1 APACHE 2.4 STATUS: -*- mode: text coding: utf-8 -*-
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/branches/2.4.x/STATUS
8 Documentation status is maintained separately and can be found at:
10 * docs/STATUS in this source tree, or
11 * http://svn.apache.org/repos/asf/httpd/httpd/branches/2.4.x/docs/STATUS
13 The current development branch of this software can be found at:
15 * http://svn.apache.org/repos/asf/httpd/httpd/trunk
17 Consult the following STATUS files for information on related projects:
19 * http://svn.apache.org/repos/asf/apr/apr/trunk/STATUS
20 * http://svn.apache.org/repos/asf/apr/apr/branches/1.4.x/STATUS
21 * http://svn.apache.org/repos/asf/apr/apr-util/branches/1.4.x/STATUS
23 Patches considered for backport are noted in their branches' STATUS:
25 * http://svn.apache.org/repos/asf/httpd/httpd/branches/2.0.x/STATUS
26 * http://svn.apache.org/repos/asf/httpd/httpd/branches/2.2.x/STATUS
27 * http://svn.apache.org/repos/asf/httpd/httpd/branches/2.4.x/STATUS
32 [NOTE that x.{odd}.z versions are strictly Alpha/Beta releases,
33 while x.{even}.z versions are Stable/GA releases.]
35 2.4.0 : In Development.
36 2.3.15 : Tagged on November 8, 2011.
37 2.3.14 : Tagged on August 1, 2011. Released Aug. 9, 2011.
38 2.3.13 : Tagged on June 28, 2011, not released.
39 2.3.12 : Tagged on May 11, 2011. Released May 23, 2011.
40 2.3.11 : Released as Beta on March 7, 2011.
41 2.3.10 : Tagged on December 13, 2010. Released Dec 21, 2010.
42 2.3.9 : Tagged on November 23, 2010, not released.
43 2.3.8 : Tagged on August 24, 2010.
44 2.3.7 : Tagged on August 19, 2010, not released.
45 2.3.6 : Released on June 21, 2010.
46 2.3.5 : Released on January 26, 2010.
47 2.3.4 : Released on December 8, 2009.
48 2.3.3 : Tagged on November 11, 2009, not released.
49 2.3.2 : Tagged on March 23, 2009, not released.
50 2.3.1 : Tagged on January 2, 2009, not released.
51 2.3.0 : Tagged on December 6, 2008, not released.
53 Contributors looking for a mission:
55 * Just do an egrep on "TODO" or "XXX" in the source.
57 * Review the bug database at: http://issues.apache.org/bugzilla/
59 * Review the "PatchAvailable" bugs in the bug database:
61 https://issues.apache.org/bugzilla/buglist.cgi?bug_status=NEW&bug_status=ASSIGNED&bug_status=REOPENED&product=Apache+httpd-2&keywords=PatchAvailable
63 After testing, you can append a comment saying "Reviewed and tested".
65 * Open bugs in the bug database.
67 * See also the STATUS file in the docs/ directory, which lists documentation-specific TODO items.
70 CURRENT RELEASE NOTES:
72 * Forward binary compatibility is expected of Apache 2.4.x releases, such
73 that no MMN major number changes will occur after 2.4.0. Such changes canr
74 only be made in the trunk.
76 * All commits to branches/2.4.x must be reflected in SVN trunk,
77 as well, if they apply. Logical progression is commit to trunk
78 then merge into branches/2.4.x, as applicable. The 2.4.x branch is
84 Jim proposes GA in late November (at the latest). Jim volunteers to RM these.
91 * Modules that are not ready for production use must be removed.
92 The same for modules without documentation.
94 See https://svn.apache.org/repos/asf/httpd/httpd/branches/2.4.x/docs/STATUS
95 under 'modules docs' for a comprehensive list of undocumented modules.
97 NEW ISSUES THAT WOULD BE NICE TO HAVE DONE IN 2.4 BUT ARE NOT BLOCKERS
99 * The mod_session* modules need to be checked that their hooks respect
100 the returning of int (HTTP status codes) and apr_status_t as appropriate,
101 and any anomolies fixed.
102 jim sez: from what I can see, mod_session* is no worse that other
103 modules that mix these 2 types... clean up is
104 forthcoming but should not be considered a blocker, imo
106 wrowe asks: what's the API change required?
107 wrowe asks; why are we shipping this if it requires apr_ssl
109 * mod_ssl's proxy support only allows one proxy client certificate per
110 frontend virtual host. Lift this restriction.
111 jim sez: Why a blocker?, pgollucci +1 jim
112 wrowe asks: what's the API change required?
114 * INCLUDE mod_fcgid with 2.4.0, esp to help php users etc to enjoy
115 a painless event mpm experience.
116 sf: can be done in 2.4.x>0
118 OLD ISSUES THAT WERE THOUGHT TO BE SHOWSTOPPERS FOR 2.2 BUT OBVIOUSLY WEREN'T:
120 * Handling of non-trailing / config by non-default handler is broken
121 http://marc.theaimsgroup.com/?l=apache-httpd-dev&m=105451701628081&w=2
122 jerenkrantz asks: Why should this block a release?
123 wsanchez agrees: this may be a change in behavior, but isn't
124 clearly wrong, and even if so, it doesn't seem like a
127 * the edge connection filter cannot be removed
128 http://marc.theaimsgroup.com/?l=apache-httpd-dev&m=105366252619530&w=2
130 jerenkrantz asks: Why should this block a release?
132 stas replies: because it requires a rewrite of the filters stack
133 implementation (you have suggested that) and once 2.2 is
134 released you can't do that anymore.
136 pgollucci: this affects mod_perl I'm pretty sure.
140 * Name the Server (version 2.4 or 3.0, depending on the final call)
141 Recent discussion indicates we should designate a (short name).
142 This is not yet a [Vote] - Your nominations please:
143 * Apache HTTP Server (httpd)
144 +1: sctemme (why mess with it?), pgollucci
146 RELEASE NON-SHOWSTOPPERS BUT WOULD BE REAL NICE TO WRAP THESE UP:
148 * Clean up all the kruft and *extremely* outdated stuff below...
150 * Maybe remove Limit/LimitExcept or at least make it log warnings when
153 * Patches submitted to the bug database:
154 http://issues.apache.org/bugzilla/buglist.cgi?bug_status=NEW&bug_status=ASSIGNED&bug_status=REOPENED&product=Apache+httpd-2&keywords=PatchAvailable
156 * Filter stacks and subrequests, redirects and fast redirects.
157 There's at least one PR that suffers from the current unclean behaviour
158 (which lets the server send garbage): PR 17629
159 nd says: Every subrequest should get its own filter stack with the
160 subreq_core filter as bottom-most. That filter does two things:
161 - swallow EOS buckets
162 - redirect the data stream to the upper request's (rr->main)
163 filter chain directly after the subrequest's starting
165 Once we have a clean solution, we can try to optimize
166 it, so that the server won't be slow down too much.
168 * RFC 2616 violations.
169 Closed PRs: 15852, 15857, 15859, 15861, 15864, 15869, 15870, 16120,
170 16125, 16135, 16136, 16137, 16138, 16139, 16140, 16518,
172 Open PRs: 15865, 15866, 15868, 16126, 16133, 16142, 16521, 42978
173 jerenkrantz says: need to decide how many we need to backport and/or
174 if these rise to showstopper status.
175 wrowe suggests: it would be nice to see "MUST" v.s. "SHOULD" v.s. "MAY"
176 out of this list, without reviewing them individually.
178 * pipes deadlock on all platforms with limited pipe buffers (e.g. both
179 Linux and Win32, as opposed to only Win32 on 1.3). The right solution
180 is either GStein's proposal for a "CGI Brigade", or OtherBill's proposal
181 for "Poll Buckets" for "Polling Filter Chains". Or maybe both :-)
183 * All handlers should always send content down even if r->header_only
184 is set. If not, it means that the HEAD requests don't generate the
185 same headers as a GET which is wrong.
187 * exec cmd and suexec arg-passing enhancements
188 Status: Patches proposed
189 Message-ID: <20020526041748.A29148@prodigy.Redbrick.DCU.IE>
190 (see the "proc.patch" and "suexec-shell.patch" links in this message)
192 * The 2.0.36 worker MPM graceless shutdown changes work but are
193 a bit clunky on some platforms; eg, on Linux, the loop to
194 join each worker thread seems to hang, and the parent ends up
195 killing off the child with SIGKILL. But at least it shuts down.
197 chrisd: Has this been fixed by the changes for PR 38737?
199 * We do not properly substitute the prefix-variables in the configuration
200 scripts or generated-configs. (i.e. if sysconfdir is etc,
201 httpd-std.conf points to conf.)
203 * If any request gets through ap_process_request_internal() and is
204 scheduled to be served by the core handler, without a flag that this
205 r->filename was tested by dir/file_walk, we need to 500 at the very
206 end of the ap_process_request_internal() processing so sub_req-esters
207 know this request cannot be run. This provides authors of older
208 modules better compatibility, while still improving the security and
211 Status: still need to decide where this goes, OtherBill comments...
212 Message-ID: <065701c14526$495203b0$96c0b0d0@roweclan.net>
213 [Deleted comments regarding the ap_run_handler phase, as irrelevant
214 as BillS points out that "common case will be caught in
215 default_handler already (with the r->finfo.filetype == 0 check)"
216 and the issue is detecting this -before- we try to run the req.]
218 gregames says: can this happen somehow without a broken module
219 being involved? If not, why waste cycles trying to defend against
220 potential broken modules? It seems futile.
221 wrowe counters: no, it shouldn't happen unless the module is broken.
222 But the right answer is to fail the request up-front in dir/file
223 walk if the path was entirely invalid; and we can't do that either
224 UNTIL 2.1 or we break modules that haven't hooked map_to_storage.
226 * Can a static httpd be built reliably?
227 Message-ID: <20020207142751.T31582@clove.org>
229 * Usage of APR_BRIGADE_NORMALIZE in core_input_filter should be
232 <Pine.LNX.4.33.0201202232430.318-100000@deepthought.cs.virginia.edu>
233 Jeff wonders if we still care about this. It is no longer an
234 API issue but simply an extra trip through the brigade.
236 * Try to get libtool inter-library dependency code working on AIX.
237 Message-ID: <cm3n10lx555.fsf@rdu163-40-092.nc.rr.com>
239 Justin says: If we get it working on AIX, we can enable this
240 on all platforms and clean up our build system somewhat.
241 Jeff says: I thought I tested a patch for you sometime in
242 January that you were going to commit within a few days.
244 * Handling of %2f in URIs. Currently both 1.3 and 2.0
245 completely disallow %2f in the request URI path (see
246 ap_unescape_url() in util.c). It's permitted and passed
247 through in the query string, however. Roy says the
248 original reason for disallowing it, from five years ago,
249 was to protect CGI scripts that applied PATH_INFO to
250 a filesystem location and which might be tricked by
251 ..%2f..%2f(...). We *should* allow path-info of the
252 form 'http://foo.com/index.cgi/path/to/path%2finfo'.
253 Since we've revamped a lot of our processing of path
254 segments, it would be nice to allow this, or at least
255 allow it conditionally with a directive.
257 OtherBill adds that %2f as the SECOND character of a multibyte
258 sequence causes the request to fail! This happens notably in
261 * There is increasing demand from module writers for an API
262 that will allow them to control the server à la apachectl.
263 Reasons include sole-function servers that need to die if
264 an external dependency (e.g., a database) fails, et cetera.
265 Perhaps something in the (ever more abused) scoreboard?
267 On the other hand, we already have a pipe that goes between parent
268 and child for graceful shutdown events, along with an API that
269 can be used to send a message down that pipe. In threaded MPMs,
270 it is easy enough to make that one pipe be used for graceful
271 and graceless events, and it is also easy to open that pipe
272 to both parent and child for writing. Then we just need to
273 figure out how to do graceless on non-threaded MPMs.
275 * Allow the DocumentRoot directive within <Location > scopes? This
276 allows the beloved (crusty) Alias /foo/ /somepath/foo/ followed
277 by a <Directory /somepath/foo> to become simply
278 <Location /foo/> DocumentRoot /somefile/foo (IMHO a bit more legible
279 and in-your-face.) DocumentRoot unset would be accepted [and would
280 not permit content to be served, only virtual resources such as
281 server-info or server-status.
282 This proposed change would _not_ depricate Alias.
283 striker: See the thread starting with Message-ID:
284 JLEGKKNELMHCJPNMOKHOGEEJFBAA.striker@apache.org.
286 * Win32: Rotatelogs sometimes is not terminated when Apache
287 goes down hard. FirstBill was looking at possibly tracking the
288 child's-child processes in the parent process.
289 stoddard: Shared scoreboard might offer a good way for the parent
290 to keep track of 'other child' processes and whack them if the child
292 Other thoughts on walking the process chain using the NT kernel
293 have also been proposed on APR.
295 * Eliminate unnecessary creation of pipes in mod_cgid
297 * Combine log_child and piped_log_spawn. Clean up http_log.c.
300 * Platforms that do not support fork (primarily Win32 and AS/400)
301 Architect start-up code that avoids initializing all the modules
302 in the parent process on platforms that do not support fork.
304 * There are still a number of places in the code where we are
305 losing error status (i.e. throwing away the error returned by a
306 system call and replacing it with a generic error code)
308 * Mass vhosting version of suEXEC.
310 * All DBMs suffer from confusion in support/dbmmanage (perl script) since
311 the dbmmanage employs the first-matched dbm format. This is not
312 necessarily the library that Apache was built with. Aught to
313 rewrite dbmmanage upon installation to bin/ with the proper library
314 for predictable mod_auth_dbm administration.
315 Questions; htdbm exists, time to kill dbmmanage, or does it remain
316 useful as a perl dbm management example? If we keep it,
317 do we address the issue above?
320 Some additional items remaining:
321 - case_preserved_filename stuff
322 (use the new canonical name stuff?)
323 - find a new home for ap_text(_header)
324 - is it possible to remove the DAV: namespace stuff from util_xml?
326 * ap_core_translate() and its use by mod_mmap_static and mod_file_cache
327 are a bit wonky. The function should probably be exposed as a utility
328 function (such as ap_translate_url2fs() or ap_validate_fs_url() or
329 something). Another approach would be a new hook phase after
330 "translate" which would allow the module to munge what the
331 translation has decided to do.
332 Status: Greg +1 (volunteers)
334 * Explore use of a post-config hook for the code in http_main.c which
335 calls ap_fixup_virutal_hosts(), ap_fini_vhost_config(), and
336 ap_sort_hooks() [to reduce the logic in main()]
338 * read the config tree just once, and process N times (as necessary)
340 * (possibly) use UUIDs in mod_unique_id and/or mod_usertrack
342 * (possibly) port the bug fix for PR 6942 (segv when LoadModule is put
343 into a VirtualHost container) to 2.0.
345 * shift stuff to mod_core.h
347 * callers of ap_run_create_request() should check the return value
348 for failure (Doug volunteers)
350 * Fix the worker MPM to use POD to kill child processes instead
351 of ap_os_killpg, regardless of how they should die.
353 chrisd: Is this done, by any chance? See r92598 and r93358.
355 * Scoreboard structures could be changed in the future such that
356 proper alignment is not maintained, leading to segfaults on
357 some systems. Cliff posted a patch to deal with this issue but
358 later recanted. See this message to dev@apr.apache.org:
360 <Pine.LNX.4.44.0203011354090.16457-200000@deepthought.cs.virginia.edu>
362 * APXS either needs to be fixed completely for use when apr is out of tree,
363 or it should drop query mode altogether, and we just grow an
364 httpd-config or similar arrangement.
365 To quote a discussion in STATUS earlier:
367 thommay: this doesn't fix all the problems with apxs and out of
368 tree apr/apr-util, but it's a good start. There's still the
369 query cases; but I'm beginning to think that in these cases
370 the app should be querying ap{r,u}-config directly
371 deprecate -q: add htpd-config: gstein, pquerna, minfrin, pgollucci
374 TODO ISSUES REMAINING IN MOD_SSL:
376 * Do we need SSL_set_read_ahead()?
378 * SSLRequire directive (parsing of) leaks memory
380 * Diffie-Hellman-Parameters for temporary keys are hardcoded in
381 ssl_engine_dh.c, while the comment in ssl_engine_kernel.c says:
382 "it is suggested that keys be changed daily or every 500
383 transactions, and more often if possible."
385 * ssl_var_lookup could be rewritten to be MUCH faster
387 * CRL callback should be pluggable
389 * init functions should return status code rather than ssl_die()
391 * ssl_engine_pphrase.c needs to be reworked so it is generic enough
392 to also decrypt proxy keys
395 * mod_proxy: Ability to run SSL over proxy gateway connections,
396 encrypting (or reencrypting) at the proxy.
398 * mod_cache: Handle ESI tags.
400 * mod_cache: Resolve issue of how to cache page fragments (or perhaps
401 -if- we want to cache page fragments). Today, mod_cache/mod_mem_cache
402 will cache #include 'virtual' requests (but not #include 'file'
403 requests). This was accomplished by making CACHE_IN a
404 CONTENT_SET-1 filter to force it to run before the SUBREQ_CORE
405 filter. But now responses cannot be cached that include the
406 effects of having been run through CONTENT_SET filters
407 (mod_deflate, mod_expires, etc). We could rerun all the
408 CONTENT_SET filters on the cached response, but this will not
409 work in all cases. For example, mod_expires relies on installing
410 the EXPIRATION filter during fixups. Contents served out of
411 mod_cache (out of the quick_handler) bypass -all- the request
412 line server hooks (Ryan really hated this. It is great for
413 performance, but bad because of the complications listed above).
415 mod_cache/mod_mem_cache/mod_cache_disk:
417 * mod_mem_cache: Consider adding a RevalidateTimeout directive to
418 specify time at which local cached content is to be revalidated
419 (ie, underlying file stat'ed to see if it has changed).
421 * mod_cache: CacheEnable/CacheDisable should accept regular expressions.
422 jerenkrantz says: Too slow. Get regexs away from speedy caches by
423 default. Introduce a new CacheEnableRegex if you want.
425 * mod_mem_cache/mod_cache_disk: Need to be able to query cache
426 status (num of entries, cache object properties, etc.).
427 mod_status could be extended to query optional hooks defined
428 by modules for the purpose of reporting module status.
429 mod_cache (et. al.) could define optional hooks that are called
430 to collect status. Status should be queryable by
432 jerenkrantz says: Yawn. Who cares.
434 * Regex containers don't work in an intutive way
435 Status: No one has come up with an efficient way to fix this
436 behavior. Dean has suggested getting rid of regex containers
438 OtherBill suggests: We at least seem to agree on eliminating
439 the <Container ~ foo> forms, and using only
440 <ContainerMatch foo> semantics.
442 * orig_ct in the byterange/multipart handling may not be
443 needed. Apache 1.3 just never stashed "multipart" into
444 r->content_type. We should probably follow suit since the
445 byterange stuff doesn't want the rest of the code to see the
446 multipart content-type; the other code should still think it is
447 dealing with the <orig_ct> stuff.
448 Status: Greg volunteers to investigate (esp. since he was most
449 likely the one to break it :-)
451 EXPERIMENTAL MODULES:
453 Experimental modules should eventually be be promoted to fully supported
454 status or removed from the repository entirely (ie, the
455 'experiment' failed). This section tracks what needs to happen to
456 get the modules promoted to fully supported status.