Skip to content
This repository


Subversion checkout URL

You can clone with HTTPS or Subversion.

Download ZIP
tag: 2.3.6
Fetching contributors…

Cannot retrieve contributors at this time

file 469 lines (357 sloc) 21.174 kb
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469
APACHE 2.3 STATUS: -*-text-*-
Last modified at [$Date$]

The current version of this file can be found at:


Documentation status is maintained seperately and can be found at:

  * docs/STATUS in this source tree, or

Consult the following STATUS files for information on related projects:


Patches considered for backport are noted in their branches' STATUS:


Release history:
    [NOTE that x.{odd}.z versions are strictly Alpha/Beta releases,
          while x.{even}.z versions are Stable/GA releases.]

    2.3.6 : Tagged June 11, 2010.
    2.3.5 : Released on January 26, 2010.
    2.3.4 : Released on December 8, 2009.
    2.3.3 : Tagged on November 11, 2009, not released.
    2.3.2 : Tagged on March 23, 2009, not released.
    2.3.1 : Tagged on January 2, 2009, not released.
    2.3.0 : Tagged on December 6, 2008, not released.

Contributors looking for a mission:

  * Just do an egrep on "TODO" or "XXX" in the source.

  * Review the bug database at:

  * Review the "PatchAvailable" bugs in the bug database:

    After testing, you can append a comment saying "Reviewed and tested".

  * Open bugs in the bug database.





  * Modules without documentation need to be moved to experimental or be

  * There is no working equivalent to 'Satisfy any' to authorize by
    user _or_ IP address:<>

  * Not all MPMs are updated to set conn_rec::current_thread correctly.
      (Prefork, Worker, Event, Simple are updated).
      jim sez: Then we just ship with those... mark any others as

  * Running the log_transaction hook from pool cleanup is fubar:

  * MPM event (maybe others, too) closes open files only after the
    connection has been closed. This could be fixed in apr-util or httpd:<>



  * Handling of non-trailing / config by non-default handler is broken
    jerenkrantz asks: Why should this block a release?
    wsanchez agrees: this may be a change in behavior, but isn't
      clearly wrong, and even if so, it doesn't seem like a

  * the edge connection filter cannot be removed

    jerenkrantz asks: Why should this block a release?

    stas replies: because it requires a rewrite of the filters stack
          implementation (you have suggested that) and once 2.2 is
          released you can't do that anymore.

    pgollucci: this affects mod_perl I'm pretty sure.


  * If the parent process dies, should the remaining child processes
    "gracefully" self-terminate. Or maybe we should make it a runtime
    option, or have a concept of 2 parent processes (one being a
    "hot spare").
    See: Message-ID: <3C58232C.FE91F19F@Golux.Com>

    Self-destruct: Ken, Martin, Lars, sctemme (parent shouldn't die, ever)
    Not self-destruct: BrianP, Ian, Cliff, BillS
    Make it runtime configurable: Aaron, jim, Justin, wrowe, rederpj, nd, pgollucci

    /* The below was a concept on *how* to handle the problem */
    Have 2 parents: +1: jim
                    -1: Justin, wrowe, rederpj, nd, pgollucci
                    +0: Lars, Martin (while standing by, could it do
                                      something useful?)

  * Make the worker MPM the default MPM for threaded Unix boxes.
    +1: Justin, Ian, Cliff, BillS, striker, wrowe, nd, pgollucci, sctemme
    +0: BrianP, Aaron (mutex contention is looking better with the
          latest code, let's continue tuning and testing), rederpj, jim
    -0: Lars

    pquerna: Do we want to change this for *2.4*?
    wrowe: Replies "yes"

  * Name the Server (version 2.4 or 3.0, depending on the final call)
    Recent discussion indicates we should designate a (short name).
    This is not yet a [Vote] - Your nominations please:
      * Apache HTTP Server (httpd)
        +1: sctemme (why mess with it?)


  * Patches submitted to the bug database:

  * Filter stacks and subrequests, redirects and fast redirects.
    There's at least one PR that suffers from the current unclean behaviour
    (which lets the server send garbage): PR 17629
    nd says: Every subrequest should get its own filter stack with the
             subreq_core filter as bottom-most. That filter does two things:
               - swallow EOS buckets
               - redirect the data stream to the upper request's (rr->main)
                 filter chain directly after the subrequest's starting
             Once we have a clean solution, we can try to optimize
             it, so that the server won't be slow down too much.

  * RFC 2616 violations.
    Closed PRs: 15852, 15857, 15859, 15861, 15864, 15869, 15870, 16120,
                16125, 16135, 16136, 16137, 16138, 16139, 16140, 16518,
    Open PRs: 15865, 15866, 15868, 16126, 16133, 16142, 16521
    jerenkrantz says: need to decide how many we need to backport and/or
                      if these rise to showstopper status.
    wrowe suggests: it would be nice to see "MUST" v.s. "SHOULD" v.s. "MAY"
                    out of this list, without reviewing them individually.

  * There is a bug in how we sort some hooks, at least the pre-config
    hook. The first time we call the hooks, they are in the correct
    order, but the second time, we don't sort them correctly. Currently,
    the modules/http/config.m4 file has been renamed to
    modules/http/config2.m4 to work around this problem, it should moved
    back when this is fixed.

      OtherBill offers that this is a SERIOUS problem. We do not sort
      correctly by the ordering arguments passed to the register hook
      functions. This was proven when I reordered the open_logs hook
      to attempt to open the error logs prior to the access logs. Possibly
      the entire sorting code needs to be refactored.

  * pipes deadlock on all platforms with limited pipe buffers (e.g. both
    Linux and Win32, as opposed to only Win32 on 1.3). The right solution
    is either GStein's proposal for a "CGI Brigade", or OtherBill's proposal
    for "Poll Buckets" for "Polling Filter Chains". Or maybe both :-)

  * All handlers should always send content down even if r->header_only
    is set. If not, it means that the HEAD requests don't generate the
    same headers as a GET which is wrong.

  * exec cmd and suexec arg-passing enhancements
    Status: Patches proposed
    Message-ID: <20020526041748.A29148@prodigy.Redbrick.DCU.IE>
    (see the "proc.patch" and "suexec-shell.patch" links in this message)

  * The 2.0.36 worker MPM graceless shutdown changes work but are
    a bit clunky on some platforms; eg, on Linux, the loop to
    join each worker thread seems to hang, and the parent ends up
    killing off the child with SIGKILL. But at least it shuts down.

    chrisd: Has this been fixed by the changes for PR 38737?

  * We do not properly substitute the prefix-variables in the configuration
    scripts or generated-configs. (i.e. if sysconfdir is etc,
    httpd-std.conf points to conf.)

  * If any request gets through ap_process_request_internal() and is
    scheduled to be served by the core handler, without a flag that this
    r->filename was tested by dir/file_walk, we need to 500 at the very
    end of the ap_process_request_internal() processing so sub_req-esters
    know this request cannot be run. This provides authors of older
    modules better compatibility, while still improving the security and
    robustness of 2.0.

    Status: still need to decide where this goes, OtherBill comments...
    Message-ID: <065701c14526$495203b0$>
    [Deleted comments regarding the ap_run_handler phase, as irrelevant
     as BillS points out that "common case will be caught in
     default_handler already (with the r->finfo.filetype == 0 check)"
     and the issue is detecting this -before- we try to run the req.]

     gregames says: can this happen somehow without a broken module
       being involved? If not, why waste cycles trying to defend against
       potential broken modules? It seems futile.
     wrowe counters: no, it shouldn't happen unless the module is broken.
       But the right answer is to fail the request up-front in dir/file
       walk if the path was entirely invalid; and we can't do that either
       UNTIL 2.1 or we break modules that haven't hooked map_to_storage.

  * Can a static httpd be built reliably?
      Message-ID: <>

  * Usage of APR_BRIGADE_NORMALIZE in core_input_filter should be
    removed if possible.
    Jeff wonders if we still care about this. It is no longer an
    API issue but simply an extra trip through the brigade.

  * Try to get libtool inter-library dependency code working on AIX.
      Message-ID: <>

    Justin says: If we get it working on AIX, we can enable this
                 on all platforms and clean up our build system somewhat.
    Jeff says: I thought I tested a patch for you sometime in
                 January that you were going to commit within a few days.

  * Handling of %2f in URIs. Currently both 1.3 and 2.0
    completely disallow %2f in the request URI path (see
    ap_unescape_url() in util.c). It's permitted and passed
    through in the query string, however. Roy says the
    original reason for disallowing it, from five years ago,
    was to protect CGI scripts that applied PATH_INFO to
    a filesystem location and which might be tricked by
    ..%2f..%2f(...). We *should* allow path-info of the
    form ''.
    Since we've revamped a lot of our processing of path
    segments, it would be nice to allow this, or at least
    allow it conditionally with a directive.

      OtherBill adds that %2f as the SECOND character of a multibyte
      sequence causes the request to fail! This happens notably in
      the ja-jis encoding.

  * There is increasing demand from module writers for an API
    that will allow them to control the server à la apachectl.
    Reasons include sole-function servers that need to die if
    an external dependency (e.g., a database) fails, et cetera.
    Perhaps something in the (ever more abused) scoreboard?
      On the other hand, we already have a pipe that goes between parent
      and child for graceful shutdown events, along with an API that
      can be used to send a message down that pipe. In threaded MPMs,
      it is easy enough to make that one pipe be used for graceful
      and graceless events, and it is also easy to open that pipe
      to both parent and child for writing. Then we just need to
      figure out how to do graceless on non-threaded MPMs.

  * Allow the DocumentRoot directive within <Location > scopes? This
    allows the beloved (crusty) Alias /foo/ /somepath/foo/ followed
    by a <Directory /somepath/foo> to become simply
    <Location /foo/> DocumentRoot /somefile/foo (IMHO a bit more legible
    and in-your-face.) DocumentRoot unset would be accepted [and would
    not permit content to be served, only virtual resources such as
    server-info or server-status.
    This proposed change would _not_ depricate Alias.
      striker: See the thread starting with Message-ID:

  * Win32: Rotatelogs sometimes is not terminated when Apache
    goes down hard. FirstBill was looking at possibly tracking the
    child's-child processes in the parent process.
      stoddard: Shared scoreboard might offer a good way for the parent
      to keep track of 'other child' processes and whack them if the child
      goes down.
      Other thoughts on walking the process chain using the NT kernel
      have also been proposed on APR.

  * Eliminate unnecessary creation of pipes in mod_cgid

  * Combine log_child and piped_log_spawn. Clean up http_log.c.
    Common logging API.

  * Platforms that do not support fork (primarily Win32 and AS/400)
    Architect start-up code that avoids initializing all the modules
    in the parent process on platforms that do not support fork.

  * There are still a number of places in the code where we are
    losing error status (i.e. throwing away the error returned by a
    system call and replacing it with a generic error code)

  * Mass vhosting version of suEXEC.

  * All DBMs suffer from confusion in support/dbmmanage (perl script) since
    the dbmmanage employs the first-matched dbm format. This is not
    necessarily the library that Apache was built with. Aught to
    rewrite dbmmanage upon installation to bin/ with the proper library
    for predictable mod_auth_dbm administration.
      Questions; htdbm exists, time to kill dbmmanage, or does it remain
                 useful as a perl dbm management example? If we keep it,
                 do we address the issue above?

  * Integrate mod_dav.
      Some additional items remaining:
      - case_preserved_filename stuff
          (use the new canonical name stuff?)
      - find a new home for ap_text(_header)
      - is it possible to remove the DAV: namespace stuff from util_xml?

  * ap_core_translate() and its use by mod_mmap_static and mod_file_cache
    are a bit wonky. The function should probably be exposed as a utility
    function (such as ap_translate_url2fs() or ap_validate_fs_url() or
    something). Another approach would be a new hook phase after
    "translate" which would allow the module to munge what the
    translation has decided to do.
      Status: Greg +1 (volunteers)

  * Explore use of a post-config hook for the code in http_main.c which
    calls ap_fixup_virutal_hosts(), ap_fini_vhost_config(), and
    ap_sort_hooks() [to reduce the logic in main()]

  * read the config tree just once, and process N times (as necessary)

  * (possibly) use UUIDs in mod_unique_id and/or mod_usertrack

  * (possibly) port the bug fix for PR 6942 (segv when LoadModule is put
    into a VirtualHost container) to 2.0.

  * shift stuff to mod_core.h

  * callers of ap_run_create_request() should check the return value
    for failure (Doug volunteers)

  * Fix the worker MPM to use POD to kill child processes instead
    of ap_os_killpg, regardless of how they should die.

      chrisd: Is this done, by any chance? See r92598 and r93358.

  * Scoreboard structures could be changed in the future such that
    proper alignment is not maintained, leading to segfaults on
    some systems. Cliff posted a patch to deal with this issue but
    later recanted. See this message to

  * APXS either needs to be fixed completely for use when apr is out of tree,
    or it should drop query mode altogether, and we just grow an
    httpd-config or similar arrangement.
    To quote a discussion in STATUS earlier:

        thommay: this doesn't fix all the problems with apxs and out of
                 tree apr/apr-util, but it's a good start. There's still the
                 query cases; but I'm beginning to think that in these cases
                 the app should be querying ap{r,u}-config directly
        deprecate -q: add htpd-config: gstein, pquerna, minfrin, pgollucci


  * Do we need SSL_set_read_ahead()?

  * the ssl_expr api is NOT THREAD SAFE. race conditions exist:
     -in ssl_expr_comp() if SSLRequire is used in .htaccess
      (ssl_expr_info is global)
     -is ssl_expr_eval() if there is an error
      (ssl_expr_error is global)

  * SSLRequire directive (parsing of) leaks memory

  * Diffie-Hellman-Parameters for temporary keys are hardcoded in
    ssl_engine_dh.c, while the comment in ssl_engine_kernel.c says:
    "it is suggested that keys be changed daily or every 500
    transactions, and more often if possible."

  * ssl_var_lookup could be rewritten to be MUCH faster

  * CRL callback should be pluggable

  * session cache store should be pluggable

  * init functions should return status code rather than ssl_die()

  * ssl_engine_pphrase.c needs to be reworked so it is generic enough
    to also decrypt proxy keys

  * output warning when allowing SSL v2.0 ? its so old

  * mod_proxy: Ability to run SSL over proxy gateway connections,
    encrypting (or reencrypting) at the proxy.

  * mod_cache: Handle ESI tags.

  * mod_cache: Resolve issue of how to cache page fragments (or perhaps
    -if- we want to cache page fragments). Today, mod_cache/mod_mem_cache
    will cache #include 'virtual' requests (but not #include 'file'
    requests). This was accomplished by making CACHE_IN a
    CONTENT_SET-1 filter to force it to run before the SUBREQ_CORE
    filter. But now responses cannot be cached that include the
    effects of having been run through CONTENT_SET filters
    (mod_deflate, mod_expires, etc). We could rerun all the
    CONTENT_SET filters on the cached response, but this will not
    work in all cases. For example, mod_expires relies on installing
    the EXPIRATION filter during fixups. Contents served out of
    mod_cache (out of the quick_handler) bypass -all- the request
    line server hooks (Ryan really hated this. It is great for
    performance, but bad because of the complications listed above).


  * mod_mem_cache: Consider adding a RevalidateTimeout directive to
    specify time at which local cached content is to be revalidated
    (ie, underlying file stat'ed to see if it has changed).

  * mod_cache: CacheEnable/CacheDisable should accept regular expressions.
    jerenkrantz says: Too slow. Get regexs away from speedy caches by
                      default. Introduce a new CacheEnableRegex if you want.

  * mod_mem_cache/mod_disk_cache: Need to be able to query cache
    status (num of entries, cache object properties, etc.).
    mod_status could be extended to query optional hooks defined
    by modules for the purpose of reporting module status.
    mod_cache (et. al.) could define optional hooks that are called
    to collect status. Status should be queryable by
    HTTP or SNMP?
    jerenkrantz says: Yawn. Who cares.

  * MaxRequestsPerChild measures connections, not requests.
      Until someone has a better way, we'll probably just rename it
  * Regex containers don't work in an intutive way
      Status: No one has come up with an efficient way to fix this
      behavior. Dean has suggested getting rid of regex containers
      OtherBill suggests: We at least seem to agree on eliminating
                          the <Container ~ foo> forms, and using only
                          <ContainerMatch foo> semantics.

  * orig_ct in the byterange/multipart handling may not be
    needed. Apache 1.3 just never stashed "multipart" into
    r->content_type. We should probably follow suit since the
    byterange stuff doesn't want the rest of the code to see the
    multipart content-type; the other code should still think it is
    dealing with the <orig_ct> stuff.
      Status: Greg volunteers to investigate (esp. since he was most
              likely the one to break it :-)


    Experimental modules should eventually be be promoted to fully supported
    status or removed from the repository entirely (ie, the
    'experiment' failed). This section tracks what needs to happen to
    get the modules promoted to fully supported status.
Something went wrong with that request. Please try again.