1 APACHE 2.3 STATUS: -*-text-*-
2 Last modified at [$Date$]
4 The current version of this file can be found at:
6 * http://svn.apache.org/repos/asf/httpd/httpd/trunk/STATUS
8 Documentation status is maintained seperately and can be found at:
10 * docs/STATUS in this source tree, or
11 * http://svn.apache.org/repos/asf/httpd/httpd/trunk/docs/STATUS
13 Consult the following STATUS files for information on related projects:
15 * http://svn.apache.org/repos/asf/apr/apr/trunk/STATUS
16 * http://svn.apache.org/repos/asf/apr/apr-util/trunk/STATUS
18 Patches considered for backport are noted in their branches' STATUS:
20 * http://svn.apache.org/repos/asf/httpd/httpd/branches/2.0.x/STATUS
21 * http://svn.apache.org/repos/asf/httpd/httpd/branches/2.2.x/STATUS
26 [NOTE that x.{odd}.z versions are strictly Alpha/Beta releases,
27 while x.{even}.z versions are Stable/GA releases.]
29 2.3.12 : In development.
30 2.3.11 : Released as Beta on March 7, 2011.
31 2.3.10 : Tagged on December 13, 2010. Released Dec 21, 2010.
32 2.3.9 : Tagged on November 23, 2010, not released.
33 2.3.8 : Tagged on August 24, 2010.
34 2.3.7 : Tagged on August 19, 2010, not released.
35 2.3.6 : Released on June 21, 2010.
36 2.3.5 : Released on January 26, 2010.
37 2.3.4 : Released on December 8, 2009.
38 2.3.3 : Tagged on November 11, 2009, not released.
39 2.3.2 : Tagged on March 23, 2009, not released.
40 2.3.1 : Tagged on January 2, 2009, not released.
41 2.3.0 : Tagged on December 6, 2008, not released.
43 Contributors looking for a mission:
45 * Just do an egrep on "TODO" or "XXX" in the source.
47 * Review the bug database at: http://issues.apache.org/bugzilla/
49 * Review the "PatchAvailable" bugs in the bug database:
51 https://issues.apache.org/bugzilla/buglist.cgi?bug_status=NEW&bug_status=ASSIGNED&bug_status=REOPENED&product=Apache+httpd-2&keywords=PatchAvailable
53 After testing, you can append a comment saying "Reviewed and tested".
55 * Open bugs in the bug database.
57 * See also the STATUS file in the docs/ directory, which lists documentation-specific TODO items.
60 CURRENT RELEASE NOTES:
65 Jim proposes another beta release the end of March, another in
66 April with a goal towards GA in May (at the latest). Jim volunteers
73 * Modules that are not ready for production use should be marked as
74 experimental or be removed. The same for modules without documentation.
79 * Review the example configuration. It should be based on current best
80 practices and not use deprecated features.
82 * Not all MPMs are updated to set conn_rec::current_thread correctly.
83 (Prefork, Worker, Event, Simple are updated).
84 jim sez: Then we just ship with those... mark any others as
85 experimental, pgollucci +1 jim
87 * The mod_session* modules need to be checked that their hooks respect
88 the returning of int (HTTP status codes) and apr_status_t as appropriate,
89 and any anomolies fixed.
90 jim sez: from what I can see, mod_session* is no worse that other
91 modules that mix these 2 types... clean up is
92 forthcoming but should not be considered a blocker, imo
95 * mod_ssl's proxy support only allows one proxy client certificate per
96 frontend virtual host. Lift this restriction.
97 jim sez: Why a blocker?, pgollucci +1 jim
102 OLD ISSUES THAT WERE THOUGHT TO BE SHOWSTOPPERS FOR 2.2 BUT OBVIOUSLY WEREN'T:
104 * Handling of non-trailing / config by non-default handler is broken
105 http://marc.theaimsgroup.com/?l=apache-httpd-dev&m=105451701628081&w=2
106 jerenkrantz asks: Why should this block a release?
107 wsanchez agrees: this may be a change in behavior, but isn't
108 clearly wrong, and even if so, it doesn't seem like a
111 * the edge connection filter cannot be removed
112 http://marc.theaimsgroup.com/?l=apache-httpd-dev&m=105366252619530&w=2
114 jerenkrantz asks: Why should this block a release?
116 stas replies: because it requires a rewrite of the filters stack
117 implementation (you have suggested that) and once 2.2 is
118 released you can't do that anymore.
120 pgollucci: this affects mod_perl I'm pretty sure.
124 * Name the Server (version 2.4 or 3.0, depending on the final call)
125 Recent discussion indicates we should designate a (short name).
126 This is not yet a [Vote] - Your nominations please:
127 * Apache HTTP Server (httpd)
128 +1: sctemme (why mess with it?), pgollucci
130 RELEASE NON-SHOWSTOPPERS BUT WOULD BE REAL NICE TO WRAP THESE UP:
132 * Clean up all the kruft and *extremely* outdated stuff below...
134 * Maybe remove Limit/LimitExcept or at least make it log warnings when
137 * Sort out modules selections for most/all/reallyall. Maybe rename
138 all -> most, reallyall -> all, and remove the old 'most'.
140 * Patches submitted to the bug database:
141 http://issues.apache.org/bugzilla/buglist.cgi?bug_status=NEW&bug_status=ASSIGNED&bug_status=REOPENED&product=Apache+httpd-2&keywords=PatchAvailable
143 * Filter stacks and subrequests, redirects and fast redirects.
144 There's at least one PR that suffers from the current unclean behaviour
145 (which lets the server send garbage): PR 17629
146 nd says: Every subrequest should get its own filter stack with the
147 subreq_core filter as bottom-most. That filter does two things:
148 - swallow EOS buckets
149 - redirect the data stream to the upper request's (rr->main)
150 filter chain directly after the subrequest's starting
152 Once we have a clean solution, we can try to optimize
153 it, so that the server won't be slow down too much.
155 * RFC 2616 violations.
156 Closed PRs: 15852, 15857, 15859, 15861, 15864, 15869, 15870, 16120,
157 16125, 16135, 16136, 16137, 16138, 16139, 16140, 16518,
159 Open PRs: 15865, 15866, 15868, 16126, 16133, 16142, 16521, 42978
160 jerenkrantz says: need to decide how many we need to backport and/or
161 if these rise to showstopper status.
162 wrowe suggests: it would be nice to see "MUST" v.s. "SHOULD" v.s. "MAY"
163 out of this list, without reviewing them individually.
165 * pipes deadlock on all platforms with limited pipe buffers (e.g. both
166 Linux and Win32, as opposed to only Win32 on 1.3). The right solution
167 is either GStein's proposal for a "CGI Brigade", or OtherBill's proposal
168 for "Poll Buckets" for "Polling Filter Chains". Or maybe both :-)
170 * All handlers should always send content down even if r->header_only
171 is set. If not, it means that the HEAD requests don't generate the
172 same headers as a GET which is wrong.
174 * exec cmd and suexec arg-passing enhancements
175 Status: Patches proposed
176 Message-ID: <20020526041748.A29148@prodigy.Redbrick.DCU.IE>
177 (see the "proc.patch" and "suexec-shell.patch" links in this message)
179 * The 2.0.36 worker MPM graceless shutdown changes work but are
180 a bit clunky on some platforms; eg, on Linux, the loop to
181 join each worker thread seems to hang, and the parent ends up
182 killing off the child with SIGKILL. But at least it shuts down.
184 chrisd: Has this been fixed by the changes for PR 38737?
186 * We do not properly substitute the prefix-variables in the configuration
187 scripts or generated-configs. (i.e. if sysconfdir is etc,
188 httpd-std.conf points to conf.)
190 * If any request gets through ap_process_request_internal() and is
191 scheduled to be served by the core handler, without a flag that this
192 r->filename was tested by dir/file_walk, we need to 500 at the very
193 end of the ap_process_request_internal() processing so sub_req-esters
194 know this request cannot be run. This provides authors of older
195 modules better compatibility, while still improving the security and
198 Status: still need to decide where this goes, OtherBill comments...
199 Message-ID: <065701c14526$495203b0$96c0b0d0@roweclan.net>
200 [Deleted comments regarding the ap_run_handler phase, as irrelevant
201 as BillS points out that "common case will be caught in
202 default_handler already (with the r->finfo.filetype == 0 check)"
203 and the issue is detecting this -before- we try to run the req.]
205 gregames says: can this happen somehow without a broken module
206 being involved? If not, why waste cycles trying to defend against
207 potential broken modules? It seems futile.
208 wrowe counters: no, it shouldn't happen unless the module is broken.
209 But the right answer is to fail the request up-front in dir/file
210 walk if the path was entirely invalid; and we can't do that either
211 UNTIL 2.1 or we break modules that haven't hooked map_to_storage.
213 * Can a static httpd be built reliably?
214 Message-ID: <20020207142751.T31582@clove.org>
216 * Usage of APR_BRIGADE_NORMALIZE in core_input_filter should be
219 <Pine.LNX.4.33.0201202232430.318-100000@deepthought.cs.virginia.edu>
220 Jeff wonders if we still care about this. It is no longer an
221 API issue but simply an extra trip through the brigade.
223 * Try to get libtool inter-library dependency code working on AIX.
224 Message-ID: <cm3n10lx555.fsf@rdu163-40-092.nc.rr.com>
226 Justin says: If we get it working on AIX, we can enable this
227 on all platforms and clean up our build system somewhat.
228 Jeff says: I thought I tested a patch for you sometime in
229 January that you were going to commit within a few days.
231 * Handling of %2f in URIs. Currently both 1.3 and 2.0
232 completely disallow %2f in the request URI path (see
233 ap_unescape_url() in util.c). It's permitted and passed
234 through in the query string, however. Roy says the
235 original reason for disallowing it, from five years ago,
236 was to protect CGI scripts that applied PATH_INFO to
237 a filesystem location and which might be tricked by
238 ..%2f..%2f(...). We *should* allow path-info of the
239 form 'http://foo.com/index.cgi/path/to/path%2finfo'.
240 Since we've revamped a lot of our processing of path
241 segments, it would be nice to allow this, or at least
242 allow it conditionally with a directive.
244 OtherBill adds that %2f as the SECOND character of a multibyte
245 sequence causes the request to fail! This happens notably in
248 * There is increasing demand from module writers for an API
249 that will allow them to control the server à la apachectl.
250 Reasons include sole-function servers that need to die if
251 an external dependency (e.g., a database) fails, et cetera.
252 Perhaps something in the (ever more abused) scoreboard?
254 On the other hand, we already have a pipe that goes between parent
255 and child for graceful shutdown events, along with an API that
256 can be used to send a message down that pipe. In threaded MPMs,
257 it is easy enough to make that one pipe be used for graceful
258 and graceless events, and it is also easy to open that pipe
259 to both parent and child for writing. Then we just need to
260 figure out how to do graceless on non-threaded MPMs.
262 * Allow the DocumentRoot directive within <Location > scopes? This
263 allows the beloved (crusty) Alias /foo/ /somepath/foo/ followed
264 by a <Directory /somepath/foo> to become simply
265 <Location /foo/> DocumentRoot /somefile/foo (IMHO a bit more legible
266 and in-your-face.) DocumentRoot unset would be accepted [and would
267 not permit content to be served, only virtual resources such as
268 server-info or server-status.
269 This proposed change would _not_ depricate Alias.
270 striker: See the thread starting with Message-ID:
271 JLEGKKNELMHCJPNMOKHOGEEJFBAA.striker@apache.org.
273 * Win32: Rotatelogs sometimes is not terminated when Apache
274 goes down hard. FirstBill was looking at possibly tracking the
275 child's-child processes in the parent process.
276 stoddard: Shared scoreboard might offer a good way for the parent
277 to keep track of 'other child' processes and whack them if the child
279 Other thoughts on walking the process chain using the NT kernel
280 have also been proposed on APR.
282 * Eliminate unnecessary creation of pipes in mod_cgid
284 * Combine log_child and piped_log_spawn. Clean up http_log.c.
287 * Platforms that do not support fork (primarily Win32 and AS/400)
288 Architect start-up code that avoids initializing all the modules
289 in the parent process on platforms that do not support fork.
291 * There are still a number of places in the code where we are
292 losing error status (i.e. throwing away the error returned by a
293 system call and replacing it with a generic error code)
295 * Mass vhosting version of suEXEC.
297 * All DBMs suffer from confusion in support/dbmmanage (perl script) since
298 the dbmmanage employs the first-matched dbm format. This is not
299 necessarily the library that Apache was built with. Aught to
300 rewrite dbmmanage upon installation to bin/ with the proper library
301 for predictable mod_auth_dbm administration.
302 Questions; htdbm exists, time to kill dbmmanage, or does it remain
303 useful as a perl dbm management example? If we keep it,
304 do we address the issue above?
307 Some additional items remaining:
308 - case_preserved_filename stuff
309 (use the new canonical name stuff?)
310 - find a new home for ap_text(_header)
311 - is it possible to remove the DAV: namespace stuff from util_xml?
313 * ap_core_translate() and its use by mod_mmap_static and mod_file_cache
314 are a bit wonky. The function should probably be exposed as a utility
315 function (such as ap_translate_url2fs() or ap_validate_fs_url() or
316 something). Another approach would be a new hook phase after
317 "translate" which would allow the module to munge what the
318 translation has decided to do.
319 Status: Greg +1 (volunteers)
321 * Explore use of a post-config hook for the code in http_main.c which
322 calls ap_fixup_virutal_hosts(), ap_fini_vhost_config(), and
323 ap_sort_hooks() [to reduce the logic in main()]
325 * read the config tree just once, and process N times (as necessary)
327 * (possibly) use UUIDs in mod_unique_id and/or mod_usertrack
329 * (possibly) port the bug fix for PR 6942 (segv when LoadModule is put
330 into a VirtualHost container) to 2.0.
332 * shift stuff to mod_core.h
334 * callers of ap_run_create_request() should check the return value
335 for failure (Doug volunteers)
337 * Fix the worker MPM to use POD to kill child processes instead
338 of ap_os_killpg, regardless of how they should die.
340 chrisd: Is this done, by any chance? See r92598 and r93358.
342 * Scoreboard structures could be changed in the future such that
343 proper alignment is not maintained, leading to segfaults on
344 some systems. Cliff posted a patch to deal with this issue but
345 later recanted. See this message to dev@apr.apache.org:
347 <Pine.LNX.4.44.0203011354090.16457-200000@deepthought.cs.virginia.edu>
349 * APXS either needs to be fixed completely for use when apr is out of tree,
350 or it should drop query mode altogether, and we just grow an
351 httpd-config or similar arrangement.
352 To quote a discussion in STATUS earlier:
354 thommay: this doesn't fix all the problems with apxs and out of
355 tree apr/apr-util, but it's a good start. There's still the
356 query cases; but I'm beginning to think that in these cases
357 the app should be querying ap{r,u}-config directly
358 deprecate -q: add htpd-config: gstein, pquerna, minfrin, pgollucci
361 TODO ISSUES REMAINING IN MOD_SSL:
363 * Do we need SSL_set_read_ahead()?
365 * SSLRequire directive (parsing of) leaks memory
367 * Diffie-Hellman-Parameters for temporary keys are hardcoded in
368 ssl_engine_dh.c, while the comment in ssl_engine_kernel.c says:
369 "it is suggested that keys be changed daily or every 500
370 transactions, and more often if possible."
372 * ssl_var_lookup could be rewritten to be MUCH faster
374 * CRL callback should be pluggable
376 * session cache store should be pluggable
378 * init functions should return status code rather than ssl_die()
380 * ssl_engine_pphrase.c needs to be reworked so it is generic enough
381 to also decrypt proxy keys
383 * output warning when allowing SSL v2.0 ? its so old
386 * mod_proxy: Ability to run SSL over proxy gateway connections,
387 encrypting (or reencrypting) at the proxy.
389 * mod_cache: Handle ESI tags.
391 * mod_cache: Resolve issue of how to cache page fragments (or perhaps
392 -if- we want to cache page fragments). Today, mod_cache/mod_mem_cache
393 will cache #include 'virtual' requests (but not #include 'file'
394 requests). This was accomplished by making CACHE_IN a
395 CONTENT_SET-1 filter to force it to run before the SUBREQ_CORE
396 filter. But now responses cannot be cached that include the
397 effects of having been run through CONTENT_SET filters
398 (mod_deflate, mod_expires, etc). We could rerun all the
399 CONTENT_SET filters on the cached response, but this will not
400 work in all cases. For example, mod_expires relies on installing
401 the EXPIRATION filter during fixups. Contents served out of
402 mod_cache (out of the quick_handler) bypass -all- the request
403 line server hooks (Ryan really hated this. It is great for
404 performance, but bad because of the complications listed above).
406 mod_cache/mod_mem_cache/mod_cache_disk:
408 * mod_mem_cache: Consider adding a RevalidateTimeout directive to
409 specify time at which local cached content is to be revalidated
410 (ie, underlying file stat'ed to see if it has changed).
412 * mod_cache: CacheEnable/CacheDisable should accept regular expressions.
413 jerenkrantz says: Too slow. Get regexs away from speedy caches by
414 default. Introduce a new CacheEnableRegex if you want.
416 * mod_mem_cache/mod_cache_disk: Need to be able to query cache
417 status (num of entries, cache object properties, etc.).
418 mod_status could be extended to query optional hooks defined
419 by modules for the purpose of reporting module status.
420 mod_cache (et. al.) could define optional hooks that are called
421 to collect status. Status should be queryable by
423 jerenkrantz says: Yawn. Who cares.
425 * Regex containers don't work in an intutive way
426 Status: No one has come up with an efficient way to fix this
427 behavior. Dean has suggested getting rid of regex containers
429 OtherBill suggests: We at least seem to agree on eliminating
430 the <Container ~ foo> forms, and using only
431 <ContainerMatch foo> semantics.
433 * orig_ct in the byterange/multipart handling may not be
434 needed. Apache 1.3 just never stashed "multipart" into
435 r->content_type. We should probably follow suit since the
436 byterange stuff doesn't want the rest of the code to see the
437 multipart content-type; the other code should still think it is
438 dealing with the <orig_ct> stuff.
439 Status: Greg volunteers to investigate (esp. since he was most
440 likely the one to break it :-)
442 EXPERIMENTAL MODULES:
444 Experimental modules should eventually be be promoted to fully supported
445 status or removed from the repository entirely (ie, the
446 'experiment' failed). This section tracks what needs to happen to
447 get the modules promoted to fully supported status.