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 2.3.2 : in development
30 2.3.1 : Tagged on January 2, 2009.
31 2.3.0 : Tagged on December 6, 2008, not released.
33 Contributors looking for a mission:
35 * Just do an egrep on "TODO" or "XXX" in the source.
37 * Review the bug database at: http://issues.apache.org/bugzilla/
39 * Review the "PatchAvailable" bugs in the bug database:
41 https://issues.apache.org/bugzilla/buglist.cgi?bug_status=NEW&bug_status=ASSIGNED&bug_status=REOPENED&product=Apache+httpd-2&keywords=PatchAvailable
43 After testing, you can append a comment saying "Reviewed and tested".
45 * Open bugs in the bug database.
48 CURRENT RELEASE NOTES:
53 * Handling of non-trailing / config by non-default handler is broken
54 http://marc.theaimsgroup.com/?l=apache-httpd-dev&m=105451701628081&w=2
55 jerenkrantz asks: Why should this block a release?
56 wsanchez agrees: this may be a change in behavior, but isn't
57 clearly wrong, and even if so, it doesn't seem like a
60 * the edge connection filter cannot be removed
61 http://marc.theaimsgroup.com/?l=apache-httpd-dev&m=105366252619530&w=2
63 jerenkrantz asks: Why should this block a release?
65 stas replies: because it requires a rewrite of the filters stack
66 implementation (you have suggested that) and once 2.2 is
67 released you can't do that anymore.
69 pgollucci: this affects mod_perl I'm pretty sure.
73 * If the parent process dies, should the remaining child processes
74 "gracefully" self-terminate. Or maybe we should make it a runtime
75 option, or have a concept of 2 parent processes (one being a
77 See: Message-ID: <3C58232C.FE91F19F@Golux.Com>
79 Self-destruct: Ken, Martin, Lars
80 Not self-destruct: BrianP, Ian, Cliff, BillS
81 Make it runtime configurable: Aaron, jim, Justin, wrowe, rederpj, nd
83 /* The below was a concept on *how* to handle the problem */
84 Have 2 parents: +1: jim
85 -1: Justin, wrowe, rederpj, nd, pgollucci
86 +0: Lars, Martin (while standing by, could it do
89 * Make the worker MPM the default MPM for threaded Unix boxes.
90 +1: Justin, Ian, Cliff, BillS, striker, wrowe, nd, pgollucci
91 +0: BrianP, Aaron (mutex contention is looking better with the
92 latest code, let's continue tuning and testing), rederpj, jim
95 pquerna: Do we want to change this for *2.4*?
98 RELEASE NON-SHOWSTOPPERS BUT WOULD BE REAL NICE TO WRAP THESE UP:
100 * Patches submitted to the bug database:
101 http://issues.apache.org/bugzilla/buglist.cgi?bug_status=NEW&bug_status=ASSIGNED&bug_status=REOPENED&product=Apache+httpd-2&keywords=PatchAvailable
103 * Filter stacks and subrequests, redirects and fast redirects.
104 There's at least one PR that suffers from the current unclean behaviour
105 (which lets the server send garbage): PR 17629
106 nd says: Every subrequest should get its own filter stack with the
107 subreq_core filter as bottom-most. That filter does two things:
108 - swallow EOS buckets
109 - redirect the data stream to the upper request's (rr->main)
110 filter chain directly after the subrequest's starting
112 Once we have a clean solution, we can try to optimize
113 it, so that the server won't be slow down too much.
115 * RFC 2616 violations.
117 Open PRs: 15852, 15859, 15861, 15864, 15865, 15866, 15868, 15869,
118 15870, 16120, 16125, 16126, 16133, 16135, 16136, 16137,
119 16138, 16139, 16140, 16142, 16518, 16520, 16521,
120 jerenkrantz says: need to decide how many we need to backport and/or
121 if these rise to showstopper status.
122 wrowe suggests: it would be nice to see "MUST" v.s. "SHOULD" v.s. "MAY"
123 out of this list, without reviewing them individually.
125 * There is a bug in how we sort some hooks, at least the pre-config
126 hook. The first time we call the hooks, they are in the correct
127 order, but the second time, we don't sort them correctly. Currently,
128 the modules/http/config.m4 file has been renamed to
129 modules/http/config2.m4 to work around this problem, it should moved
130 back when this is fixed.
132 OtherBill offers that this is a SERIOUS problem. We do not sort
133 correctly by the ordering arguments passed to the register hook
134 functions. This was proven when I reordered the open_logs hook
135 to attempt to open the error logs prior to the access logs. Possibly
136 the entire sorting code needs to be refactored.
138 * pipes deadlock on all platforms with limited pipe buffers (e.g. both
139 Linux and Win32, as opposed to only Win32 on 1.3). The right solution
140 is either GStein's proposal for a "CGI Brigade", or OtherBill's proposal
141 for "Poll Buckets" for "Polling Filter Chains". Or maybe both :-)
143 * All handlers should always send content down even if r->header_only
144 is set. If not, it means that the HEAD requests don't generate the
145 same headers as a GET which is wrong.
147 * exec cmd and suexec arg-passing enhancements
148 Status: Patches proposed
149 Message-ID: <20020526041748.A29148@prodigy.Redbrick.DCU.IE>
150 (see the "proc.patch" and "suexec-shell.patch" links in this message)
152 * The 2.0.36 worker MPM graceless shutdown changes work but are
153 a bit clunky on some platforms; eg, on Linux, the loop to
154 join each worker thread seems to hang, and the parent ends up
155 killing off the child with SIGKILL. But at least it shuts down.
157 chrisd: Has this been fixed by the changes for PR 38737?
159 * --enable-mods-shared="foo1 foo2" is busted on Darwin. Pier
160 posted a patch (Message-ID: <B8DBBE8D.575A%pier@betaversion.org>).
162 * We do not properly substitute the prefix-variables in the configuration
163 scripts or generated-configs. (i.e. if sysconfdir is etc,
164 httpd-std.conf points to conf.)
166 * If any request gets through ap_process_request_internal() and is
167 scheduled to be served by the core handler, without a flag that this
168 r->filename was tested by dir/file_walk, we need to 500 at the very
169 end of the ap_process_request_internal() processing so sub_req-esters
170 know this request cannot be run. This provides authors of older
171 modules better compatibility, while still improving the security and
174 Status: still need to decide where this goes, OtherBill comments...
175 Message-ID: <065701c14526$495203b0$96c0b0d0@roweclan.net>
176 [Deleted comments regarding the ap_run_handler phase, as irrelevant
177 as BillS points out that "common case will be caught in
178 default_handler already (with the r->finfo.filetype == 0 check)"
179 and the issue is detecting this -before- we try to run the req.]
181 gregames says: can this happen somehow without a broken module
182 being involved? If not, why waste cycles trying to defend against
183 potential broken modules? It seems futile.
184 wrowe counters: no, it shouldn't happen unless the module is broken.
185 But the right answer is to fail the request up-front in dir/file
186 walk if the path was entirely invalid; and we can't do that either
187 UNTIL 2.1 or we break modules that haven't hooked map_to_storage.
189 * With AP_MODE_EXHAUSTIVE in the core, it is finally clear to me
190 how the Perchild MPM should be re-written. It hasn't worked
191 correctly since filters were added because it wasn't possible to
192 get the content that had already been written and the socket at
193 the same time. This mode lets us do that, so the MPM can be
196 * Can a static httpd be built reliably?
197 Message-ID: <20020207142751.T31582@clove.org>
199 * Usage of APR_BRIGADE_NORMALIZE in core_input_filter should be
202 <Pine.LNX.4.33.0201202232430.318-100000@deepthought.cs.virginia.edu>
203 Jeff wonders if we still care about this. It is no longer an
204 API issue but simply an extra trip through the brigade.
206 * Get perchild to work on platforms other than Linux. This
207 will require a portable mechanism to pass data and file/socket
208 descriptors between vhost child groups. An API was proposed
210 Message-ID: <20020111115006.K1529@clove.org>
212 * Try to get libtool inter-library dependency code working on AIX.
213 Message-ID: <cm3n10lx555.fsf@rdu163-40-092.nc.rr.com>
215 Justin says: If we get it working on AIX, we can enable this
216 on all platforms and clean up our build system somewhat.
217 Jeff says: I thought I tested a patch for you sometime in
218 January that you were going to commit within a few days.
220 * Handling of %2f in URIs. Currently both 1.3 and 2.0
221 completely disallow %2f in the request URI path (see
222 ap_unescape_url() in util.c). It's permitted and passed
223 through in the query string, however. Roy says the
224 original reason for disallowing it, from five years ago,
225 was to protect CGI scripts that applied PATH_INFO to
226 a filesystem location and which might be tricked by
227 ..%2f..%2f(...). We *should* allow path-info of the
228 form 'http://foo.com/index.cgi/path/to/path%2finfo'.
229 Since we've revamped a lot of our processing of path
230 segments, it would be nice to allow this, or at least
231 allow it conditionally with a directive.
233 OtherBill adds that %2f as the SECOND character of a multibyte
234 sequence causes the request to fail! This happens notably in
237 * There is increasing demand from module writers for an API
238 that will allow them to control the server à la apachectl.
239 Reasons include sole-function servers that need to die if
240 an external dependency (e.g., a database) fails, et cetera.
241 Perhaps something in the (ever more abused) scoreboard?
243 On the other hand, we already have a pipe that goes between parent
244 and child for graceful shutdown events, along with an API that
245 can be used to send a message down that pipe. In threaded MPMs,
246 it is easy enough to make that one pipe be used for graceful
247 and graceless events, and it is also easy to open that pipe
248 to both parent and child for writing. Then we just need to
249 figure out how to do graceless on non-threaded MPMs.
251 * Allow the DocumentRoot directive within <Location > scopes? This
252 allows the beloved (crusty) Alias /foo/ /somepath/foo/ followed
253 by a <Directory /somepath/foo> to become simply
254 <Location /foo/> DocumentRoot /somefile/foo (IMHO a bit more legible
255 and in-your-face.) DocumentRoot unset would be accepted [and would
256 not permit content to be served, only virtual resources such as
257 server-info or server-status.
258 This proposed change would _not_ depricate Alias.
259 striker: See the thread starting with Message-ID:
260 JLEGKKNELMHCJPNMOKHOGEEJFBAA.striker@apache.org.
262 * Win32: Rotatelogs sometimes is not terminated when Apache
263 goes down hard. FirstBill was looking at possibly tracking the
264 child's-child processes in the parent process.
265 stoddard: Shared scoreboard might offer a good way for the parent
266 to keep track of 'other child' processes and whack them if the child
268 Other thoughts on walking the process chain using the NT kernel
269 have also been proposed on APR.
271 * Eliminate unnecessary creation of pipes in mod_cgid
273 * Combine log_child and piped_log_spawn. Clean up http_log.c.
276 * Platforms that do not support fork (primarily Win32 and AS/400)
277 Architect start-up code that avoids initializing all the modules
278 in the parent process on platforms that do not support fork.
280 * There are still a number of places in the code where we are
281 losing error status (i.e. throwing away the error returned by a
282 system call and replacing it with a generic error code)
284 * Mass vhosting version of suEXEC.
286 * All DBMs suffer from confusion in support/dbmmanage (perl script) since
287 the dbmmanage employs the first-matched dbm format. This is not
288 necessarily the library that Apache was built with. Aught to
289 rewrite dbmmanage upon installation to bin/ with the proper library
290 for predictable mod_auth_dbm administration.
291 Questions; htdbm exists, time to kill dbmmanage, or does it remain
292 useful as a perl dbm management example? If we keep it,
293 do we address the issue above?
296 Some additional items remaining:
297 - case_preserved_filename stuff
298 (use the new canonical name stuff?)
299 - find a new home for ap_text(_header)
300 - is it possible to remove the DAV: namespace stuff from util_xml?
302 * ap_core_translate() and its use by mod_mmap_static and mod_file_cache
303 are a bit wonky. The function should probably be exposed as a utility
304 function (such as ap_translate_url2fs() or ap_validate_fs_url() or
305 something). Another approach would be a new hook phase after
306 "translate" which would allow the module to munge what the
307 translation has decided to do.
308 Status: Greg +1 (volunteers)
310 * Explore use of a post-config hook for the code in http_main.c which
311 calls ap_fixup_virutal_hosts(), ap_fini_vhost_config(), and
312 ap_sort_hooks() [to reduce the logic in main()]
314 * read the config tree just once, and process N times (as necessary)
316 * (possibly) use UUIDs in mod_unique_id and/or mod_usertrack
318 * (possibly) port the bug fix for PR 6942 (segv when LoadModule is put
319 into a VirtualHost container) to 2.0.
321 * shift stuff to mod_core.h
323 * callers of ap_run_create_request() should check the return value
324 for failure (Doug volunteers)
326 * Fix the worker MPM to use POD to kill child processes instead
327 of ap_os_killpg, regardless of how they should die.
329 chrisd: Is this done, by any chance? See r92598 and r93358.
331 * Scoreboard structures could be changed in the future such that
332 proper alignment is not maintained, leading to segfaults on
333 some systems. Cliff posted a patch to deal with this issue but
334 later recanted. See this message to dev@apr.apache.org:
336 <Pine.LNX.4.44.0203011354090.16457-200000@deepthought.cs.virginia.edu>
338 * APXS either needs to be fixed completely for use when apr is out of tree,
339 or it should drop query mode altogether, and we just grow an
340 httpd-config or similar arrangement.
341 To quote a discussion in STATUS earlier:
343 thommay: this doesn't fix all the problems with apxs and out of
344 tree apr/apr-util, but it's a good start. There's still the
345 query cases; but I'm beginning to think that in these cases
346 the app should be querying ap{r,u}-config directly
347 deprecate -q: add htpd-config: gstein, pquerna, minfrin, pgollucci
350 TODO ISSUES REMAINING IN MOD_SSL:
352 * Do we need SSL_set_read_ahead()?
354 * the ssl_expr api is NOT THREAD SAFE. race conditions exist:
355 -in ssl_expr_comp() if SSLRequire is used in .htaccess
356 (ssl_expr_info is global)
357 -is ssl_expr_eval() if there is an error
358 (ssl_expr_error is global)
360 * SSLRequire directive (parsing of) leaks memory
362 * Diffie-Hellman-Parameters for temporary keys are hardcoded in
363 ssl_engine_dh.c, while the comment in ssl_engine_kernel.c says:
364 "it is suggested that keys be changed daily or every 500
365 transactions, and more often if possible."
367 * ssl_var_lookup could be rewritten to be MUCH faster
369 * CRL callback should be pluggable
371 * session cache store should be pluggable
373 * init functions should return status code rather than ssl_die()
375 * ssl_engine_pphrase.c needs to be reworked so it is generic enough
376 to also decrypt proxy keys
378 * output warning when allowing SSL v2.0 ? its so old
381 * mod_proxy: Ability to run SSL over proxy gateway connections,
382 encrypting (or reencrypting) at the proxy.
384 * mod_cache: Handle ESI tags.
386 * mod_cache: Resolve issue of how to cache page fragements (or perhaps
387 -if- we want to cache page fragements). Today, mod_cache/mod_mem_cache
388 will cache #include 'virtual' requests (but not #include 'file'
389 requests). This was accomplished by making CACHE_IN a
390 CONTENT_SET-1 filter to force it to run before the SUBREQ_CORE
391 filter. But now responses cannot be cached that include the
392 effects of having been run through CONTENT_SET filters
393 (mod_deflate, mod_expires, etc). We could rerun all the
394 CONTENT_SET filters on the cached response, but this will not
395 work in all cases. For example, mod_expires relies on installing
396 the EXPIRATION filter during fixups. Contents served out of
397 mod_cache (out of the quick_handler) bypass -all- the request
398 line server hooks (Ryan really hated this. It is great for
399 performance, but bad because of the complications listed above).
401 mod_cache/mod_mem_cache/mod_disk_cache:
403 * mod_mem_cache: Consider adding a RevalidateTimeout directive to
404 specify time at which local cached content is to be revalidated
405 (ie, underlying file stat'ed to see if it has changed).
407 * mod_cache: CacheEnable/CacheDisable should accept regular expressions.
408 jerenkrantz says: Too slow. Get regexs away from speedy caches by
409 default. Introduce a new CacheEnableRegex if you want.
411 * mod_mem_cache/mod_disk_cache: Need to be able to query cache
412 status (num of entries, cache object properties, etc.).
413 mod_status could be extended to query optional hooks defined
414 by modules for the purpose of reporting module status.
415 mod_cache (et. al.) could define optional hooks that are called
416 to collect status. Status should be queryable by
418 jerenkrantz says: Yawn. Who cares.
420 * MaxRequestsPerChild measures connections, not requests.
421 Until someone has a better way, we'll probably just rename it
422 "MaxConnectionsPerChild".
424 * Regex containers don't work in an intutive way
425 Status: No one has come up with an efficient way to fix this
426 behavior. Dean has suggested getting rid of regex containers
428 OtherBill suggests: We at least seem to agree on eliminating
429 the <Container ~ foo> forms, and using only
430 <ContainerMatch foo> semantics.
432 * orig_ct in the byterange/multipart handling may not be
433 needed. Apache 1.3 just never stashed "multipart" into
434 r->content_type. We should probably follow suit since the
435 byterange stuff doesn't want the rest of the code to see the
436 multipart content-type; the other code should still think it is
437 dealing with the <orig_ct> stuff.
438 Status: Greg volunteers to investigate (esp. since he was most
439 likely the one to break it :-)
441 EXPERIMENTAL MODULES:
443 Experimental modules should eventually be be promoted to fully supported
444 status or removed from the repository entirely (ie, the
445 'experiment' failed). This section tracks what needs to happen to
446 get the modules promoted to fully supported status.