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 2.3.5 : Tagged on January 21, 2010.
30 2.3.4 : Tagged on November 24, 2009.
31 2.3.3 : Tagged on November 11, 2009, not released.
32 2.3.2 : Tagged on March 23, 2009, not released.
33 2.3.1 : Tagged on January 2, 2009, not released.
34 2.3.0 : Tagged on December 6, 2008, not released.
36 Contributors looking for a mission:
38 * Just do an egrep on "TODO" or "XXX" in the source.
40 * Review the bug database at: http://issues.apache.org/bugzilla/
42 * Review the "PatchAvailable" bugs in the bug database:
44 https://issues.apache.org/bugzilla/buglist.cgi?bug_status=NEW&bug_status=ASSIGNED&bug_status=REOPENED&product=Apache+httpd-2&keywords=PatchAvailable
46 After testing, you can append a comment saying "Reviewed and tested".
48 * Open bugs in the bug database.
51 CURRENT RELEASE NOTES:
56 * Not all MPMs are updated to set conn_rec::current_thread correctly.
57 (Prefork, Worker, Event, Simple are updated).
59 * Handling of non-trailing / config by non-default handler is broken
60 http://marc.theaimsgroup.com/?l=apache-httpd-dev&m=105451701628081&w=2
61 jerenkrantz asks: Why should this block a release?
62 wsanchez agrees: this may be a change in behavior, but isn't
63 clearly wrong, and even if so, it doesn't seem like a
66 * the edge connection filter cannot be removed
67 http://marc.theaimsgroup.com/?l=apache-httpd-dev&m=105366252619530&w=2
69 jerenkrantz asks: Why should this block a release?
71 stas replies: because it requires a rewrite of the filters stack
72 implementation (you have suggested that) and once 2.2 is
73 released you can't do that anymore.
75 pgollucci: this affects mod_perl I'm pretty sure.
79 * If the parent process dies, should the remaining child processes
80 "gracefully" self-terminate. Or maybe we should make it a runtime
81 option, or have a concept of 2 parent processes (one being a
83 See: Message-ID: <3C58232C.FE91F19F@Golux.Com>
85 Self-destruct: Ken, Martin, Lars
86 Not self-destruct: BrianP, Ian, Cliff, BillS
87 Make it runtime configurable: Aaron, jim, Justin, wrowe, rederpj, nd
89 /* The below was a concept on *how* to handle the problem */
90 Have 2 parents: +1: jim
91 -1: Justin, wrowe, rederpj, nd, pgollucci
92 +0: Lars, Martin (while standing by, could it do
95 * Make the worker MPM the default MPM for threaded Unix boxes.
96 +1: Justin, Ian, Cliff, BillS, striker, wrowe, nd, pgollucci
97 +0: BrianP, Aaron (mutex contention is looking better with the
98 latest code, let's continue tuning and testing), rederpj, jim
101 pquerna: Do we want to change this for *2.4*?
104 RELEASE NON-SHOWSTOPPERS BUT WOULD BE REAL NICE TO WRAP THESE UP:
106 * Patches submitted to the bug database:
107 http://issues.apache.org/bugzilla/buglist.cgi?bug_status=NEW&bug_status=ASSIGNED&bug_status=REOPENED&product=Apache+httpd-2&keywords=PatchAvailable
109 * Filter stacks and subrequests, redirects and fast redirects.
110 There's at least one PR that suffers from the current unclean behaviour
111 (which lets the server send garbage): PR 17629
112 nd says: Every subrequest should get its own filter stack with the
113 subreq_core filter as bottom-most. That filter does two things:
114 - swallow EOS buckets
115 - redirect the data stream to the upper request's (rr->main)
116 filter chain directly after the subrequest's starting
118 Once we have a clean solution, we can try to optimize
119 it, so that the server won't be slow down too much.
121 * RFC 2616 violations.
123 Open PRs: 15852, 15859, 15861, 15864, 15865, 15866, 15868, 15869,
124 15870, 16120, 16125, 16126, 16133, 16135, 16136, 16137,
125 16138, 16139, 16140, 16142, 16518, 16520, 16521,
126 jerenkrantz says: need to decide how many we need to backport and/or
127 if these rise to showstopper status.
128 wrowe suggests: it would be nice to see "MUST" v.s. "SHOULD" v.s. "MAY"
129 out of this list, without reviewing them individually.
131 * There is a bug in how we sort some hooks, at least the pre-config
132 hook. The first time we call the hooks, they are in the correct
133 order, but the second time, we don't sort them correctly. Currently,
134 the modules/http/config.m4 file has been renamed to
135 modules/http/config2.m4 to work around this problem, it should moved
136 back when this is fixed.
138 OtherBill offers that this is a SERIOUS problem. We do not sort
139 correctly by the ordering arguments passed to the register hook
140 functions. This was proven when I reordered the open_logs hook
141 to attempt to open the error logs prior to the access logs. Possibly
142 the entire sorting code needs to be refactored.
144 * pipes deadlock on all platforms with limited pipe buffers (e.g. both
145 Linux and Win32, as opposed to only Win32 on 1.3). The right solution
146 is either GStein's proposal for a "CGI Brigade", or OtherBill's proposal
147 for "Poll Buckets" for "Polling Filter Chains". Or maybe both :-)
149 * All handlers should always send content down even if r->header_only
150 is set. If not, it means that the HEAD requests don't generate the
151 same headers as a GET which is wrong.
153 * exec cmd and suexec arg-passing enhancements
154 Status: Patches proposed
155 Message-ID: <20020526041748.A29148@prodigy.Redbrick.DCU.IE>
156 (see the "proc.patch" and "suexec-shell.patch" links in this message)
158 * The 2.0.36 worker MPM graceless shutdown changes work but are
159 a bit clunky on some platforms; eg, on Linux, the loop to
160 join each worker thread seems to hang, and the parent ends up
161 killing off the child with SIGKILL. But at least it shuts down.
163 chrisd: Has this been fixed by the changes for PR 38737?
165 * --enable-mods-shared="foo1 foo2" is busted on Darwin. Pier
166 posted a patch (Message-ID: <B8DBBE8D.575A%pier@betaversion.org>).
168 * We do not properly substitute the prefix-variables in the configuration
169 scripts or generated-configs. (i.e. if sysconfdir is etc,
170 httpd-std.conf points to conf.)
172 * If any request gets through ap_process_request_internal() and is
173 scheduled to be served by the core handler, without a flag that this
174 r->filename was tested by dir/file_walk, we need to 500 at the very
175 end of the ap_process_request_internal() processing so sub_req-esters
176 know this request cannot be run. This provides authors of older
177 modules better compatibility, while still improving the security and
180 Status: still need to decide where this goes, OtherBill comments...
181 Message-ID: <065701c14526$495203b0$96c0b0d0@roweclan.net>
182 [Deleted comments regarding the ap_run_handler phase, as irrelevant
183 as BillS points out that "common case will be caught in
184 default_handler already (with the r->finfo.filetype == 0 check)"
185 and the issue is detecting this -before- we try to run the req.]
187 gregames says: can this happen somehow without a broken module
188 being involved? If not, why waste cycles trying to defend against
189 potential broken modules? It seems futile.
190 wrowe counters: no, it shouldn't happen unless the module is broken.
191 But the right answer is to fail the request up-front in dir/file
192 walk if the path was entirely invalid; and we can't do that either
193 UNTIL 2.1 or we break modules that haven't hooked map_to_storage.
195 * Can a static httpd be built reliably?
196 Message-ID: <20020207142751.T31582@clove.org>
198 * Usage of APR_BRIGADE_NORMALIZE in core_input_filter should be
201 <Pine.LNX.4.33.0201202232430.318-100000@deepthought.cs.virginia.edu>
202 Jeff wonders if we still care about this. It is no longer an
203 API issue but simply an extra trip through the brigade.
205 * Try to get libtool inter-library dependency code working on AIX.
206 Message-ID: <cm3n10lx555.fsf@rdu163-40-092.nc.rr.com>
208 Justin says: If we get it working on AIX, we can enable this
209 on all platforms and clean up our build system somewhat.
210 Jeff says: I thought I tested a patch for you sometime in
211 January that you were going to commit within a few days.
213 * Handling of %2f in URIs. Currently both 1.3 and 2.0
214 completely disallow %2f in the request URI path (see
215 ap_unescape_url() in util.c). It's permitted and passed
216 through in the query string, however. Roy says the
217 original reason for disallowing it, from five years ago,
218 was to protect CGI scripts that applied PATH_INFO to
219 a filesystem location and which might be tricked by
220 ..%2f..%2f(...). We *should* allow path-info of the
221 form 'http://foo.com/index.cgi/path/to/path%2finfo'.
222 Since we've revamped a lot of our processing of path
223 segments, it would be nice to allow this, or at least
224 allow it conditionally with a directive.
226 OtherBill adds that %2f as the SECOND character of a multibyte
227 sequence causes the request to fail! This happens notably in
230 * There is increasing demand from module writers for an API
231 that will allow them to control the server à la apachectl.
232 Reasons include sole-function servers that need to die if
233 an external dependency (e.g., a database) fails, et cetera.
234 Perhaps something in the (ever more abused) scoreboard?
236 On the other hand, we already have a pipe that goes between parent
237 and child for graceful shutdown events, along with an API that
238 can be used to send a message down that pipe. In threaded MPMs,
239 it is easy enough to make that one pipe be used for graceful
240 and graceless events, and it is also easy to open that pipe
241 to both parent and child for writing. Then we just need to
242 figure out how to do graceless on non-threaded MPMs.
244 * Allow the DocumentRoot directive within <Location > scopes? This
245 allows the beloved (crusty) Alias /foo/ /somepath/foo/ followed
246 by a <Directory /somepath/foo> to become simply
247 <Location /foo/> DocumentRoot /somefile/foo (IMHO a bit more legible
248 and in-your-face.) DocumentRoot unset would be accepted [and would
249 not permit content to be served, only virtual resources such as
250 server-info or server-status.
251 This proposed change would _not_ depricate Alias.
252 striker: See the thread starting with Message-ID:
253 JLEGKKNELMHCJPNMOKHOGEEJFBAA.striker@apache.org.
255 * Win32: Rotatelogs sometimes is not terminated when Apache
256 goes down hard. FirstBill was looking at possibly tracking the
257 child's-child processes in the parent process.
258 stoddard: Shared scoreboard might offer a good way for the parent
259 to keep track of 'other child' processes and whack them if the child
261 Other thoughts on walking the process chain using the NT kernel
262 have also been proposed on APR.
264 * Eliminate unnecessary creation of pipes in mod_cgid
266 * Combine log_child and piped_log_spawn. Clean up http_log.c.
269 * Platforms that do not support fork (primarily Win32 and AS/400)
270 Architect start-up code that avoids initializing all the modules
271 in the parent process on platforms that do not support fork.
273 * There are still a number of places in the code where we are
274 losing error status (i.e. throwing away the error returned by a
275 system call and replacing it with a generic error code)
277 * Mass vhosting version of suEXEC.
279 * All DBMs suffer from confusion in support/dbmmanage (perl script) since
280 the dbmmanage employs the first-matched dbm format. This is not
281 necessarily the library that Apache was built with. Aught to
282 rewrite dbmmanage upon installation to bin/ with the proper library
283 for predictable mod_auth_dbm administration.
284 Questions; htdbm exists, time to kill dbmmanage, or does it remain
285 useful as a perl dbm management example? If we keep it,
286 do we address the issue above?
289 Some additional items remaining:
290 - case_preserved_filename stuff
291 (use the new canonical name stuff?)
292 - find a new home for ap_text(_header)
293 - is it possible to remove the DAV: namespace stuff from util_xml?
295 * ap_core_translate() and its use by mod_mmap_static and mod_file_cache
296 are a bit wonky. The function should probably be exposed as a utility
297 function (such as ap_translate_url2fs() or ap_validate_fs_url() or
298 something). Another approach would be a new hook phase after
299 "translate" which would allow the module to munge what the
300 translation has decided to do.
301 Status: Greg +1 (volunteers)
303 * Explore use of a post-config hook for the code in http_main.c which
304 calls ap_fixup_virutal_hosts(), ap_fini_vhost_config(), and
305 ap_sort_hooks() [to reduce the logic in main()]
307 * read the config tree just once, and process N times (as necessary)
309 * (possibly) use UUIDs in mod_unique_id and/or mod_usertrack
311 * (possibly) port the bug fix for PR 6942 (segv when LoadModule is put
312 into a VirtualHost container) to 2.0.
314 * shift stuff to mod_core.h
316 * callers of ap_run_create_request() should check the return value
317 for failure (Doug volunteers)
319 * Fix the worker MPM to use POD to kill child processes instead
320 of ap_os_killpg, regardless of how they should die.
322 chrisd: Is this done, by any chance? See r92598 and r93358.
324 * Scoreboard structures could be changed in the future such that
325 proper alignment is not maintained, leading to segfaults on
326 some systems. Cliff posted a patch to deal with this issue but
327 later recanted. See this message to dev@apr.apache.org:
329 <Pine.LNX.4.44.0203011354090.16457-200000@deepthought.cs.virginia.edu>
331 * APXS either needs to be fixed completely for use when apr is out of tree,
332 or it should drop query mode altogether, and we just grow an
333 httpd-config or similar arrangement.
334 To quote a discussion in STATUS earlier:
336 thommay: this doesn't fix all the problems with apxs and out of
337 tree apr/apr-util, but it's a good start. There's still the
338 query cases; but I'm beginning to think that in these cases
339 the app should be querying ap{r,u}-config directly
340 deprecate -q: add htpd-config: gstein, pquerna, minfrin, pgollucci
343 TODO ISSUES REMAINING IN MOD_SSL:
345 * Do we need SSL_set_read_ahead()?
347 * the ssl_expr api is NOT THREAD SAFE. race conditions exist:
348 -in ssl_expr_comp() if SSLRequire is used in .htaccess
349 (ssl_expr_info is global)
350 -is ssl_expr_eval() if there is an error
351 (ssl_expr_error is global)
353 * SSLRequire directive (parsing of) leaks memory
355 * Diffie-Hellman-Parameters for temporary keys are hardcoded in
356 ssl_engine_dh.c, while the comment in ssl_engine_kernel.c says:
357 "it is suggested that keys be changed daily or every 500
358 transactions, and more often if possible."
360 * ssl_var_lookup could be rewritten to be MUCH faster
362 * CRL callback should be pluggable
364 * session cache store should be pluggable
366 * init functions should return status code rather than ssl_die()
368 * ssl_engine_pphrase.c needs to be reworked so it is generic enough
369 to also decrypt proxy keys
371 * output warning when allowing SSL v2.0 ? its so old
374 * mod_proxy: Ability to run SSL over proxy gateway connections,
375 encrypting (or reencrypting) at the proxy.
377 * mod_cache: Handle ESI tags.
379 * mod_cache: Resolve issue of how to cache page fragements (or perhaps
380 -if- we want to cache page fragements). Today, mod_cache/mod_mem_cache
381 will cache #include 'virtual' requests (but not #include 'file'
382 requests). This was accomplished by making CACHE_IN a
383 CONTENT_SET-1 filter to force it to run before the SUBREQ_CORE
384 filter. But now responses cannot be cached that include the
385 effects of having been run through CONTENT_SET filters
386 (mod_deflate, mod_expires, etc). We could rerun all the
387 CONTENT_SET filters on the cached response, but this will not
388 work in all cases. For example, mod_expires relies on installing
389 the EXPIRATION filter during fixups. Contents served out of
390 mod_cache (out of the quick_handler) bypass -all- the request
391 line server hooks (Ryan really hated this. It is great for
392 performance, but bad because of the complications listed above).
394 mod_cache/mod_mem_cache/mod_disk_cache:
396 * mod_mem_cache: Consider adding a RevalidateTimeout directive to
397 specify time at which local cached content is to be revalidated
398 (ie, underlying file stat'ed to see if it has changed).
400 * mod_cache: CacheEnable/CacheDisable should accept regular expressions.
401 jerenkrantz says: Too slow. Get regexs away from speedy caches by
402 default. Introduce a new CacheEnableRegex if you want.
404 * mod_mem_cache/mod_disk_cache: Need to be able to query cache
405 status (num of entries, cache object properties, etc.).
406 mod_status could be extended to query optional hooks defined
407 by modules for the purpose of reporting module status.
408 mod_cache (et. al.) could define optional hooks that are called
409 to collect status. Status should be queryable by
411 jerenkrantz says: Yawn. Who cares.
413 * MaxRequestsPerChild measures connections, not requests.
414 Until someone has a better way, we'll probably just rename it
415 "MaxConnectionsPerChild".
417 * Regex containers don't work in an intutive way
418 Status: No one has come up with an efficient way to fix this
419 behavior. Dean has suggested getting rid of regex containers
421 OtherBill suggests: We at least seem to agree on eliminating
422 the <Container ~ foo> forms, and using only
423 <ContainerMatch foo> semantics.
425 * orig_ct in the byterange/multipart handling may not be
426 needed. Apache 1.3 just never stashed "multipart" into
427 r->content_type. We should probably follow suit since the
428 byterange stuff doesn't want the rest of the code to see the
429 multipart content-type; the other code should still think it is
430 dealing with the <orig_ct> stuff.
431 Status: Greg volunteers to investigate (esp. since he was most
432 likely the one to break it :-)
434 EXPERIMENTAL MODULES:
436 Experimental modules should eventually be be promoted to fully supported
437 status or removed from the repository entirely (ie, the
438 'experiment' failed). This section tracks what needs to happen to
439 get the modules promoted to fully supported status.