1 APACHE 2.5 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 separately 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
17 Patches considered for backport are noted in their branches' STATUS:
19 * http://svn.apache.org/repos/asf/httpd/httpd/branches/2.0.x/STATUS
20 * http://svn.apache.org/repos/asf/httpd/httpd/branches/2.2.x/STATUS
21 * http://svn.apache.org/repos/asf/httpd/httpd/branches/2.4.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 2.5.1 : In development
30 2.5.0 : Tagged on November 8, 2017. 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.
46 * See also the STATUS file in the docs/ directory, which lists documentation-specific TODO items.
49 CURRENT RELEASE NOTES:
58 THINGS THAT SHOULD BE CONSIDERED EARLY IN THE 2.6/3.0 DEVELOPMENT CYCLE:
60 * Seriously ramp up/replace test framework and cases to have better
61 coverage of existing special cases and behaviours users rely on.
63 * Add performance testing to the test framework.
65 * Competely untangle core filesystem behavior where a filesystem htdocs/
66 resource wasn't indicated by the request URI.
68 * Refactor r->uri into a %escaped raw form presented by the client, and
69 a distinct decoded field used only for local filesystem access.
71 * Change default prefix from /usr/local/apache2 to something corresponding
72 to the project name. Rename apachectl to httpdctl.
74 * Change merge order of <Location> to be most specific match last. This
75 is more consistent with <Directory> and allows some optimizations for the
78 * Detect Lua 5.2.0 during configure and add LUA_COMPAT_ALL to CPPFLAGS.
79 Maybe it even suffices to add LUA_COMPAT_MODULE and individually
80 care about the two remaining incompatible code lines (one with lua_strlen,
83 * Event's timeout_mutex to enter keepalive state probably needs some
86 * Better H2 integration?
87 - adding handling of slave connections to mpm, no extra H2 workers,
88 triggering "events" read/write/timer from main/slave
89 - add slave writes/done/abort to events that wake up master connection
90 - disentangle core filters to server one purpose only, so that H2
91 versions can reuse them properly.
93 * Remove mod_access_compat?
95 * Ditch platforms/89/old prereqs or anything else?
97 * Leverage libmill? Drop serf?
99 * Better abstraction of slave connections and "requests".
100 - add abstraction for "response" as something that can be passed
101 through filters. To be serialized into the correct HTTP bytes on
103 - solve multi-threaded access to master connection props/module conf
106 * make mod_ssl more "core"?
108 * add high-level server configuration directives that can steer/influence
109 module defaults/warn/rejects related to security
111 * Ditch HTTP/0.9? At least, make HttpProtocolOptions Require1.0 the default.
113 * Restructure merge fn table/indexes to ignore modules with no directives,
114 and permit modules with dozens upon dozens of merge values to split these
115 into multiple functional config groups to avoid excessive merging.
116 Retitle from 'per-dir' to 'per-location' to better reflect the always-run
117 sections (location, ifexpr etc), while we phase out the file-oriented
120 * New versioning or release cadence.
122 * Ditch old APIs when we have the chance with 3.x. Consolidate current
123 functionality into APIs with stronger guarantees. (Specific examples TBD.)
125 * Remove as many undesirable-but-kept-for-backwards-compatibility behaviors
126 as possible from current config directives. (Specific examples TBD.)
128 * True event-loop/asynchronous support in the server core.
130 * Modify configuration syntax to separate meta-directives from runtime
131 directives (e.g. If vs. IfVersion). Allow as much static analysis of the
132 configuration as possible without needing to start the server to figure out
135 * Support JSON-like configuration files
137 * Opaque data structures w/ getters/setters
139 * Generic interface to enable runtime changes (adjusting log level, modifying
140 balancer information, toggling flags on/off). Perhaps modules can register
141 callbacks for making these changes?
143 * REST-based administration for existing (balancer/etc) and new dynamic
144 runtime changes (see above)
146 * Improve the look of generated pages (status, load-balancer...) with dynamic
147 update of the values. Generate HTML5 pages, instead of 3.2, Get rid of XHTML
148 in the generated pages.
150 * Add performance monitoring of the server, of each module (?), in order to help
151 understanding what worth looking at in order to improve overall performance.
152 (https://cdn.wp.nginx.com/wp-content/uploads/2016/12/Amplify-Dashboards-page-base-for-filters.png)
154 * Drop CGI-1.1-incompatible behaviors kept for compatibility reasons with
155 "broken" server implementations (PR 51517). (Note that many of them are
156 "broken" *because* of our behaviors.)
158 * Add a "normalized" list of headers for a HTTP response, rather then relying
159 on r->headers_out and r->err_headers_out, since mod_headers' behavior is
160 not really user friendly in some scenarios (example in PR 62380).
163 OLD ISSUES THAT WERE THOUGHT TO BE SHOWSTOPPERS FOR 2.4 BUT OBVIOUSLY WEREN'T:
165 * Handling of non-trailing / config by non-default handler is broken
166 http://marc.theaimsgroup.com/?l=apache-httpd-dev&m=105451701628081&w=2
167 jerenkrantz asks: Why should this block a release?
168 wsanchez agrees: this may be a change in behavior, but isn't
169 clearly wrong, and even if so, it doesn't seem like a
172 * the edge connection filter cannot be removed
173 http://marc.theaimsgroup.com/?l=apache-httpd-dev&m=105366252619530&w=2
174 http://mail-archives.apache.org/mod_mbox/httpd-dev/200501.mbox/%3C41E30B42.4060202@stason.org%3E
175 jerenkrantz asks: Why should this block a release?
176 stas replies: because it requires a rewrite of the filters stack
177 implementation (you have suggested that) and once 2.2 is
178 released you can't do that anymore.
179 pgollucci: this affects mod_perl I'm pretty sure.
182 RELEASE NON-SHOWSTOPPERS BUT WOULD BE REAL NICE TO WRAP THESE UP:
184 * Clean up all the kruft and *extremely* outdated stuff below...
186 * Maybe remove Limit/LimitExcept or at least make it log warnings when
189 * Patches submitted to the bug database:
190 http://issues.apache.org/bugzilla/buglist.cgi?bug_status=NEW&bug_status=ASSIGNED&bug_status=REOPENED&product=Apache+httpd-2&keywords=PatchAvailable
192 * Filter stacks and subrequests, redirects and fast redirects.
193 There's at least one PR that suffers from the current unclean behaviour
194 (which lets the server send garbage): PR 17629
195 nd says: Every subrequest should get its own filter stack with the
196 subreq_core filter as bottom-most. That filter does two things:
197 - swallow EOS buckets
198 - redirect the data stream to the upper request's (rr->main)
199 filter chain directly after the subrequest's starting
201 Once we have a clean solution, we can try to optimize
202 it, so that the server won't be slow down too much.
204 * RFC 2616 violations.
205 Closed PRs: 15852, 15857, 15859, 15861, 15864, 15869, 15870, 16120,
206 16125, 16135, 16136, 16137, 16138, 16139, 16140, 16518,
208 Open PRs: 15865, 15866, 15868, 16126, 16133, 16142, 16521, 42978
209 jerenkrantz says: need to decide how many we need to backport and/or
210 if these rise to showstopper status.
211 wrowe suggests: it would be nice to see "MUST" v.s. "SHOULD" v.s. "MAY"
212 out of this list, without reviewing them individually.
213 wrowe asks: what is lingering after 2.4.25 release? Offhand, only
216 * pipes deadlock on all platforms with limited pipe buffers (e.g. both
217 Linux and Win32, as opposed to only Win32 on 1.3). The right solution
218 is either GStein's proposal for a "CGI Brigade", or OtherBill's proposal
219 for "Poll Buckets" for "Polling Filter Chains". Or maybe both :-)
221 * All handlers should always send content down even if r->header_only
222 is set. If not, it means that the HEAD requests don't generate the
223 same headers as a GET which is wrong.
225 * exec cmd and suexec arg-passing enhancements
226 Status: Patches proposed
227 Message-ID: <20020526041748.A29148@prodigy.Redbrick.DCU.IE>
228 (see the "proc.patch" and "suexec-shell.patch" links in this message)
230 * The 2.0.36 worker MPM graceless shutdown changes work but are
231 a bit clunky on some platforms; eg, on Linux, the loop to
232 join each worker thread seems to hang, and the parent ends up
233 killing off the child with SIGKILL. But at least it shuts down.
235 chrisd: Has this been fixed by the changes for PR 38737?
237 * We do not properly substitute the prefix-variables in the configuration
238 scripts or generated-configs. (i.e. if sysconfdir is etc,
239 httpd-std.conf points to conf.)
241 * If any request gets through ap_process_request_internal() and is
242 scheduled to be served by the core handler, without a flag that this
243 r->filename was tested by dir/file_walk, we need to 500 at the very
244 end of the ap_process_request_internal() processing so sub_req-esters
245 know this request cannot be run. This provides authors of older
246 modules better compatibility, while still improving the security and
249 Status: still need to decide where this goes, OtherBill comments...
250 Message-ID: <065701c14526$495203b0$96c0b0d0@roweclan.net>
251 [Deleted comments regarding the ap_run_handler phase, as irrelevant
252 as BillS points out that "common case will be caught in
253 default_handler already (with the r->finfo.filetype == 0 check)"
254 and the issue is detecting this -before- we try to run the req.]
256 gregames says: can this happen somehow without a broken module
257 being involved? If not, why waste cycles trying to defend against
258 potential broken modules? It seems futile.
259 wrowe counters: no, it shouldn't happen unless the module is broken.
260 But the right answer is to fail the request up-front in dir/file
261 walk if the path was entirely invalid; and we can't do that either
262 UNTIL 2.1 or we break modules that haven't hooked map_to_storage.
264 * Can a static httpd be built reliably?
265 Message-ID: <20020207142751.T31582@clove.org>
267 * Usage of APR_BRIGADE_NORMALIZE in core_input_filter should be
270 <Pine.LNX.4.33.0201202232430.318-100000@deepthought.cs.virginia.edu>
271 Jeff wonders if we still care about this. It is no longer an
272 API issue but simply an extra trip through the brigade.
274 * Try to get libtool inter-library dependency code working on AIX.
275 Message-ID: <cm3n10lx555.fsf@rdu163-40-092.nc.rr.com>
277 Justin says: If we get it working on AIX, we can enable this
278 on all platforms and clean up our build system somewhat.
279 Jeff says: I thought I tested a patch for you sometime in
280 January that you were going to commit within a few days.
282 * Handling of %2f in URIs. Currently both 1.3 and 2.0
283 completely disallow %2f in the request URI path (see
284 ap_unescape_url() in util.c). It's permitted and passed
285 through in the query string, however. Roy says the
286 original reason for disallowing it, from five years ago,
287 was to protect CGI scripts that applied PATH_INFO to
288 a filesystem location and which might be tricked by
289 ..%2f..%2f(...). We *should* allow path-info of the
290 form 'http://foo.com/index.cgi/path/to/path%2finfo'.
291 Since we've revamped a lot of our processing of path
292 segments, it would be nice to allow this, or at least
293 allow it conditionally with a directive.
295 OtherBill adds that %2f as the SECOND character of a multibyte
296 sequence causes the request to fail! This happens notably in
299 * There is increasing demand from module writers for an API
300 that will allow them to control the server à la apachectl.
301 Reasons include sole-function servers that need to die if
302 an external dependency (e.g., a database) fails, et cetera.
303 Perhaps something in the (ever more abused) scoreboard?
305 On the other hand, we already have a pipe that goes between parent
306 and child for graceful shutdown events, along with an API that
307 can be used to send a message down that pipe. In threaded MPMs,
308 it is easy enough to make that one pipe be used for graceful
309 and graceless events, and it is also easy to open that pipe
310 to both parent and child for writing. Then we just need to
311 figure out how to do graceless on non-threaded MPMs.
313 * Allow the DocumentRoot directive within <Location > scopes? This
314 allows the beloved (crusty) Alias /foo/ /somepath/foo/ followed
315 by a <Directory /somepath/foo> to become simply
316 <Location /foo/> DocumentRoot /somefile/foo (IMHO a bit more legible
317 and in-your-face.) DocumentRoot unset would be accepted [and would
318 not permit content to be served, only virtual resources such as
319 server-info or server-status.
320 This proposed change would _not_ depricate Alias.
321 striker: See the thread starting with Message-ID:
322 JLEGKKNELMHCJPNMOKHOGEEJFBAA.striker@apache.org.
324 * Win32: Rotatelogs sometimes is not terminated when Apache
325 goes down hard. FirstBill was looking at possibly tracking the
326 child's-child processes in the parent process.
327 stoddard: Shared scoreboard might offer a good way for the parent
328 to keep track of 'other child' processes and whack them if the child
330 Other thoughts on walking the process chain using the NT kernel
331 have also been proposed on APR.
333 * Eliminate unnecessary creation of pipes in mod_cgid
335 * Combine log_child and piped_log_spawn. Clean up http_log.c.
338 * Platforms that do not support fork (primarily Win32 and AS/400)
339 Architect start-up code that avoids initializing all the modules
340 in the parent process on platforms that do not support fork.
342 * There are still a number of places in the code where we are
343 losing error status (i.e. throwing away the error returned by a
344 system call and replacing it with a generic error code)
346 * Mass vhosting version of suEXEC.
348 * All DBMs suffer from confusion in support/dbmmanage (perl script) since
349 the dbmmanage employs the first-matched dbm format. This is not
350 necessarily the library that Apache was built with. Aught to
351 rewrite dbmmanage upon installation to bin/ with the proper library
352 for predictable mod_auth_dbm administration.
353 Questions; htdbm exists, time to kill dbmmanage, or does it remain
354 useful as a perl dbm management example? If we keep it,
355 do we address the issue above?
358 Some additional items remaining:
359 - case_preserved_filename stuff
360 (use the new canonical name stuff?)
361 - find a new home for ap_text(_header)
362 - is it possible to remove the DAV: namespace stuff from util_xml?
364 * ap_core_translate() and its use by mod_mmap_static and mod_file_cache
365 are a bit wonky. The function should probably be exposed as a utility
366 function (such as ap_translate_url2fs() or ap_validate_fs_url() or
367 something). Another approach would be a new hook phase after
368 "translate" which would allow the module to munge what the
369 translation has decided to do.
370 Status: Greg +1 (volunteers)
372 * Explore use of a post-config hook for the code in http_main.c which
373 calls ap_fixup_virutal_hosts(), ap_fini_vhost_config(), and
374 ap_sort_hooks() [to reduce the logic in main()]
376 * read the config tree just once, and process N times (as necessary)
378 * (possibly) use UUIDs in mod_unique_id and/or mod_usertrack
380 * (possibly) port the bug fix for PR 6942 (segv when LoadModule is put
381 into a VirtualHost container) to 2.0.
383 * shift stuff to mod_core.h
385 * callers of ap_run_create_request() should check the return value
386 for failure (Doug volunteers)
388 * Fix the worker MPM to use POD to kill child processes instead
389 of ap_os_killpg, regardless of how they should die.
391 chrisd: Is this done, by any chance? See r92598 and r93358.
393 * Scoreboard structures could be changed in the future such that
394 proper alignment is not maintained, leading to segfaults on
395 some systems. Cliff posted a patch to deal with this issue but
396 later recanted. See this message to dev@apr.apache.org:
398 <Pine.LNX.4.44.0203011354090.16457-200000@deepthought.cs.virginia.edu>
400 * APXS either needs to be fixed completely for use when apr is out of tree,
401 or it should drop query mode altogether, and we just grow an
402 httpd-config or similar arrangement.
403 To quote a discussion in STATUS earlier:
405 thommay: this doesn't fix all the problems with apxs and out of
406 tree apr/apr-util, but it's a good start. There's still the
407 query cases; but I'm beginning to think that in these cases
408 the app should be querying ap{r,u}-config directly
409 deprecate -q: add htpd-config: gstein, pquerna, minfrin, pgollucci
412 TODO ISSUES REMAINING IN MOD_SSL:
414 * SSLRequire directive (parsing of) leaks memory
416 * ssl_var_lookup could be rewritten to be MUCH faster
419 * mod_proxy: Ability to run SSL over proxy gateway connections,
420 encrypting (or reencrypting) at the proxy.
422 * mod_cache: Handle ESI tags.
424 * mod_cache: Resolve issue of how to cache page fragments (or perhaps
425 -if- we want to cache page fragments). Today, mod_cache/mod_mem_cache
426 will cache #include 'virtual' requests (but not #include 'file'
427 requests). This was accomplished by making CACHE_IN a
428 CONTENT_SET-1 filter to force it to run before the SUBREQ_CORE
429 filter. But now responses cannot be cached that include the
430 effects of having been run through CONTENT_SET filters
431 (mod_deflate, mod_expires, etc). We could rerun all the
432 CONTENT_SET filters on the cached response, but this will not
433 work in all cases. For example, mod_expires relies on installing
434 the EXPIRATION filter during fixups. Contents served out of
435 mod_cache (out of the quick_handler) bypass -all- the request
436 line server hooks (Ryan really hated this. It is great for
437 performance, but bad because of the complications listed above).
439 mod_cache/mod_mem_cache/mod_cache_disk:
441 * mod_mem_cache: Consider adding a RevalidateTimeout directive to
442 specify time at which local cached content is to be revalidated
443 (ie, underlying file stat'ed to see if it has changed).
445 * mod_mem_cache/mod_cache_disk: Need to be able to query cache
446 status (num of entries, cache object properties, etc.).
447 mod_status could be extended to query optional hooks defined
448 by modules for the purpose of reporting module status.
449 mod_cache (et. al.) could define optional hooks that are called
450 to collect status. Status should be queryable by
452 jerenkrantz says: Yawn. Who cares.
454 * Regex containers don't work in an intutive way
455 Status: No one has come up with an efficient way to fix this
456 behavior. Dean has suggested getting rid of regex containers
458 OtherBill suggests: We at least seem to agree on eliminating
459 the <Container ~ foo> forms, and using only
460 <ContainerMatch foo> semantics.
462 * orig_ct in the byterange/multipart handling may not be
463 needed. Apache 1.3 just never stashed "multipart" into
464 r->content_type. We should probably follow suit since the
465 byterange stuff doesn't want the rest of the code to see the
466 multipart content-type; the other code should still think it is
467 dealing with the <orig_ct> stuff.
468 Status: Greg volunteers to investigate (esp. since he was most
469 likely the one to break it :-)
471 EXPERIMENTAL MODULES:
473 Experimental modules should eventually be be promoted to fully supported
474 status or removed from the repository entirely (ie, the
475 'experiment' failed). This section tracks what needs to happen to
476 get the modules promoted to fully supported status.