : released October 8, 2000
26 2.0a6 : released August 18, 2000
27 2.0a5 : released August 4, 2000
28 2.0a4 : released June 7, 2000
29 2.0a3 : released April 28, 2000
30 2.0a2 : released March 31, 2000
31 2.0a1 : released March 10, 2000
33 Please consult the following STATUS files for information
37 * srclib/apr-util/STATUS
42 * ap_directory_walk skips some per-dir config merge functions
43 if there is no "<Directory />" block in the configuration
44 Message-ID: <m3itbdiijq.fsf@rdu163-40-092.nc.rr.com>
46 * Test suite failures:
47 o perchild doesn't even build
48 o both worker and prefork are failing some of the 'chunked' subtests
49 o worker is also failing some of the 'cgi' subtests
50 (see <URL:http://Source-Zone.Org/Apache/regression/>):
52 * If any request gets to the core handler, without a flag that this
53 r->filename was tested by dir/file_walk, we need to 500 at the very
54 end of the ap_process_request_internal() processing. This provides
55 authors of older modules better compatibility, while still improving
56 the security and robustness of 2.0.
57 Status: still need to decide where this goes, OtherBill comments...
58 Message-ID: <065701c14526$495203b0$96c0b0d0@roweclan.net>
59 we need to look at halting this in the 'default handler' case,
60 and that implies pushing the 'handler election' into the request
61 internal processing phase from the run request phase.
63 * There is a bug in how we sort some hooks, at least the pre-config
64 hook. The first time we call the hooks, they are in the correct
65 order, but the second time, we don't sort them correctly. Currently,
66 the modules/http/config.m4 file has been renamed to
67 modules/http/config2.m4 to work around this problem, it should moved
68 back when this is fixed. rbb
70 * The Add...Filter and Set...Filter directives do not allow the
71 administrator to order filters, beyond the order of filename (mime)
72 extensions. It isn't clear if Set...Filter(s) should be inserted
73 before or after the Add...Filter(s) which are ordered by sequence of
74 filename extensions. At minimum, some sort of +-[0-10] syntax seems
75 like the quickest fix for a 2.0 gold release.
77 * mod_negotiation needs a new option or directive, something like
78 ForceLanguagePriority, to fall back to the LanguagePriority
79 directive instead of returning a "no acceptable variant" error.
80 Status: Bill has some code in his tree that accomplishes
81 this, and will commit it Friday after it's tested.
83 * Fold mod_auth_db features back into mod_auth_dbm, and depricate it.
84 This can't wait until we have a 2.0-gold release, if folks need
85 to move over to auth_dbm, we can't do that to them after 2.0 gold.
86 Status: Ian says.. auth_dbm can now handle multiple DBM types,
87 is this still an issue?
88 Vote: Remove mod_auth_db
91 * Convert all instances of the old apr_lock_t type to the new
92 types (once they are fully supported in APR).
93 Status: Aaron is working on converting INTRAPROCESS
94 to apr_thread_mutex_t types. Full replacements for
95 LOCKALL and CROSS_PROCESS are not yet complete on all
96 platforms, and should only be used in MPMs like worker
97 with limited OS exposure.
99 * ap_create_scoreboard() can exit the process, leaving stuff like
100 mod_cgid's daemon process stranded. Either ap_create_scoreboard()
101 needs to be called at a different time or the pre-mpm hook needs
102 to be able to return an error code.
104 RELEASE NON-SHOWSTOPPERS BUT WOULD BE REAL NICE TO WRAP THESE UP:
106 * Handling of %2f in URIs. Currently both 1.3 and 2.0
107 completely disallow %2f in the request URI path (see
108 ap_unescape_url() in util.c). It's permitted and passed
109 through in the query string, however. Roy says the
110 original reason for disallowing it, from five years ago,
111 was to protect CGI scripts that applied PATH_INFO to
112 a filesystem location and which might be tricked by
113 ..%2f..%2f(...). We *should* allow path-info of the
114 form 'http://foo.com/index.cgi/path/to/path%2finfo'.
115 Since we've revamped a lot of our processing of path
116 segments, it would be nice to allow this, or at least
117 allow it conditionally with a directive.
119 * FreeBSD, threads, and worker MPM. All seems to work fine
120 if you only have one worker process with many threads. Add
121 a second worker process and the accept lock seems to be
122 lost. This might be an APR issue with how it deals with
123 the child_init hook (i.e. the fcntl lock needs to be resynced).
124 More examination and analysis is required.
126 * There is increasing demand from module writers for an API
127 that will allow them to control the server à la apachectl.
128 Reasons include sole-function servers that need to die if
129 an external dependency (e.g., a database) fails, et cetera.
130 Perhaps something in the (ever more abused) scoreboard?
131 rbb: I don't believe the scoreboard is the correct mechanism
132 for this. We already have a pipe that goes between parent
133 and child for graceful shutdown events, along with an API that
134 can be used to send a message down that pipe. In threaded MPMs,
135 it is easy enough to make that one pipe be used for graceful
136 and graceless events, and it is also easy to open that pipe
137 to both parent and child for writing. Then we just need to figure
138 out how to do graceless on non-threaded MPMs.
140 * revamp the input filter behavior, per discussions since
141 February (and especially at the hackathon last
142 April). Specifically, ap_get_brigade will return a brigade with
143 *up to* a specific number of bytes, or a "line" of data. The
144 read may be blocking or nonblocking. ap_getline() will be
145 refactored into apr_brigade_getline(), and then DECHUNK can use
146 f->next (ap_getline will always read "top of input stack"). Also
147 fix the bug where request body content will end up closing the
148 connection (buggering up persistent conns).
149 Status: Justin is working on this as fast as he can.
150 The core input filters, HTTP-related filters, mod_ssl, and
151 mod_proxy are switched to the new logic.
152 However, ap_getline() still needs to be refactored out. But,
153 there's a problem there: ap_getline() peeks ahead for MIME
154 continuation (first character on line is space or \t) and
155 stores unused data in core_request_config which violates the
156 abstraction. That's cheating. So, we may not be able to
157 implement this without setting some data aside (yuck!).
158 I believe this is OtherBill's main complaint with the current
160 AIUI (correct me if I'm wrong!), OtherBill believes we
161 should have a pushback option so that we can return unread
162 data - this would solve this case. However, my question to
163 him is how do we handle stuff like mod_ssl - we can't "unread"
164 data. So, do we have two brigades for each filter? An in
165 brigade and a returned brigade? That seems messy. To
166 everyone else, can we refactor ap_getline() without pushback
169 - socket bucket and core input filter changes. see end of
170 message ID (Feb 27): <20010227075326.S2297@lyra.org>
172 - fix up ap_get_brigade() semantics, fix bug in DECHUNK /
173 ap_getline. many messages (plus their threads) (Apr/May):
174 Message-ID: <20010402101207.J27539@lyra.org>
175 Message-ID: <3AF7F921.D2EEC41A@algroup.co.uk>
176 Message-ID: <20010508190029.E18404@lyra.org>
178 - further work with combining/tweaking the builtin filters:
179 Message-ID: <20010509115445.D1374@lyra.org>
181 - thoughts on filter modes:
182 Message-ID: <021b01c14dee$09782af0$93c0b0d0@roweclan.net>
184 * Allow the DocumentRoot directive within <Location > scopes? This
185 allows the beloved (crusty) Alias /foo/ /somepath/foo/ followed
186 by a <Directory /somepath/foo> to become simply
187 <Location /foo/> DocumentRoot /somefile/foo (IMHO a bit more legible
188 and in-your-face.) DocumentRoot unset would be accepted [and would
189 not permit content to be served, only virtual resources such as
190 server-info or server-status.
191 This proposed change would _not_ depricate Alias.
193 * Win32: Rotatelogs sometimes is not terminated when Apache
194 goes down hard. FirstBill was looking at possibly tracking the
195 child's-child processes in the parent process.
196 OtherBill asks, wasn't this fixed?
198 * Win32: Add a simple hold console open patch (wait for close or
199 the ESC key, with a nice message) if the server died a bad
200 death (non-zero exit code) in console mode.
201 Resolution: bring forward same ugly hacks from 1.3.13-.20
203 * Port of mod_ssl to Apache 2.0:
205 The current porting state is summarized in modules/ssl/README. The
206 remaining work includes:
207 (1) stablizing/optimizing the SSL filter logic
208 (2) Enabling the various SSL caching mechanisms (shmcb, shmht)
209 (3) Enabling SSL extentions
210 (4) Trying to seperate the https filter logic from mod_ssl -
211 This is to facilitate other modules that wish to use the https
212 filter or the mod_ssl logic or both as required.
213 Justin: mod_ssl filter logic is redone, so that should be fine.
214 Madhu has submitted a patch for SSL caching - however, I
215 am -0 on that patch as I *think* we could implement the
216 shared memory another way that is much cleaner (i.e.
217 treat shmem directly as a dbm via APR routines). Justin
218 also thinks that the https filter logic may be sufficiently
219 decoupled now, but isn't really sure.
221 * Performance: Get the SINGLE_LISTEN_UNSERIALIZED_ACCEPT
222 optimization working in worker. prefork's new design for how
223 to notice data on the pod should be sufficient.
225 * Performance & Debug: Eliminate most (and perhaps all) of the
226 malloc/free calls in the bucket brigade code. Need some
227 light weight memory management functions that allow freeing
228 memory (putting it back into a memory pool) when it is no
229 longer needed. Enabling simple debugging features like guard
230 bands, double free detection, etc. would be cool but certainly
231 not a hard requirement.
233 Status: Cliff started to implement this using SMS as has
234 been discussed at length for months, but since
235 SMS is not being used anywhere else in the server,
236 several people expressed the opinion that we should
237 get rid of it entirely, meaning that the buckets
238 need their own memory management (free list) functions.
239 Cliff will implement that this weekend so we at least
240 have something to look at/compare with.
242 * Eliminate unnecessary creation of pipes in mod_cgid
244 * the autoconf setup should be fixed to default to using the
245 "Apache" layout from config.layout, and each variable settable
246 in a layout should be overridable on the command line. Plus,
247 what we do right now just doesn't seem to fully fit into how autoconf
248 works, eg. AC_PREFIX_DEFAULT issues.
249 Message-ID: <Pine.BSF.4.20.0104031557420.20876-100000@alive.znep.com>
251 * Combine log_child and piped_log_spawn. Clean up http_log.c.
254 * Document mod_file_cache.
256 * OS/2: Make mod_status work for spmt_os2 MPM.
258 * Platforms that do not support fork (primarily Win32 and AS/400)
259 Architect start-up code that avoids initializing all the modules
260 in the parent process on platforms that do not support fork.
262 * Win32: Migrate the MPM over to use APR thread/process calls. This
263 would eliminate some code in the Win32 branch that essentially
264 duplicates what is in APR.
266 * There are still a number of places in the code where we are
267 losing error status (i.e. throwing away the error returned by a
268 system call and replacing it with a generic error code)
270 * Mass vhosting version of suEXEC.
272 * All DBMs suffer from confusion in support/dbmmanage (perl script) since
273 the dbmmanage employs the first-matched dbm format. This is not
274 necessarily the library that Apache was built with. Aught to
275 rewrite dbmmanage upon installation to bin/ with the proper library
276 for predictable mod_auth_db/dbm administration.
277 Status: Mladen Turk has posted several patches and ideas.
278 Key question, part of htpasswd, or a seperate utility?
280 prefer seperate: OtherBill
282 * use apu_dbm in mod_auth_dbm
283 Status: Greg +1 (low-priority volunteer)
284 Justin says: "Seems like this is already there, so should we just
285 remove the other DBM code in that file? If you want
286 to use gdbm, or dbm, etc, you should tell apr-util."
287 Will says: "bs - I may choose the fastest - most efficient native
288 dbm implementation, for shared proc caches, ssl session
289 caching, etc, but that has nothing to do with maintaining
290 a userlist via dbm, which has to remain readable between
291 builds/machines, etc. The use-multiple database schema
292 for apr-util would let us do this with just apr, though."
293 Ian says: "multi-dbm is in, but it still has ndbm support hardcoded
294 is this still required? isn't ndbm supported via gdbm?"
298 Some additional items remaining:
299 - case_preserved_filename stuff
300 (use the new canonical name stuff?)
301 - find a new home for ap_text(_header)
302 - is it possible to remove the DAV: namespace stuff from util_xml?
304 * ap_core_translate() and its use by mod_mmap_static and mod_file_cache
305 are a bit wonky. The function should probably be exposed as a utility
306 function (such as ap_translate_url2fs() or ap_validate_fs_url() or
307 something). Another approach would be a new hook phase after
308 "translate" which would allow the module to munge what the
309 translation has decided to do.
310 Status: Greg +1 (volunteers), Ryan +1
312 * Explore use of a post-config hook for the code in http_main.c which
313 calls ap_fixup_virutal_hosts(), ap_fini_vhost_config(), and
314 ap_sort_hooks() [to reduce the logic in main()]
316 * read the config tree just once, and process N times (as necessary)
318 * (possibly) use UUIDs in mod_unique_id and/or mod_usertrack
320 * (possibly) port the bug fix for PR 6942 (segv when LoadModule is put
321 into a VirtualHost container) to 2.0.
323 * shift stuff to mod_core.h
325 * callers of ap_run_create_request() should check the return value
326 for failure (Doug volunteers)
328 * Win32: Get Apache working on Windows 95/98. The following work
329 (at least) needs to be done:
330 - winnt MPM: Fix 95/98 code paths in the winnt MPM. There is some NT
331 specific code that is still not in NT only code paths
332 - IOL binds to APR sendfile, implemented with TransmitFile, which
333 is not available on 95/98.
334 - Document warning that OSR2 is required (for Crypt functions, in
335 rand.c, at least.) This could be resolved with an SSL library, or
336 randomization in APR itself.
337 - Bring the Win9xConHook.dll from 1.3 into 2.0 (no sense till it
338 actually works) and add in a splash of Win9x service code.
340 * In order to use a DSO version of mod_ssl we have to link with
341 -lssl and -lcrypto. A workaround is in place right now where the
342 entire EXTRA_LIBS macro is being appended to the objects list, but
343 this is a hack. We should either revamp the APACHE_CHECK_SSL_TOOLKIT
344 autoconf function or come up with some other autoconf checks to
345 search for libssl and libcrypto and properly add them to mod_ssl's
348 * Make the worker MPM the default MPM for threaded Unix boxes.
349 +1: Justin, Jeff, Ian
350 -0: Aaron (premature decision, needs more discussion)
351 -0: Cliff (I think the default config should be the safest possible)
353 * Fix the worker MPM to use POD to kill child processes instead
354 of ap_os_killpg, regardless of how they should die. (Ryan Bloom)
356 PRs that have been suspended forever waiting for someone to
357 put them into 'the next release':
360 missing call to "setlocale();"
364 Additional status for XBitHack directive
368 Mod_proxy doesn't allow change of error pages
372 Modified PATH environemnt variable is not passed, instead
377 Proxy doesn't deliver documents if not connected
381 proxy converts ~name to %7Ename when name starts with a dot (.)
385 mod_access syntax allows hosts that should be restricted
388 * PR#557: mod_auth-any
389 ~UserHome directories are not honored in absolute pathname
394 Proxy FTP Authentication Fails
397 * PR#623: mod_include
398 A smarter "Last Modified" value for SSI documents (see PR number 600)
402 Request of "Options SymLinksIfGroupMatch"
406 Proxy doesn't do links right for OpenVMS files through ftp:
410 imap should read <MAP><AREA>*</MAP> too!
414 RLimitCPU and RLimitMEM don't apply to all children like they should
418 Uses cwd before filling it in, doesn't use syslog
422 it is useful to allow specifiction that root-owned symlinks
423 should always be followed
427 Controlling Access to Remote Proxies would be nice...
431 Adding authentication "on the fly" through the proxy module
434 * PR#1004: apache-api
435 request_config field in request_rec is moderately bogus
439 DoS attacks involving memory consumption
442 * PR#1050: mod_log-any
443 Logging of virtual server to error_log as well
447 ProxyRemote make a dead cycle.
450 * PR#1117: mod_auth-any
451 Using NIS passwd.byname dbm files with AuthDBMUserFile
455 suexec does not parse arguments to #exec cmd
458 * PR#1145: mod_include
459 Allow for Last-Modified: without resorting to XBitHack
462 * PR#1158: apache-api
463 improvements to child spawning API
467 ``nph-'' not honored (no buffering) for ProxyRemote mapping
471 Apache cannot handle continuation line in headers
475 setlogin() is not called, causing problems with e.g. identd
479 regerror() exists, use it
482 * PR#1233: apache-api
483 there is no way to keep per-connection per-module state
486 * PR#1263: mod_autoexec
487 Add frame-safe anchor attribute to mod_autoindex links
491 CGI scripts running as Apache user: security (suexec etc.)
495 Error messages could be easier to spot in cgi.log file for suexec.c
498 * PR#1287: mod_access
499 add allow,deny/deny,allow warning to mod_access
503 Need to know "hit-rate" on proxy cache
506 * PR#1358: mod_log-any
507 Selective url-encode of log fields (or maybe a pseudo
511 * PR#1383: mod_headers
512 I make mod_headers to modify request headers as well as
517 Proxy transfer logging
521 No HTTP_X_FORWARDED_FOR set...
525 ProxyRemote proxy requests fail authentication by firewall
528 * PR#1582: mod_rewrite
529 mod_rewrite forms REQUEST_URI different than mod_cgi does
532 * PR#1677: mod_headers
533 mod_headers should allow mod_log_config-style formats in
538 mod_proxy to support persistent conns?
541 * PR#1803: mod_include
542 patches to mod_include to allow for file tests
545 * PR#1809: mod_auth-any
546 Suggestion for improving authentication modules and core source
547 code, problem with 401 and ErrorDocument
551 listing of proxy cache content
555 Allow modules to set user:group for execution.
558 * PR#2024: apache-api
559 adding auth_why to conn_rec
562 * PR#2073: mod_log-any
563 pipelined connections are not logged correctly
566 * PR#2074: mod_rewrite
567 mod_rewrite doesn't pass Proxy Throughput on internal subrequests
571 HTTP Server Rebuild Line Needs Changing for the better
574 * PR#2138: mod_status
575 mod_status always displays 256 possible connection slots
578 * PR#2221: documentation
579 Make online documentation search link back to my installation
583 Can not POST to ErrorDocument - Apache/1.3b6
587 patterns in ProxyRemote
590 * PR#2343: mod_status
591 Status module averages are for entire uptime
595 suexec for general access of user content?
599 Proposal for TimeZone directive
603 /server-info doesn't check for the virtual host to list the info
607 problem specifying ndbm library for build ?with autoconfigure
611 A small addition to rotatelogs.c to improve program functionality.
615 AllowOverride FileInfo is too coarse
619 TimeOut applies to output of CGI scripts
622 * PR#2512: mod_access
623 <IfDenied> directive wanted
627 CGI's for general use still have to be run as another user
632 Cache file names in Proxy module
636 [PATCH] User/Group for <Directory> and <Location> i.e. not only
637 in global and <Virtual>.
641 mailto tags and bundling bug report script
645 Support for System Resource Controller
649 When will Apache support P3P? Any Plans?
653 Feedback/Comment on APACI
657 Inclusion of RPM spec file in CVS/distributions
661 Propose that Apache recommend $UNIQUE_ID for all "session id"
666 suggestion: power up your Include directive :)
670 cannot limit some HTTP methods
673 * PR#3143: apache-api
674 No module specific data hook for per-connection data
677 * PR#3191: mod_negotiation
678 no way to set global quality-of-source (qs) coneg values
683 Accessing URL through proxy server corrupts data.
687 Some anonymous FTP URLs ask for authentication
691 New ErrorDocumentMatch directive
695 Need to be able to override shebang line to make CGI scripts
700 "Files" and "FilesMatch" regexp does not recognize bang as
704 * PR#4448: mod_log-any
705 Please allow CGI env variables (QUERY_STRING, ...) to be logged
709 * PR#4459: mod_include
710 Suggestion for better handling of Last-modified headers
714 mod_cgi prevents handling of OPTIONS requests
717 * PR#5713: os-windows
718 [PATCH] install as win32 service with domain account
722 AllowOverride should have a 'CheckNone' and 'AllowNone' argument
723 instead of only 'None'
726 Other bugs that need fixing:
728 * MaxRequestsPerChild measures connections, not requests.
729 Until someone has a better way, we'll probably just rename it
730 "MaxConnectionsPerChild".
732 * Regex containers don't work in an intutive way
733 Status: No one has come up with an efficient way to fix this
734 behavior. Dean has suggested getting rid of regex containers
737 * SIGSEGV on Linux (glibc 2.1.2) isn't caught properly by a
738 sigwaiting thread. We need to work around this, perhaps unless
739 there is hope soon for a fixed glibc.
741 * orig_ct in the byterange/multipart handling may not be
742 needed. Apache 1.3 just never stashed "multipart" into
743 r->content_type. We should probably follow suit since the
744 byterange stuff doesn't want the rest of the code to see the
745 multipart content-type; the other code should still think it is
746 dealing with the <orig_ct> stuff.
747 Status: Greg volunteers to investigate (esp. since he was most
748 likely the one to break it :-)
750 Other features that need writing:
752 * Finish infrastructure in core for async MPMs
755 * TODO in source -- just do an egrep on "TODO" and see what's there
759 * Jon Travis's <jtravis@covalent.net> patch to deal with thread-safe
760 issues with inet_ntoa. See message <20001201163220.A12827@covalent.net>
761 Status: This is being set aside until the IPv6 work is finished
762 so that we know exactly what is required.
764 * Martin Sojka <msojka@gmx.de>'s patch to add error reporting for failed
765 htpasswd actions due to a full /tmp volume (other programs may have
770 * Mike Abbott's <mja@trudge.engr.sgi.com> patches to improve
772 Status: These were written for 1.3, and are awaiting a port to
775 * Jim Winstead's <jimw@trainedmonkey.com> patch to add CookieDomain and
776 other small mod_usertrack features
778 * Dan Rench's <drench@xnet.com> patch to add allow the errmsg and timefmt
779 of SSI's to be modified in the config file. Patch is available in
784 * Which MPMs will be included with Apache 2.0?