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