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/1.3.x/STATUS
21 * http://svn.apache.org/repos/asf/httpd/httpd/branches/2.0.x/STATUS
22 * 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 2.3.6 : In development
30 2.3.5 : Released on January 26, 2010.
31 2.3.4 : Released on December 8, 2009.
32 2.3.3 : Tagged on November 11, 2009, not released.
33 2.3.2 : Tagged on March 23, 2009, not released.
34 2.3.1 : Tagged on January 2, 2009, not released.
35 2.3.0 : Tagged on December 6, 2008, not released.
37 Contributors looking for a mission:
39 * Just do an egrep on "TODO" or "XXX" in the source.
41 * Review the bug database at: http://issues.apache.org/bugzilla/
43 * Review the "PatchAvailable" bugs in the bug database:
45 https://issues.apache.org/bugzilla/buglist.cgi?bug_status=NEW&bug_status=ASSIGNED&bug_status=REOPENED&product=Apache+httpd-2&keywords=PatchAvailable
47 After testing, you can append a comment saying "Reviewed and tested".
49 * Open bugs in the bug database.
52 CURRENT RELEASE NOTES:
61 * Modules without documentation need to be moved to experimental or be
64 * There is no working equivalent to 'Satisfy any' to authorize by
66 http://mail-archives.apache.org/mod_mbox/httpd-dev/200912.mbox/<4B28E73C.4050209%40kippdata.de>
68 * Not all MPMs are updated to set conn_rec::current_thread correctly.
69 (Prefork, Worker, Event, Simple are updated).
74 OLD ISSUES THAT WERE THOUGHT TO BE SHOWSTOPPERS FOR 2.2 BUT OBVIOUSLY WEREN'T:
76 * Handling of non-trailing / config by non-default handler is broken
77 http://marc.theaimsgroup.com/?l=apache-httpd-dev&m=105451701628081&w=2
78 jerenkrantz asks: Why should this block a release?
79 wsanchez agrees: this may be a change in behavior, but isn't
80 clearly wrong, and even if so, it doesn't seem like a
83 * the edge connection filter cannot be removed
84 http://marc.theaimsgroup.com/?l=apache-httpd-dev&m=105366252619530&w=2
86 jerenkrantz asks: Why should this block a release?
88 stas replies: because it requires a rewrite of the filters stack
89 implementation (you have suggested that) and once 2.2 is
90 released you can't do that anymore.
92 pgollucci: this affects mod_perl I'm pretty sure.
96 * If the parent process dies, should the remaining child processes
97 "gracefully" self-terminate. Or maybe we should make it a runtime
98 option, or have a concept of 2 parent processes (one being a
100 See: Message-ID: <3C58232C.FE91F19F@Golux.Com>
102 Self-destruct: Ken, Martin, Lars
103 Not self-destruct: BrianP, Ian, Cliff, BillS
104 Make it runtime configurable: Aaron, jim, Justin, wrowe, rederpj, nd, pgollucci
106 /* The below was a concept on *how* to handle the problem */
107 Have 2 parents: +1: jim
108 -1: Justin, wrowe, rederpj, nd, pgollucci
109 +0: Lars, Martin (while standing by, could it do
112 * Make the worker MPM the default MPM for threaded Unix boxes.
113 +1: Justin, Ian, Cliff, BillS, striker, wrowe, nd, pgollucci
114 +0: BrianP, Aaron (mutex contention is looking better with the
115 latest code, let's continue tuning and testing), rederpj, jim
118 pquerna: Do we want to change this for *2.4*?
121 * Name the Server (version 2.4 or 3.0, depending on the final call)
122 Recent discussion indicates we should designate a (short name).
123 This is not yet a [Vote] - Your nominations please:
124 * Apache HTTP Server (httpd)
126 RELEASE NON-SHOWSTOPPERS BUT WOULD BE REAL NICE TO WRAP THESE UP:
128 * Patches submitted to the bug database:
129 http://issues.apache.org/bugzilla/buglist.cgi?bug_status=NEW&bug_status=ASSIGNED&bug_status=REOPENED&product=Apache+httpd-2&keywords=PatchAvailable
131 * Filter stacks and subrequests, redirects and fast redirects.
132 There's at least one PR that suffers from the current unclean behaviour
133 (which lets the server send garbage): PR 17629
134 nd says: Every subrequest should get its own filter stack with the
135 subreq_core filter as bottom-most. That filter does two things:
136 - swallow EOS buckets
137 - redirect the data stream to the upper request's (rr->main)
138 filter chain directly after the subrequest's starting
140 Once we have a clean solution, we can try to optimize
141 it, so that the server won't be slow down too much.
143 * RFC 2616 violations.
145 Open PRs: 15852, 15859, 15861, 15864, 15865, 15866, 15868, 15869,
146 15870, 16120, 16125, 16126, 16133, 16135, 16136, 16137,
147 16138, 16139, 16140, 16142, 16518, 16520, 16521,
148 jerenkrantz says: need to decide how many we need to backport and/or
149 if these rise to showstopper status.
150 wrowe suggests: it would be nice to see "MUST" v.s. "SHOULD" v.s. "MAY"
151 out of this list, without reviewing them individually.
153 * There is a bug in how we sort some hooks, at least the pre-config
154 hook. The first time we call the hooks, they are in the correct
155 order, but the second time, we don't sort them correctly. Currently,
156 the modules/http/config.m4 file has been renamed to
157 modules/http/config2.m4 to work around this problem, it should moved
158 back when this is fixed.
160 OtherBill offers that this is a SERIOUS problem. We do not sort
161 correctly by the ordering arguments passed to the register hook
162 functions. This was proven when I reordered the open_logs hook
163 to attempt to open the error logs prior to the access logs. Possibly
164 the entire sorting code needs to be refactored.
166 * pipes deadlock on all platforms with limited pipe buffers (e.g. both
167 Linux and Win32, as opposed to only Win32 on 1.3). The right solution
168 is either GStein's proposal for a "CGI Brigade", or OtherBill's proposal
169 for "Poll Buckets" for "Polling Filter Chains". Or maybe both :-)
171 * All handlers should always send content down even if r->header_only
172 is set. If not, it means that the HEAD requests don't generate the
173 same headers as a GET which is wrong.
175 * exec cmd and suexec arg-passing enhancements
176 Status: Patches proposed
177 Message-ID: <20020526041748.A29148@prodigy.Redbrick.DCU.IE>
178 (see the "proc.patch" and "suexec-shell.patch" links in this message)
180 * The 2.0.36 worker MPM graceless shutdown changes work but are
181 a bit clunky on some platforms; eg, on Linux, the loop to
182 join each worker thread seems to hang, and the parent ends up
183 killing off the child with SIGKILL. But at least it shuts down.
185 chrisd: Has this been fixed by the changes for PR 38737?
187 * We do not properly substitute the prefix-variables in the configuration
188 scripts or generated-configs. (i.e. if sysconfdir is etc,
189 httpd-std.conf points to conf.)
191 * If any request gets through ap_process_request_internal() and is
192 scheduled to be served by the core handler, without a flag that this
193 r->filename was tested by dir/file_walk, we need to 500 at the very
194 end of the ap_process_request_internal() processing so sub_req-esters
195 know this request cannot be run. This provides authors of older
196 modules better compatibility, while still improving the security and
199 Status: still need to decide where this goes, OtherBill comments...
200 Message-ID: <065701c14526$495203b0$96c0b0d0@roweclan.net>
201 [Deleted comments regarding the ap_run_handler phase, as irrelevant
202 as BillS points out that "common case will be caught in
203 default_handler already (with the r->finfo.filetype == 0 check)"
204 and the issue is detecting this -before- we try to run the req.]
206 gregames says: can this happen somehow without a broken module
207 being involved? If not, why waste cycles trying to defend against
208 potential broken modules? It seems futile.
209 wrowe counters: no, it shouldn't happen unless the module is broken.
210 But the right answer is to fail the request up-front in dir/file
211 walk if the path was entirely invalid; and we can't do that either
212 UNTIL 2.1 or we break modules that haven't hooked map_to_storage.
214 * Can a static httpd be built reliably?
215 Message-ID: <20020207142751.T31582@clove.org>
217 * Usage of APR_BRIGADE_NORMALIZE in core_input_filter should be
220 <Pine.LNX.4.33.0201202232430.318-100000@deepthought.cs.virginia.edu>
221 Jeff wonders if we still care about this. It is no longer an
222 API issue but simply an extra trip through the brigade.
224 * Try to get libtool inter-library dependency code working on AIX.
225 Message-ID: <cm3n10lx555.fsf@rdu163-40-092.nc.rr.com>
227 Justin says: If we get it working on AIX, we can enable this
228 on all platforms and clean up our build system somewhat.
229 Jeff says: I thought I tested a patch for you sometime in
230 January that you were going to commit within a few days.
232 * Handling of %2f in URIs. Currently both 1.3 and 2.0
233 completely disallow %2f in the request URI path (see
234 ap_unescape_url() in util.c). It's permitted and passed
235 through in the query string, however. Roy says the
236 original reason for disallowing it, from five years ago,
237 was to protect CGI scripts that applied PATH_INFO to
238 a filesystem location and which might be tricked by
239 ..%2f..%2f(...). We *should* allow path-info of the
240 form 'http://foo.com/index.cgi/path/to/path%2finfo'.
241 Since we've revamped a lot of our processing of path
242 segments, it would be nice to allow this, or at least
243 allow it conditionally with a directive.
245 OtherBill adds that %2f as the SECOND character of a multibyte
246 sequence causes the request to fail! This happens notably in
249 * There is increasing demand from module writers for an API
250 that will allow them to control the server à la apachectl.
251 Reasons include sole-function servers that need to die if
252 an external dependency (e.g., a database) fails, et cetera.
253 Perhaps something in the (ever more abused) scoreboard?
255 On the other hand, we already have a pipe that goes between parent
256 and child for graceful shutdown events, along with an API that
257 can be used to send a message down that pipe. In threaded MPMs,
258 it is easy enough to make that one pipe be used for graceful
259 and graceless events, and it is also easy to open that pipe
260 to both parent and child for writing. Then we just need to
261 figure out how to do graceless on non-threaded MPMs.
263 * Allow the DocumentRoot directive within <Location > scopes? This
264 allows the beloved (crusty) Alias /foo/ /somepath/foo/ followed
265 by a <Directory /somepath/foo> to become simply
266 <Location /foo/> DocumentRoot /somefile/foo (IMHO a bit more legible
267 and in-your-face.) DocumentRoot unset would be accepted [and would
268 not permit content to be served, only virtual resources such as
269 server-info or server-status.
270 This proposed change would _not_ depricate Alias.
271 striker: See the thread starting with Message-ID:
272 JLEGKKNELMHCJPNMOKHOGEEJFBAA.striker@apache.org.
274 * Win32: Rotatelogs sometimes is not terminated when Apache
275 goes down hard. FirstBill was looking at possibly tracking the
276 child's-child processes in the parent process.
277 stoddard: Shared scoreboard might offer a good way for the parent
278 to keep track of 'other child' processes and whack them if the child
280 Other thoughts on walking the process chain using the NT kernel
281 have also been proposed on APR.
283 * Eliminate unnecessary creation of pipes in mod_cgid
285 * Combine log_child and piped_log_spawn. Clean up http_log.c.
288 * Platforms that do not support fork (primarily Win32 and AS/400)
289 Architect start-up code that avoids initializing all the modules
290 in the parent process on platforms that do not support fork.
292 * There are still a number of places in the code where we are
293 losing error status (i.e. throwing away the error returned by a
294 system call and replacing it with a generic error code)
296 * Mass vhosting version of suEXEC.
298 * All DBMs suffer from confusion in support/dbmmanage (perl script) since
299 the dbmmanage employs the first-matched dbm format. This is not
300 necessarily the library that Apache was built with. Aught to
301 rewrite dbmmanage upon installation to bin/ with the proper library
302 for predictable mod_auth_dbm administration.
303 Questions; htdbm exists, time to kill dbmmanage, or does it remain
304 useful as a perl dbm management example? If we keep it,
305 do we address the issue above?
308 Some additional items remaining:
309 - case_preserved_filename stuff
310 (use the new canonical name stuff?)
311 - find a new home for ap_text(_header)
312 - is it possible to remove the DAV: namespace stuff from util_xml?
314 * ap_core_translate() and its use by mod_mmap_static and mod_file_cache
315 are a bit wonky. The function should probably be exposed as a utility
316 function (such as ap_translate_url2fs() or ap_validate_fs_url() or
317 something). Another approach would be a new hook phase after
318 "translate" which would allow the module to munge what the
319 translation has decided to do.
320 Status: Greg +1 (volunteers)
322 * Explore use of a post-config hook for the code in http_main.c which
323 calls ap_fixup_virutal_hosts(), ap_fini_vhost_config(), and
324 ap_sort_hooks() [to reduce the logic in main()]
326 * read the config tree just once, and process N times (as necessary)
328 * (possibly) use UUIDs in mod_unique_id and/or mod_usertrack
330 * (possibly) port the bug fix for PR 6942 (segv when LoadModule is put
331 into a VirtualHost container) to 2.0.
333 * shift stuff to mod_core.h
335 * callers of ap_run_create_request() should check the return value
336 for failure (Doug volunteers)
338 * Fix the worker MPM to use POD to kill child processes instead
339 of ap_os_killpg, regardless of how they should die.
341 chrisd: Is this done, by any chance? See r92598 and r93358.
343 * Scoreboard structures could be changed in the future such that
344 proper alignment is not maintained, leading to segfaults on
345 some systems. Cliff posted a patch to deal with this issue but
346 later recanted. See this message to dev@apr.apache.org:
348 <Pine.LNX.4.44.0203011354090.16457-200000@deepthought.cs.virginia.edu>
350 * APXS either needs to be fixed completely for use when apr is out of tree,
351 or it should drop query mode altogether, and we just grow an
352 httpd-config or similar arrangement.
353 To quote a discussion in STATUS earlier:
355 thommay: this doesn't fix all the problems with apxs and out of
356 tree apr/apr-util, but it's a good start. There's still the
357 query cases; but I'm beginning to think that in these cases
358 the app should be querying ap{r,u}-config directly
359 deprecate -q: add htpd-config: gstein, pquerna, minfrin, pgollucci
362 TODO ISSUES REMAINING IN MOD_SSL:
364 * Do we need SSL_set_read_ahead()?
366 * the ssl_expr api is NOT THREAD SAFE. race conditions exist:
367 -in ssl_expr_comp() if SSLRequire is used in .htaccess
368 (ssl_expr_info is global)
369 -is ssl_expr_eval() if there is an error
370 (ssl_expr_error is global)
372 * SSLRequire directive (parsing of) leaks memory
374 * Diffie-Hellman-Parameters for temporary keys are hardcoded in
375 ssl_engine_dh.c, while the comment in ssl_engine_kernel.c says:
376 "it is suggested that keys be changed daily or every 500
377 transactions, and more often if possible."
379 * ssl_var_lookup could be rewritten to be MUCH faster
381 * CRL callback should be pluggable
383 * session cache store should be pluggable
385 * init functions should return status code rather than ssl_die()
387 * ssl_engine_pphrase.c needs to be reworked so it is generic enough
388 to also decrypt proxy keys
390 * output warning when allowing SSL v2.0 ? its so old
393 * mod_proxy: Ability to run SSL over proxy gateway connections,
394 encrypting (or reencrypting) at the proxy.
396 * mod_cache: Handle ESI tags.
398 * mod_cache: Resolve issue of how to cache page fragements (or perhaps
399 -if- we want to cache page fragements). Today, mod_cache/mod_mem_cache
400 will cache #include 'virtual' requests (but not #include 'file'
401 requests). This was accomplished by making CACHE_IN a
402 CONTENT_SET-1 filter to force it to run before the SUBREQ_CORE
403 filter. But now responses cannot be cached that include the
404 effects of having been run through CONTENT_SET filters
405 (mod_deflate, mod_expires, etc). We could rerun all the
406 CONTENT_SET filters on the cached response, but this will not
407 work in all cases. For example, mod_expires relies on installing
408 the EXPIRATION filter during fixups. Contents served out of
409 mod_cache (out of the quick_handler) bypass -all- the request
410 line server hooks (Ryan really hated this. It is great for
411 performance, but bad because of the complications listed above).
413 mod_cache/mod_mem_cache/mod_disk_cache:
415 * mod_mem_cache: Consider adding a RevalidateTimeout directive to
416 specify time at which local cached content is to be revalidated
417 (ie, underlying file stat'ed to see if it has changed).
419 * mod_cache: CacheEnable/CacheDisable should accept regular expressions.
420 jerenkrantz says: Too slow. Get regexs away from speedy caches by
421 default. Introduce a new CacheEnableRegex if you want.
423 * mod_mem_cache/mod_disk_cache: Need to be able to query cache
424 status (num of entries, cache object properties, etc.).
425 mod_status could be extended to query optional hooks defined
426 by modules for the purpose of reporting module status.
427 mod_cache (et. al.) could define optional hooks that are called
428 to collect status. Status should be queryable by
430 jerenkrantz says: Yawn. Who cares.
432 * MaxRequestsPerChild measures connections, not requests.
433 Until someone has a better way, we'll probably just rename it
434 "MaxConnectionsPerChild".
436 * Regex containers don't work in an intutive way
437 Status: No one has come up with an efficient way to fix this
438 behavior. Dean has suggested getting rid of regex containers
440 OtherBill suggests: We at least seem to agree on eliminating
441 the <Container ~ foo> forms, and using only
442 <ContainerMatch foo> semantics.
444 * orig_ct in the byterange/multipart handling may not be
445 needed. Apache 1.3 just never stashed "multipart" into
446 r->content_type. We should probably follow suit since the
447 byterange stuff doesn't want the rest of the code to see the
448 multipart content-type; the other code should still think it is
449 dealing with the <orig_ct> stuff.
450 Status: Greg volunteers to investigate (esp. since he was most
451 likely the one to break it :-)
453 EXPERIMENTAL MODULES:
455 Experimental modules should eventually be be promoted to fully supported
456 status or removed from the repository entirely (ie, the
457 'experiment' failed). This section tracks what needs to happen to
458 get the modules promoted to fully supported status.