1 APACHE 2.0 STATUS: -*-text-*-
2 Last modified at [$Date: 2001/10/11 03:42:53 $]
6 2.0.25 : rolled August 29, 2001
7 2.0.24 : rolled August 18, 2001
8 2.0.23 : rolled August 9, 2001
9 2.0.22 : rolled July 29, 2001
10 2.0.21 : rolled July 20, 2001
11 2.0.20 : rolled July 8, 2001
12 2.0.19 : rolled June 27, 2001
13 2.0.18 : rolled May 18, 2001
14 2.0.17 : rolled April 17, 2001
15 2.0.16 : rolled April 4, 2001
16 2.0.15 : rolled March 21, 2001
17 2.0.14 : rolled March 7, 2001
18 2.0a9 : released December 12, 2000
19 2.0a8 : released November 20, 2000
20 2.0a7 : released October 8, 2000
21 2.0a6 : released August 18, 2000
22 2.0a5 : released August 4, 2000
23 2.0a4 : released June 7, 2000
24 2.0a3 : released April 28, 2000
25 2.0a2 : released March 31, 2000
26 2.0a1 : released March 10, 2000
28 Please consult the following STATUS files for information
32 * srclib/apr-util/STATUS
37 * If any request gets to the core handler, without a flag that this
38 r->filename was tested by dir/file_walk, we need to 500 at the very
39 end of the ap_process_request_internal() processing. This provides
40 authors of older modules better compatibility, while still improving
41 the security and robustness of 2.0.
42 Status: still need to decide where this goes, OtherBill comments...
43 Message-ID: <065701c14526$495203b0$96c0b0d0@roweclan.net>
44 we need to look at halting this in the 'default handler' case,
45 and that implies pushing the 'handler election' into the request
46 internal processing phase from the run request phase.
48 * There is a bug in how we sort some hooks, at least the pre-config
49 hook. The first time we call the hooks, they are in the correct
50 order, but the second time, we don't sort them correctly. Currently,
51 the modules/http/config.m4 file has been renamed to
52 modules/http/config2.m4 to work around this problem, it should moved
53 back when this is fixed. rbb
55 * The Add...Filter and Set...Filter directives do not allow the
56 administrator to order filters, beyond the order of filename (mime)
57 extensions. It isn't clear if Set...Filter(s) should be inserted
58 before or after the Add...Filter(s) which are ordered by sequence of
59 filename extensions. Add...FilterByType will add to this quandry.
60 Some sort of resolution needs to be proposed,
62 * mod_dir should normally redirect ALL directory requests which do
63 not include a trailing slash on the URI. However, if a "notes"
64 flag is set (say, via BrowserMatch), this behavior will be
65 disabled for non-GET requests.
66 Status: Greg volunteers
67 MsgId: <20010227104646.E2297@lyra.org>
68 MsgId: <3A9C0097.9C83F07C@Golux.Com>
70 * mod_negotiation needs a new option or directive, something like
71 ForceLanguagePriority, to fall back to the LanguagePriority
72 directive instead of returning a "no acceptable variant" error.
73 Status: Bill has some code in his tree that accomplishes
74 this, and will commit it Friday after it's tested.
76 * Usability: Sanitize the MPM config directives. MaxClients in
77 the threaded MPM is totally misleading now as it has little to
78 do with limiting the number of clients (it limits the number
79 of child processes). Bill proposed nomenclature change to
80 something like "StartWorkers, MaxWorkers, etc." that could
81 apply to most all the MPMs (with some notable exceptions).
82 Bill would be happy with changing MaxClients to MaxServers
83 to make it agree with the operation of the StartServers
86 * revamp the input filter semantics, per discussions since
87 February (and especially at the hackathon last
88 April). Specifically, ap_get_brigade will return a brigade with
89 *up to* a specific number of bytes, or a "line" of data. The
90 read may be blocking or nonblocking. ap_getline() will be
91 refactored into apr_brigade_getline(), and then DECHUNK can use
92 f->next (ap_getline will always read "top of input stack"). Also
93 fix the bug where request body content will end up closing the
94 connection (buggering up persistent conns).
95 Status: Justin is working on this as fast as he can.
96 The core input filters, HTTP-related filters, and mod_ssl
97 are switched to the new logic. At this point, proxy may be
98 broken (Ian says it works, but dechunking is a bit shaky).
100 - socket bucket and core input filter changes. see end of
101 message ID (Feb 27): <20010227075326.S2297@lyra.org>
103 - fix up ap_get_brigade() semantics, fix bug in DECHUNK /
104 ap_getline. many messages (plus their threads) (Apr/May):
105 Message-ID: <20010402101207.J27539@lyra.org>
106 Message-ID: <3AF7F921.D2EEC41A@algroup.co.uk>
107 Message-ID: <20010508190029.E18404@lyra.org>
109 - further work with combining/tweaking the builtin filters:
110 Message-ID: <20010509115445.D1374@lyra.org>
112 - thoughts on filter modes:
113 Message-ID: <021b01c14dee$09782af0$93c0b0d0@roweclan.net>
115 * Fold mod_auth_db features back into mod_auth_dbm, and depricate it.
116 This can't wait until we have a 2.0-gold release, if folks need
117 to move over to auth_dbm, we can't do that to them after 2.0 gold.
119 RELEASE NON-SHOWSTOPPERS BUT WOULD BE REAL NICE TO WRAP THESE UP:
121 * Allow the DocumentRoot directive within <Location > scopes? This
122 allows the beloved (crusty) Alias /foo/ /somepath/foo/ followed
123 by a <Directory /somepath/foo> to become simply
124 <Location /foo/> DocumentRoot /somefile/foo (IMHO a bit more legible
125 and in-your-face.) DocumentRoot unset would be accepted [and would
126 not permit content to be served, only virtual resources such as
127 server-info or server-status.
128 This proposed change would _not_ depricate Alias.
130 * add mod_gz to httpd-2.0 (in modules/experimental/)
131 +1: Greg, Justin, ben, Ken, Jeff, Bill
132 0: Cliff (there's now another candidate to be evaluated)
133 0: Jim (premature decision at present, IMO)
136 non-binding: Ian (+1)
138 * daedalus: mod_cgid and suexec have a problem co-existing. suexec
139 sees a null command string sometimes. The problem happens when
140 you access bugs.apache.org, then click on the "search the bug db"
143 * Win32: Rotatelogs sometimes is not terminated when Apache
144 goes down hard. FirstBill was looking at possibly tracking the
145 child's-child processes in the parent process.
146 OtherBill asks, wasn't this fixed?
148 * Win32: Add a simple hold console open patch (wait for close or
149 the ESC key, with a nice message) if the server died a bad
150 death (non-zero exit code) in console mode.
151 Resolution: bring forward same ugly hacks from 1.3.13-.20
153 * Port of mod_ssl to Apache 2.0:
155 The current porting state is summarized in modules/ssl/README. The
156 remaining work includes:
157 (1) stablizing/optimizing the SSL filter logic
158 (2) Enabling the various SSL caching mechanisms (shmcb, shmht)
159 (3) Enabling SSL extentions
160 (4) Trying to seperate the https filter logic from mod_ssl -
161 This is to facilitate other modules that wish to use the https
162 filter or the mod_ssl logic or both as required.
164 * Performance: Get the SINGLE_LISTEN_UNSERIALIZED_ACCEPT
165 optimization working in threaded. prefork's new design for how
166 to notice data on the pod should be sufficient.
168 * Performance & Debug: Eliminate most (and perhaps all) of the
169 malloc/free calls in the bucket brigade code. Need some
170 light weight memory management functions that allow freeing
171 memory (putting it back into a memory pool) when it is no
172 longer needed. Enabling simple debugging features like guard
173 bands, double free detection, etc. would be cool but certainly
174 not a hard requirement.
176 Status: Cliff started to implement this using SMS as has
177 been discussed at length for months, but since
178 SMS is not being used anywhere else in the server,
179 several people expressed the opinion that we should
180 get rid of it entirely, meaning that the buckets
181 need their own memory management (free list) functions.
182 Cliff will implement that this weekend so we at least
183 have something to look at/compare with.
185 * Eliminate unnecessary creation of pipes in mod_cgid
187 * the autoconf setup should be fixed to default to using the
188 "Apache" layout from config.layout, and each variable settable
189 in a layout should be overridable on the command line. Plus,
190 what we do right now just doesn't seem to fully fit into how autoconf
191 works, eg. AC_PREFIX_DEFAULT issues.
192 Message-ID: <Pine.BSF.4.20.0104031557420.20876-100000@alive.znep.com>
194 * Combine log_child and piped_log_spawn. Clean up http_log.c.
197 * Document mod_file_cache.
199 * OS/2: Make mod_status work for spmt_os2 MPM.
201 * Platforms that do not support fork (primarily Win32 and AS/400)
202 Architect start-up code that avoids initializing all the modules
203 in the parent process on platforms that do not support fork.
205 * Win32: Migrate the MPM over to use APR thread/process calls. This
206 would eliminate some code in the Win32 branch that essentially
207 duplicates what is in APR.
209 * There are still a number of places in the code where we are
210 losing error status (i.e. throwing away the error returned by a
211 system call and replacing it with a generic error code)
213 * Mass vhosting version of suEXEC.
215 * All DBMs suffer from confusion in support/dbmmanage (perl script) since
216 the dbmmanage employs the first-matched dbm format. This is not
217 necessarily the library that Apache was built with. Aught to
218 rewrite dbmmanage upon installation to bin/ with the proper library
219 for predictable mod_auth_db/dbm administration.
220 Status: Mladen Turk has posted several patches and ideas.
221 Key question, part of htpasswd, or a seperate utility?
223 prefer seperate: OtherBill
226 * use apu_dbm in mod_auth_dbm
227 Status: Greg +1 (low-priority volunteer)
228 Justin says: "Seems like this is already there, so should we just
229 remove the other DBM code in that file? If you want
230 to use gdbm, or dbm, etc, you should tell apr-util."
231 Will says: "bs - I may choose the fastest - most efficient native
232 dbm implementation, for shared proc caches, ssl session
233 caching, etc, but that has nothing to do with maintaining
234 a userlist via dbm, which has to remain readable between
235 builds/machines, etc. The use-multiple database schema
236 for apr-util would let us do this with just apr, though."
239 Some additional items remaining:
240 - case_preserved_filename stuff
241 (use the new canonical name stuff?)
242 - find a new home for ap_text(_header)
243 - is it possible to remove the DAV: namespace stuff from util_xml?
245 * ap_core_translate() and its use by mod_mmap_static and mod_file_cache
246 are a bit wonky. The function should probably be exposed as a utility
247 function (such as ap_translate_url2fs() or ap_validate_fs_url() or
248 something). Another approach would be a new hook phase after
249 "translate" which would allow the module to munge what the
250 translation has decided to do.
251 Status: Greg +1 (volunteers), Ryan +1
253 * Explore use of a post-config hook for the code in http_main.c which
254 calls ap_fixup_virutal_hosts(), ap_fini_vhost_config(), and
255 ap_sort_hooks() [to reduce the logic in main()]
257 * read the config tree just once, and process N times (as necessary)
259 * (possibly) use UUIDs in mod_unique_id and/or mod_usertrack
261 * (possibly) port the bug fix for PR 6942 (segv when LoadModule is put
262 into a VirtualHost container) to 2.0.
264 * shift stuff to mod_core.h
266 * APR-ize resolver stuff in mod_unique_id (Jeff volunteers)
268 * callers of ap_run_create_request() should check the return value
269 for failure (Doug volunteers)
271 * when prefork bails out due to an initialization error in the detached
272 process (e.g., mutex init failure), other children (cgid, at least)
273 are left hanging around
275 * Win32: Get Apache working on Windows 95/98. The following work
276 (at least) needs to be done:
277 - winnt MPM: Fix 95/98 code paths in the winnt MPM. There is some NT
278 specific code that is still not in NT only code paths
279 - IOL binds to APR sendfile, implemented with TransmitFile, which
280 is not available on 95/98.
281 - Document warning that OSR2 is required (for Crypt functions, in
282 rand.c, at least.) This could be resolved with an SSL library, or
283 randomization in APR itself.
284 - Bring the Win9xConHook.dll from 1.3 into 2.0 (no sense till it
285 actually works) and add in a splash of Win9x service code.
287 PRs that have been suspended forever waiting for someone to
288 put them into 'the next release':
291 missing call to "setlocale();"
295 Additional status for XBitHack directive
299 Mod_proxy doesn't allow change of error pages
303 Modified PATH environemnt variable is not passed, instead
308 Proxy doesn't deliver documents if not connected
312 proxy converts ~name to %7Ename when name starts with a dot (.)
316 mod_access syntax allows hosts that should be restricted
319 * PR#557: mod_auth-any
320 ~UserHome directories are not honored in absolute pathname
325 Proxy FTP Authentication Fails
328 * PR#623: mod_include
329 A smarter "Last Modified" value for SSI documents (see PR number 600)
333 Request of "Options SymLinksIfGroupMatch"
337 Proxy doesn't do links right for OpenVMS files through ftp:
341 imap should read <MAP><AREA>*</MAP> too!
345 RLimitCPU and RLimitMEM don't apply to all children like they should
349 Uses cwd before filling it in, doesn't use syslog
353 it is useful to allow specifiction that root-owned symlinks
354 should always be followed
358 Controlling Access to Remote Proxies would be nice...
362 Adding authentication "on the fly" through the proxy module
365 * PR#1004: apache-api
366 request_config field in request_rec is moderately bogus
370 DoS attacks involving memory consumption
373 * PR#1050: mod_log-any
374 Logging of virtual server to error_log as well
378 ProxyRemote make a dead cycle.
381 * PR#1117: mod_auth-any
382 Using NIS passwd.byname dbm files with AuthDBMUserFile
386 suexec does not parse arguments to #exec cmd
389 * PR#1145: mod_include
390 Allow for Last-Modified: without resorting to XBitHack
393 * PR#1158: apache-api
394 improvements to child spawning API
398 ``nph-'' not honored (no buffering) for ProxyRemote mapping
402 Apache cannot handle continuation line in headers
406 setlogin() is not called, causing problems with e.g. identd
410 regerror() exists, use it
413 * PR#1233: apache-api
414 there is no way to keep per-connection per-module state
417 * PR#1263: mod_autoexec
418 Add frame-safe anchor attribute to mod_autoindex links
422 CGI scripts running as Apache user: security (suexec etc.)
426 Error messages could be easier to spot in cgi.log file for suexec.c
429 * PR#1287: mod_access
430 add allow,deny/deny,allow warning to mod_access
434 Need to know "hit-rate" on proxy cache
437 * PR#1358: mod_log-any
438 Selective url-encode of log fields (or maybe a pseudo
442 * PR#1383: mod_headers
443 I make mod_headers to modify request headers as well as
448 Proxy transfer logging
452 No HTTP_X_FORWARDED_FOR set...
456 ProxyRemote proxy requests fail authentication by firewall
459 * PR#1582: mod_rewrite
460 mod_rewrite forms REQUEST_URI different than mod_cgi does
463 * PR#1677: mod_headers
464 mod_headers should allow mod_log_config-style formats in
469 mod_proxy to support persistent conns?
472 * PR#1803: mod_include
473 patches to mod_include to allow for file tests
476 * PR#1809: mod_auth-any
477 Suggestion for improving authentication modules and core source
478 code, problem with 401 and ErrorDocument
482 listing of proxy cache content
486 Allow modules to set user:group for execution.
489 * PR#2024: apache-api
490 adding auth_why to conn_rec
493 * PR#2073: mod_log-any
494 pipelined connections are not logged correctly
497 * PR#2074: mod_rewrite
498 mod_rewrite doesn't pass Proxy Throughput on internal subrequests
502 HTTP Server Rebuild Line Needs Changing for the better
505 * PR#2138: mod_status
506 mod_status always displays 256 possible connection slots
509 * PR#2221: documentation
510 Make online documentation search link back to my installation
514 Can not POST to ErrorDocument - Apache/1.3b6
518 patterns in ProxyRemote
521 * PR#2343: mod_status
522 Status module averages are for entire uptime
526 suexec for general access of user content?
530 Proposal for TimeZone directive
534 /server-info doesn't check for the virtual host to list the info
538 problem specifying ndbm library for build ?with autoconfigure
542 A small addition to rotatelogs.c to improve program functionality.
546 AllowOverride FileInfo is too coarse
550 TimeOut applies to output of CGI scripts
553 * PR#2512: mod_access
554 <IfDenied> directive wanted
558 CGI's for general use still have to be run as another user
563 Cache file names in Proxy module
567 [PATCH] User/Group for <Directory> and <Location> i.e. not only
568 in global and <Virtual>.
572 mailto tags and bundling bug report script
576 Support for System Resource Controller
580 When will Apache support P3P? Any Plans?
584 Feedback/Comment on APACI
588 Inclusion of RPM spec file in CVS/distributions
592 Propose that Apache recommend $UNIQUE_ID for all "session id"
597 suggestion: power up your Include directive :)
601 cannot limit some HTTP methods
604 * PR#3143: apache-api
605 No module specific data hook for per-connection data
608 * PR#3191: mod_negotiation
609 no way to set global quality-of-source (qs) coneg values
614 Accessing URL through proxy server corrupts data.
618 Some anonymous FTP URLs ask for authentication
622 New ErrorDocumentMatch directive
626 Need to be able to override shebang line to make CGI scripts
631 "Files" and "FilesMatch" regexp does not recognize bang as
635 * PR#4448: mod_log-any
636 Please allow CGI env variables (QUERY_STRING, ...) to be logged
640 * PR#4459: mod_include
641 Suggestion for better handling of Last-modified headers
645 mod_cgi prevents handling of OPTIONS requests
648 * PR#5713: os-windows
649 [PATCH] install as win32 service with domain account
653 AllowOverride should have a 'CheckNone' and 'AllowNone' argument
654 instead of only 'None'
657 Other bugs that need fixing:
659 * MaxRequestsPerChild measures connections, not requests.
660 Until someone has a better way, we'll probably just rename it
661 "MaxConnectionsPerChild".
663 * Regex containers don't work in an intutive way
664 Status: No one has come up with an efficient way to fix this
665 behavior. Dean has suggested getting rid of regex containers
668 * SIGSEGV on Linux (glibc 2.1.2) isn't caught properly by a
669 sigwaiting thread. We need to work around this, perhaps unless
670 there is hope soon for a fixed glibc.
672 * orig_ct in the byterange/multipart handling may not be
673 needed. Apache 1.3 just never stashed "multipart" into
674 r->content_type. We should probably follow suit since the
675 byterange stuff doesn't want the rest of the code to see the
676 multipart content-type; the other code should still think it is
677 dealing with the <orig_ct> stuff.
678 Status: Greg volunteers to investigate (esp. since he was most
679 likely the one to break it :-)
681 Other features that need writing:
683 * Finish infrastructure in core for async MPMs
686 * TODO in source -- just do an egrep on "TODO" and see what's there
690 * Jon Travis's <jtravis@covalent.net> patch to deal with thread-safe
691 issues with inet_ntoa. See message <20001201163220.A12827@covalent.net>
692 Status: This is being set aside until the IPv6 work is finished
693 so that we know exactly what is required.
695 * Martin Sojka <msojka@gmx.de>'s patch to add error reporting for failed
696 htpasswd actions due to a full /tmp volume (other programs may have
701 * Mike Abbott's <mja@trudge.engr.sgi.com> patches to improve
703 Status: These were written for 1.3, and are awaiting a port to
706 * Jim Winstead's <jimw@trainedmonkey.com> patch to add CookieDomain and
707 other small mod_usertrack features
709 * Dan Rench's <drench@xnet.com> patch to add allow the errmsg and timefmt
710 of SSI's to be modified in the config file. Patch is available in
715 * Which MPMs will be included with Apache 2.0?