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 2.3.1 : in development
30 2.3.0 : Tagged on December 6, 2008, not released.
32 Contributors looking for a mission:
34 * Just do an egrep on "TODO" or "XXX" in the source.
36 * Review the bug database at: http://issues.apache.org/bugzilla/
38 * Review the "PatchAvailable" bugs in the bug database:
40 https://issues.apache.org/bugzilla/buglist.cgi?bug_status=NEW&bug_status=ASSIGNED&bug_status=REOPENED&product=Apache+httpd-2&keywords=PatchAvailable
42 After testing, you can append a comment saying "Reviewed and tested".
44 * Open bugs in the bug database.
47 CURRENT RELEASE NOTES:
52 * Handling of non-trailing / config by non-default handler is broken
53 http://marc.theaimsgroup.com/?l=apache-httpd-dev&m=105451701628081&w=2
54 jerenkrantz asks: Why should this block a release?
55 wsanchez agrees: this may be a change in behavior, but isn't
56 clearly wrong, and even if so, it doesn't seem like a
59 * the edge connection filter cannot be removed
60 http://marc.theaimsgroup.com/?l=apache-httpd-dev&m=105366252619530&w=2
62 jerenkrantz asks: Why should this block a release?
64 stas replies: because it requires a rewrite of the filters stack
65 implementation (you have suggested that) and once 2.2 is
66 released you can't do that anymore.
68 pgollucci: this affects mod_perl I'm pretty sure.
72 * If the parent process dies, should the remaining child processes
73 "gracefully" self-terminate. Or maybe we should make it a runtime
74 option, or have a concept of 2 parent processes (one being a
76 See: Message-ID: <3C58232C.FE91F19F@Golux.Com>
78 Self-destruct: Ken, Martin, Lars
79 Not self-destruct: BrianP, Ian, Cliff, BillS
80 Make it runtime configurable: Aaron, jim, Justin, wrowe, rederpj, nd
82 /* The below was a concept on *how* to handle the problem */
83 Have 2 parents: +1: jim
84 -1: Justin, wrowe, rederpj, nd, pgollucci
85 +0: Lars, Martin (while standing by, could it do
88 * Make the worker MPM the default MPM for threaded Unix boxes.
89 +1: Justin, Ian, Cliff, BillS, striker, wrowe, nd, pgollucci
90 +0: BrianP, Aaron (mutex contention is looking better with the
91 latest code, let's continue tuning and testing), rederpj, jim
94 pquerna: Do we want to change this for *2.4*?
97 RELEASE NON-SHOWSTOPPERS BUT WOULD BE REAL NICE TO WRAP THESE UP:
99 * Patches submitted to the bug database:
100 http://issues.apache.org/bugzilla/buglist.cgi?bug_status=NEW&bug_status=ASSIGNED&bug_status=REOPENED&product=Apache+httpd-2&keywords=PatchAvailable
102 * Filter stacks and subrequests, redirects and fast redirects.
103 There's at least one PR that suffers from the current unclean behaviour
104 (which lets the server send garbage): PR 17629
105 nd says: Every subrequest should get its own filter stack with the
106 subreq_core filter as bottom-most. That filter does two things:
107 - swallow EOS buckets
108 - redirect the data stream to the upper request's (rr->main)
109 filter chain directly after the subrequest's starting
111 Once we have a clean solution, we can try to optimize
112 it, so that the server won't be slow down too much.
114 * RFC 2616 violations.
116 Open PRs: 15852, 15859, 15861, 15864, 15865, 15866, 15868, 15869,
117 15870, 16120, 16125, 16126, 16133, 16135, 16136, 16137,
118 16138, 16139, 16140, 16142, 16518, 16520, 16521,
119 jerenkrantz says: need to decide how many we need to backport and/or
120 if these rise to showstopper status.
121 wrowe suggests: it would be nice to see "MUST" v.s. "SHOULD" v.s. "MAY"
122 out of this list, without reviewing them individually.
124 * There is a bug in how we sort some hooks, at least the pre-config
125 hook. The first time we call the hooks, they are in the correct
126 order, but the second time, we don't sort them correctly. Currently,
127 the modules/http/config.m4 file has been renamed to
128 modules/http/config2.m4 to work around this problem, it should moved
129 back when this is fixed.
131 OtherBill offers that this is a SERIOUS problem. We do not sort
132 correctly by the ordering arguments passed to the register hook
133 functions. This was proven when I reordered the open_logs hook
134 to attempt to open the error logs prior to the access logs. Possibly
135 the entire sorting code needs to be refactored.
137 * pipes deadlock on all platforms with limited pipe buffers (e.g. both
138 Linux and Win32, as opposed to only Win32 on 1.3). The right solution
139 is either GStein's proposal for a "CGI Brigade", or OtherBill's proposal
140 for "Poll Buckets" for "Polling Filter Chains". Or maybe both :-)
142 * All handlers should always send content down even if r->header_only
143 is set. If not, it means that the HEAD requests don't generate the
144 same headers as a GET which is wrong.
146 * exec cmd and suexec arg-passing enhancements
147 Status: Patches proposed
148 Message-ID: <20020526041748.A29148@prodigy.Redbrick.DCU.IE>
149 (see the "proc.patch" and "suexec-shell.patch" links in this message)
151 * The 2.0.36 worker MPM graceless shutdown changes work but are
152 a bit clunky on some platforms; eg, on Linux, the loop to
153 join each worker thread seems to hang, and the parent ends up
154 killing off the child with SIGKILL. But at least it shuts down.
156 chrisd: Has this been fixed by the changes for PR 38737?
158 * --enable-mods-shared="foo1 foo2" is busted on Darwin. Pier
159 posted a patch (Message-ID: <B8DBBE8D.575A%pier@betaversion.org>).
161 * We do not properly substitute the prefix-variables in the configuration
162 scripts or generated-configs. (i.e. if sysconfdir is etc,
163 httpd-std.conf points to conf.)
165 * If any request gets through ap_process_request_internal() and is
166 scheduled to be served by the core handler, without a flag that this
167 r->filename was tested by dir/file_walk, we need to 500 at the very
168 end of the ap_process_request_internal() processing so sub_req-esters
169 know this request cannot be run. This provides authors of older
170 modules better compatibility, while still improving the security and
173 Status: still need to decide where this goes, OtherBill comments...
174 Message-ID: <065701c14526$495203b0$96c0b0d0@roweclan.net>
175 [Deleted comments regarding the ap_run_handler phase, as irrelevant
176 as BillS points out that "common case will be caught in
177 default_handler already (with the r->finfo.filetype == 0 check)"
178 and the issue is detecting this -before- we try to run the req.]
180 gregames says: can this happen somehow without a broken module
181 being involved? If not, why waste cycles trying to defend against
182 potential broken modules? It seems futile.
183 wrowe counters: no, it shouldn't happen unless the module is broken.
184 But the right answer is to fail the request up-front in dir/file
185 walk if the path was entirely invalid; and we can't do that either
186 UNTIL 2.1 or we break modules that haven't hooked map_to_storage.
188 * With AP_MODE_EXHAUSTIVE in the core, it is finally clear to me
189 how the Perchild MPM should be re-written. It hasn't worked
190 correctly since filters were added because it wasn't possible to
191 get the content that had already been written and the socket at
192 the same time. This mode lets us do that, so the MPM can be
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 * Get perchild to work on platforms other than Linux. This
206 will require a portable mechanism to pass data and file/socket
207 descriptors between vhost child groups. An API was proposed
209 Message-ID: <20020111115006.K1529@clove.org>
211 * Try to get libtool inter-library dependency code working on AIX.
212 Message-ID: <cm3n10lx555.fsf@rdu163-40-092.nc.rr.com>
214 Justin says: If we get it working on AIX, we can enable this
215 on all platforms and clean up our build system somewhat.
216 Jeff says: I thought I tested a patch for you sometime in
217 January that you were going to commit within a few days.
219 * Handling of %2f in URIs. Currently both 1.3 and 2.0
220 completely disallow %2f in the request URI path (see
221 ap_unescape_url() in util.c). It's permitted and passed
222 through in the query string, however. Roy says the
223 original reason for disallowing it, from five years ago,
224 was to protect CGI scripts that applied PATH_INFO to
225 a filesystem location and which might be tricked by
226 ..%2f..%2f(...). We *should* allow path-info of the
227 form 'http://foo.com/index.cgi/path/to/path%2finfo'.
228 Since we've revamped a lot of our processing of path
229 segments, it would be nice to allow this, or at least
230 allow it conditionally with a directive.
232 OtherBill adds that %2f as the SECOND character of a multibyte
233 sequence causes the request to fail! This happens notably in
236 * There is increasing demand from module writers for an API
237 that will allow them to control the server à la apachectl.
238 Reasons include sole-function servers that need to die if
239 an external dependency (e.g., a database) fails, et cetera.
240 Perhaps something in the (ever more abused) scoreboard?
242 On the other hand, we already have a pipe that goes between parent
243 and child for graceful shutdown events, along with an API that
244 can be used to send a message down that pipe. In threaded MPMs,
245 it is easy enough to make that one pipe be used for graceful
246 and graceless events, and it is also easy to open that pipe
247 to both parent and child for writing. Then we just need to
248 figure out how to do graceless on non-threaded MPMs.
250 * Allow the DocumentRoot directive within <Location > scopes? This
251 allows the beloved (crusty) Alias /foo/ /somepath/foo/ followed
252 by a <Directory /somepath/foo> to become simply
253 <Location /foo/> DocumentRoot /somefile/foo (IMHO a bit more legible
254 and in-your-face.) DocumentRoot unset would be accepted [and would
255 not permit content to be served, only virtual resources such as
256 server-info or server-status.
257 This proposed change would _not_ depricate Alias.
258 striker: See the thread starting with Message-ID:
259 JLEGKKNELMHCJPNMOKHOGEEJFBAA.striker@apache.org.
261 * Win32: Rotatelogs sometimes is not terminated when Apache
262 goes down hard. FirstBill was looking at possibly tracking the
263 child's-child processes in the parent process.
264 stoddard: Shared scoreboard might offer a good way for the parent
265 to keep track of 'other child' processes and whack them if the child
267 Other thoughts on walking the process chain using the NT kernel
268 have also been proposed on APR.
270 * Eliminate unnecessary creation of pipes in mod_cgid
272 * Combine log_child and piped_log_spawn. Clean up http_log.c.
275 * Platforms that do not support fork (primarily Win32 and AS/400)
276 Architect start-up code that avoids initializing all the modules
277 in the parent process on platforms that do not support fork.
279 * There are still a number of places in the code where we are
280 losing error status (i.e. throwing away the error returned by a
281 system call and replacing it with a generic error code)
283 * Mass vhosting version of suEXEC.
285 * All DBMs suffer from confusion in support/dbmmanage (perl script) since
286 the dbmmanage employs the first-matched dbm format. This is not
287 necessarily the library that Apache was built with. Aught to
288 rewrite dbmmanage upon installation to bin/ with the proper library
289 for predictable mod_auth_dbm administration.
290 Questions; htdbm exists, time to kill dbmmanage, or does it remain
291 useful as a perl dbm management example? If we keep it,
292 do we address the issue above?
295 Some additional items remaining:
296 - case_preserved_filename stuff
297 (use the new canonical name stuff?)
298 - find a new home for ap_text(_header)
299 - is it possible to remove the DAV: namespace stuff from util_xml?
301 * ap_core_translate() and its use by mod_mmap_static and mod_file_cache
302 are a bit wonky. The function should probably be exposed as a utility
303 function (such as ap_translate_url2fs() or ap_validate_fs_url() or
304 something). Another approach would be a new hook phase after
305 "translate" which would allow the module to munge what the
306 translation has decided to do.
307 Status: Greg +1 (volunteers)
309 * Explore use of a post-config hook for the code in http_main.c which
310 calls ap_fixup_virutal_hosts(), ap_fini_vhost_config(), and
311 ap_sort_hooks() [to reduce the logic in main()]
313 * read the config tree just once, and process N times (as necessary)
315 * (possibly) use UUIDs in mod_unique_id and/or mod_usertrack
317 * (possibly) port the bug fix for PR 6942 (segv when LoadModule is put
318 into a VirtualHost container) to 2.0.
320 * shift stuff to mod_core.h
322 * callers of ap_run_create_request() should check the return value
323 for failure (Doug volunteers)
325 * Fix the worker MPM to use POD to kill child processes instead
326 of ap_os_killpg, regardless of how they should die.
328 chrisd: Is this done, by any chance? See r92598 and r93358.
330 * Scoreboard structures could be changed in the future such that
331 proper alignment is not maintained, leading to segfaults on
332 some systems. Cliff posted a patch to deal with this issue but
333 later recanted. See this message to dev@apr.apache.org:
335 <Pine.LNX.4.44.0203011354090.16457-200000@deepthought.cs.virginia.edu>
337 * APXS either needs to be fixed completely for use when apr is out of tree,
338 or it should drop query mode altogether, and we just grow an
339 httpd-config or similar arrangement.
340 To quote a discussion in STATUS earlier:
342 thommay: this doesn't fix all the problems with apxs and out of
343 tree apr/apr-util, but it's a good start. There's still the
344 query cases; but I'm beginning to think that in these cases
345 the app should be querying ap{r,u}-config directly
346 deprecate -q: add htpd-config: gstein, pquerna, minfrin, pgollucci
349 TODO ISSUES REMAINING IN MOD_SSL:
351 * Do we need SSL_set_read_ahead()?
353 * the ssl_expr api is NOT THREAD SAFE. race conditions exist:
354 -in ssl_expr_comp() if SSLRequire is used in .htaccess
355 (ssl_expr_info is global)
356 -is ssl_expr_eval() if there is an error
357 (ssl_expr_error is global)
359 * SSLRequire directive (parsing of) leaks memory
361 * Diffie-Hellman-Parameters for temporary keys are hardcoded in
362 ssl_engine_dh.c, while the comment in ssl_engine_kernel.c says:
363 "it is suggested that keys be changed daily or every 500
364 transactions, and more often if possible."
366 * ssl_var_lookup could be rewritten to be MUCH faster
368 * CRL callback should be pluggable
370 * session cache store should be pluggable
372 * init functions should return status code rather than ssl_die()
374 * ssl_engine_pphrase.c needs to be reworked so it is generic enough
375 to also decrypt proxy keys
377 * output warning when allowing SSL v2.0 ? its so old
380 * mod_proxy: Ability to run SSL over proxy gateway connections,
381 encrypting (or reencrypting) at the proxy.
383 * mod_cache: Handle ESI tags.
385 * mod_cache: Resolve issue of how to cache page fragements (or perhaps
386 -if- we want to cache page fragements). Today, mod_cache/mod_mem_cache
387 will cache #include 'virtual' requests (but not #include 'file'
388 requests). This was accomplished by making CACHE_IN a
389 CONTENT_SET-1 filter to force it to run before the SUBREQ_CORE
390 filter. But now responses cannot be cached that include the
391 effects of having been run through CONTENT_SET filters
392 (mod_deflate, mod_expires, etc). We could rerun all the
393 CONTENT_SET filters on the cached response, but this will not
394 work in all cases. For example, mod_expires relies on installing
395 the EXPIRATION filter during fixups. Contents served out of
396 mod_cache (out of the quick_handler) bypass -all- the request
397 line server hooks (Ryan really hated this. It is great for
398 performance, but bad because of the complications listed above).
400 mod_cache/mod_mem_cache/mod_disk_cache:
402 * mod_mem_cache: Consider adding a RevalidateTimeout directive to
403 specify time at which local cached content is to be revalidated
404 (ie, underlying file stat'ed to see if it has changed).
406 * mod_cache: CacheEnable/CacheDisable should accept regular expressions.
407 jerenkrantz says: Too slow. Get regexs away from speedy caches by
408 default. Introduce a new CacheEnableRegex if you want.
410 * mod_mem_cache/mod_disk_cache: Need to be able to query cache
411 status (num of entries, cache object properties, etc.).
412 mod_status could be extended to query optional hooks defined
413 by modules for the purpose of reporting module status.
414 mod_cache (et. al.) could define optional hooks that are called
415 to collect status. Status should be queryable by
417 jerenkrantz says: Yawn. Who cares.
419 * MaxRequestsPerChild measures connections, not requests.
420 Until someone has a better way, we'll probably just rename it
421 "MaxConnectionsPerChild".
423 * Regex containers don't work in an intutive way
424 Status: No one has come up with an efficient way to fix this
425 behavior. Dean has suggested getting rid of regex containers
427 OtherBill suggests: We at least seem to agree on eliminating
428 the <Container ~ foo> forms, and using only
429 <ContainerMatch foo> semantics.
431 * orig_ct in the byterange/multipart handling may not be
432 needed. Apache 1.3 just never stashed "multipart" into
433 r->content_type. We should probably follow suit since the
434 byterange stuff doesn't want the rest of the code to see the
435 multipart content-type; the other code should still think it is
436 dealing with the <orig_ct> stuff.
437 Status: Greg volunteers to investigate (esp. since he was most
438 likely the one to break it :-)
440 EXPERIMENTAL MODULES:
442 Experimental modules should eventually be be promoted to fully supported
443 status or removed from the repository entirely (ie, the
444 'experiment' failed). This section tracks what needs to happen to
445 get the modules promoted to fully supported status.