STATUS revision 9180a5933673ffb1af633c255ceee029340f3b1e
45e9809aff7304721fddb95654901b32195c9c7avboxsyncAPACHE 2.0 STATUS: -*-text-*-
45e9809aff7304721fddb95654901b32195c9c7avboxsyncLast modified at [$Date: 2001/07/09 02:51:07 $]
45e9809aff7304721fddb95654901b32195c9c7avboxsync 2.0.20 : rolled July 8, 2001
45e9809aff7304721fddb95654901b32195c9c7avboxsync 2.0.19 : rolled June 27, 2001
45e9809aff7304721fddb95654901b32195c9c7avboxsync 2.0.18 : rolled May 18, 2001
45e9809aff7304721fddb95654901b32195c9c7avboxsync 2.0.17 : rolled April 17, 2001
45e9809aff7304721fddb95654901b32195c9c7avboxsync 2.0.16 : rolled April 4, 2001
45e9809aff7304721fddb95654901b32195c9c7avboxsync 2.0.15 : rolled March 21, 2001
45e9809aff7304721fddb95654901b32195c9c7avboxsync 2.0.14 : rolled March 7, 2001
45e9809aff7304721fddb95654901b32195c9c7avboxsync 2.0a9 : released December 12, 2000
45e9809aff7304721fddb95654901b32195c9c7avboxsync 2.0a8 : released November 20, 2000
45e9809aff7304721fddb95654901b32195c9c7avboxsync 2.0a7 : released October 8, 2000
45e9809aff7304721fddb95654901b32195c9c7avboxsync 2.0a6 : released August 18, 2000
45e9809aff7304721fddb95654901b32195c9c7avboxsync 2.0a5 : released August 4, 2000
45e9809aff7304721fddb95654901b32195c9c7avboxsync 2.0a4 : released June 7, 2000
45e9809aff7304721fddb95654901b32195c9c7avboxsync 2.0a3 : released April 28, 2000
45e9809aff7304721fddb95654901b32195c9c7avboxsync 2.0a2 : released March 31, 2000
45e9809aff7304721fddb95654901b32195c9c7avboxsync 2.0a1 : released March 10, 2000
45e9809aff7304721fddb95654901b32195c9c7avboxsyncDAEDALUS 2.0 PROBLEMS:
45e9809aff7304721fddb95654901b32195c9c7avboxsync * mod_cgid and suexec have a problem co-existing. suexec sees a null
45e9809aff7304721fddb95654901b32195c9c7avboxsync command string sometimes.
45e9809aff7304721fddb95654901b32195c9c7avboxsync * core dump from 20010418 running 2_0_16
45e9809aff7304721fddb95654901b32195c9c7avboxsync #0 0x2813a3c8 in kill () from /usr/lib/libc.so.4
45e9809aff7304721fddb95654901b32195c9c7avboxsync #1 0x2817609e in abort () from /usr/lib/libc.so.4
45e9809aff7304721fddb95654901b32195c9c7avboxsync #2 0x8065299 in ap_log_assert (szExp=0x80aaa60 "total_bytes_left > 0 && tmplen > 0", szFile=0x80aa2aa "core.c", nLine=2555)
45e9809aff7304721fddb95654901b32195c9c7avboxsync #3 0x8075227 in sendfile_it_all (c=0x81470fc, fd=0x814759c, hdtr=0xbfbff670, file_offset=1929216, file_bytes_left=261949,
45e9809aff7304721fddb95654901b32195c9c7avboxsync total_bytes_left=261949, flags=0) at core.c:2555
45e9809aff7304721fddb95654901b32195c9c7avboxsync #4 0x80761e2 in core_output_filter (f=0x814737c, b=0x814764c) at core.c:3172
45e9809aff7304721fddb95654901b32195c9c7avboxsync #5 0x806d227 in ap_pass_brigade (next=0x814737c, bb=0x81e80fc) at util_filter.c:240
45e9809aff7304721fddb95654901b32195c9c7avboxsync #6 0x805e696 in check_pipeline_flush (r=0x820803c) at http_request.c:388
45e9809aff7304721fddb95654901b32195c9c7avboxsync #7 0x805e707 in ap_process_request (r=0x820803c) at http_request.c:432
45e9809aff7304721fddb95654901b32195c9c7avboxsync #8 0x805a1a9 in ap_process_http_connection (c=0x81470fc) at http_core.c:280
45e9809aff7304721fddb95654901b32195c9c7avboxsync #9 0x806bc60 in ap_run_process_connection (c=0x81470fc) at connection.c:82
45e9809aff7304721fddb95654901b32195c9c7avboxsync #10 0x806be84 in ap_process_connection (c=0x81470fc) at connection.c:216
45e9809aff7304721fddb95654901b32195c9c7avboxsync #11 0x805fbba in child_main (child_num_arg=272) at prefork.c:807
45e9809aff7304721fddb95654901b32195c9c7avboxsync #12 0x805fd20 in make_child (s=0x80c64fc, slot=272) at prefork.c:880
45e9809aff7304721fddb95654901b32195c9c7avboxsync #13 0x805ffec in perform_idle_server_maintenance () at prefork.c:1021
45e9809aff7304721fddb95654901b32195c9c7avboxsync #14 0x80603d1 in ap_mpm_run (_pconf=0x80c600c, plog=0x80f300c, s=0x80c64fc) at prefork.c:1191
45e9809aff7304721fddb95654901b32195c9c7avboxsync #15 0x80660cd in main (argc=1, argv=0xbfbffadc) at main.c:425
45e9809aff7304721fddb95654901b32195c9c7avboxsync #16 0x8059bf9 in _start ()
45e9809aff7304721fddb95654901b32195c9c7avboxsync The input data (received in one read from TCP layer):
45e9809aff7304721fddb95654901b32195c9c7avboxsync Via: 1.0 MDRPRXY01, 1.0 NS2
45e9809aff7304721fddb95654901b32195c9c7avboxsync Connection: Keep-Alive
45e9809aff7304721fddb95654901b32195c9c7avboxsync User-Agent: Mozilla/4.0 (compatible; MSIE 5.5; Windows NT 4.0)
45e9809aff7304721fddb95654901b32195c9c7avboxsync Accept: application/vnd.ms-excel, application/msword, application/vnd.ms-powerpoint, image/gif, image/x-xbitmap, image/jpeg,
45e9809aff7304721fddb95654901b32195c9c7avboxsync Accept-Language: en-us,tscii;q=0.5
45e9809aff7304721fddb95654901b32195c9c7avboxsync Referer: http://jakarta.apache.org/log4j/docs/download.html
45e9809aff7304721fddb95654901b32195c9c7avboxsync Accept-Encoding: gzip, deflate
45e9809aff7304721fddb95654901b32195c9c7avboxsync The confusion was because apr_sendfile() returned APR_SUCCESS
45e9809aff7304721fddb95654901b32195c9c7avboxsync but zero bytes sent. Presumably the FreeBSD kernel sendfile()
45e9809aff7304721fddb95654901b32195c9c7avboxsync did the same thing (not 100% sure).
45e9809aff7304721fddb95654901b32195c9c7avboxsync Also happened on 20010605...
45e9809aff7304721fddb95654901b32195c9c7avboxsync /usr/local/apache2b-vhost-trap/corefiles/httpd.core.12
45e9809aff7304721fddb95654901b32195c9c7avboxsync GET /builds/jakarta-turbine/release/2.1/tdk-2.1.zip HTTP/1.1
45e9809aff7304721fddb95654901b32195c9c7avboxsync Accept: image/gif, image/x-xbitmap, image/jpeg, image/pjpeg,
45e9809aff7304721fddb95654901b32195c9c7avboxsync application/vnd.ms-powerpoint, application/vnd.ms-excel,
45e9809aff7304721fddb95654901b32195c9c7avboxsync Referer: http://jakarta.apache.org/builds/jakarta-turbine/release/2.1/
45e9809aff7304721fddb95654901b32195c9c7avboxsync Accept-Language: en-gb
45e9809aff7304721fddb95654901b32195c9c7avboxsync Accept-Encoding: gzip, deflate
45e9809aff7304721fddb95654901b32195c9c7avboxsync User-Agent: Mozilla/4.0 (compatible; MSIE 5.01; Windows NT 5.0)
45e9809aff7304721fddb95654901b32195c9c7avboxsync Connection: Keep-Alive
45e9809aff7304721fddb95654901b32195c9c7avboxsync Again, it would seem that FreeBSD sendfile() returned rc 0 with
45e9809aff7304721fddb95654901b32195c9c7avboxsync no bytes sent. (Other eyes welcome, of course... make sure you
45e9809aff7304721fddb95654901b32195c9c7avboxsync look at 2_0_16 sources.)
45e9809aff7304721fddb95654901b32195c9c7avboxsync * core dump from 20010521 and 20010529 running 2_0_16 - the "3030" problem
45e9809aff7304721fddb95654901b32195c9c7avboxsync #0 0x80987e8 in apr_cvt (arg=1.3980432860952889e-76,
45e9809aff7304721fddb95654901b32195c9c7avboxsync ndigits=808464432, decpt=0x30303030,
45e9809aff7304721fddb95654901b32195c9c7avboxsync sign=0x30303030, eflag=808464432,
45e9809aff7304721fddb95654901b32195c9c7avboxsync buf=0x30303030 <Address 0x30303030 out of bounds>) at apr_snprintf.c:177
45e9809aff7304721fddb95654901b32195c9c7avboxsync #1 0x30303030 in ?? ()
45e9809aff7304721fddb95654901b32195c9c7avboxsync Cannot access memory at address 0x30303030.
45e9809aff7304721fddb95654901b32195c9c7avboxsync In both coredumps the request is /server-status?auto.
45e9809aff7304721fddb95654901b32195c9c7avboxsync It is unclear whether the apr_*printf function was passed bad
45e9809aff7304721fddb95654901b32195c9c7avboxsync data or it screwed up on its own. 0x30 is '0'. There is a
45e9809aff7304721fddb95654901b32195c9c7avboxsync string of 200-300 '0' characters in the dump, apparently
45e9809aff7304721fddb95654901b32195c9c7avboxsync overlaying enough of the stack to cause serious problems :)
45e9809aff7304721fddb95654901b32195c9c7avboxsyncRELEASE SHOWSTOPPERS:
45e9809aff7304721fddb95654901b32195c9c7avboxsync WARNING: ALWAYS check srclib/apr/STATUS and srclib/apr-util/STATUS
45e9809aff7304721fddb95654901b32195c9c7avboxsync * cgi on all platforms is broken. Specifically, with the addition of
45e9809aff7304721fddb95654901b32195c9c7avboxsync filters, we have lost the ability to flush partially written buffers
45e9809aff7304721fddb95654901b32195c9c7avboxsync received from CGI scripts to the network. Apache 1.3 does a
45e9809aff7304721fddb95654901b32195c9c7avboxsync non-blocking read on the pipe from the CGI. If bytes are present,
45e9809aff7304721fddb95654901b32195c9c7avboxsync Apache 1.3 keeps reading until one of three conditions are
45e9809aff7304721fddb95654901b32195c9c7avboxsync 1. output buffer fills up (Apache should flush to network)
45e9809aff7304721fddb95654901b32195c9c7avboxsync 2. receives an EOF on the pipe (ditto & cleanup)
45e9809aff7304721fddb95654901b32195c9c7avboxsync 3. receives EWOULDBLOCK on the non-blocking read
45e9809aff7304721fddb95654901b32195c9c7avboxsync In case 3, Apache should flush the network buffer then do a
45e9809aff7304721fddb95654901b32195c9c7avboxsync blocking read on the pipe. Apache 2.0 always does a blocking read
45e9809aff7304721fddb95654901b32195c9c7avboxsync (in the content length filter), thus never gets the chance to
45e9809aff7304721fddb95654901b32195c9c7avboxsync flush the network i/o buffer.
45e9809aff7304721fddb95654901b32195c9c7avboxsync * Rotatelogs on Windows sometimes is not terminated when Apache
45e9809aff7304721fddb95654901b32195c9c7avboxsync goes down hard. When this occurs, Apache will not serve requests
45e9809aff7304721fddb95654901b32195c9c7avboxsync after it is restarted. The problem is that the old rotatelogs
45e9809aff7304721fddb95654901b32195c9c7avboxsync is holding the listening socket. netstat -an shows two listeners
45e9809aff7304721fddb95654901b32195c9c7avboxsync on the webserver port.
45e9809aff7304721fddb95654901b32195c9c7avboxsync * There is a bug in how we sort some hooks, at least the pre-config
45e9809aff7304721fddb95654901b32195c9c7avboxsync hook. The first time we call the hooks, they are in the correct
45e9809aff7304721fddb95654901b32195c9c7avboxsync order, but the second time, we don't sort them correctly. Currently,
45e9809aff7304721fddb95654901b32195c9c7avboxsync the modules/http/config.m4 file has been renamed to
45e9809aff7304721fddb95654901b32195c9c7avboxsync modules/http/config2.m4 to work around this problem, it should moved
45e9809aff7304721fddb95654901b32195c9c7avboxsync back when this is fixed. rbb
45e9809aff7304721fddb95654901b32195c9c7avboxsync * Root all file systems with <Directory /> for WIN32/OS2/NW permissions
45e9809aff7304721fddb95654901b32195c9c7avboxsync Status: patch brought forward from 1.3.14
45e9809aff7304721fddb95654901b32195c9c7avboxsync WIN32 and OS2 need review [William Rowe, Brian Harvard]
45e9809aff7304721fddb95654901b32195c9c7avboxsync * Win32: Get Apache working on Windows 95/98. The following work
45e9809aff7304721fddb95654901b32195c9c7avboxsync (at least) needs to be done:
45e9809aff7304721fddb95654901b32195c9c7avboxsync - winnt MPM: Fix 95/98 code paths in the winnt MPM. There is some NT
45e9809aff7304721fddb95654901b32195c9c7avboxsync specific code that is still not in NT only code paths
45e9809aff7304721fddb95654901b32195c9c7avboxsync - IOL binds to APR sendfile, implemented with TransmitFile, which
45e9809aff7304721fddb95654901b32195c9c7avboxsync is not available on 95/98.
45e9809aff7304721fddb95654901b32195c9c7avboxsync - Document warning that OSR2 is required (for Crypt functions, in
45e9809aff7304721fddb95654901b32195c9c7avboxsync * Win32: Test access logging with multiple threads. Will the
45e9809aff7304721fddb95654901b32195c9c7avboxsync native file I/O calls serialize automagically like the
45e9809aff7304721fddb95654901b32195c9c7avboxsync CRT calls or do we need to add region locking each time
45e9809aff7304721fddb95654901b32195c9c7avboxsync we access the logs?
45e9809aff7304721fddb95654901b32195c9c7avboxsync * Win32: Complete the revamp the service environment and relocation
45e9809aff7304721fddb95654901b32195c9c7avboxsync into the WinNT MPM. Changes ServerRoot service registry
45e9809aff7304721fddb95654901b32195c9c7avboxsync parameter into ConfigArgs for multiple service startup parameters.
45e9809aff7304721fddb95654901b32195c9c7avboxsync Problems to fix in the revamp: -k shutdown/restart are broken,
45e9809aff7304721fddb95654901b32195c9c7avboxsync signals are not being acknowledged. Close window and shutdown
45e9809aff7304721fddb95654901b32195c9c7avboxsync also seem out of sorts.
45e9809aff7304721fddb95654901b32195c9c7avboxsync OtherBill is working on this and will finish over the weekend.
45e9809aff7304721fddb95654901b32195c9c7avboxsync * We need a thread-safe resolver, at least on Unix.
45e9809aff7304721fddb95654901b32195c9c7avboxsync Status: The best known candidate would be something from
45e9809aff7304721fddb95654901b32195c9c7avboxsync Status: Greg asks, "why? doesn't gethostbyname_r() handle this?"
45e9809aff7304721fddb95654901b32195c9c7avboxsync * The AddInputFilter and AddOutputFilter directives do not allow the
45e9809aff7304721fddb95654901b32195c9c7avboxsync administrator to remove or reorder filters. Once a filter is added
45e9809aff7304721fddb95654901b32195c9c7avboxsync in a container, it is present in any subcontainers. It can only be
45e9809aff7304721fddb95654901b32195c9c7avboxsync added to the filter chain after any filters specified in enclosing
45e9809aff7304721fddb95654901b32195c9c7avboxsync containers.
45e9809aff7304721fddb95654901b32195c9c7avboxsync * remove the --disable-shared from the subdir config of APR(UTIL)
45e9809aff7304721fddb95654901b32195c9c7avboxsync before the final release. (in fact, it might even be nice to
45e9809aff7304721fddb95654901b32195c9c7avboxsync allow for Apache config/build against an already-installed
45e9809aff7304721fddb95654901b32195c9c7avboxsync Note: we need to do a "make install" for APR(UTIL) so the shared
45e9809aff7304721fddb95654901b32195c9c7avboxsync libraries can be installed properly. We could also use that
45e9809aff7304721fddb95654901b32195c9c7avboxsync point to install include files (rather than have Apache
45e9809aff7304721fddb95654901b32195c9c7avboxsync know everything that needs to be installed from the
45e9809aff7304721fddb95654901b32195c9c7avboxsync sub-packages). The original impetus for doing the
45e9809aff7304721fddb95654901b32195c9c7avboxsync disable-shared was because the shared lib wasn't getting
45e9809aff7304721fddb95654901b32195c9c7avboxsync installed and a "make clean" in aprutil would make Apache
45e9809aff7304721fddb95654901b32195c9c7avboxsync fail to load.
45e9809aff7304721fddb95654901b32195c9c7avboxsync * ap_vrprintf() needs to handle more than 4K
45e9809aff7304721fddb95654901b32195c9c7avboxsync Status: Greg volunteers
45e9809aff7304721fddb95654901b32195c9c7avboxsync * mod_dir should normally redirect ALL directory requests which do
45e9809aff7304721fddb95654901b32195c9c7avboxsync not include a trailing slash on the URI. However, if a "notes"
45e9809aff7304721fddb95654901b32195c9c7avboxsync flag is set (say, via BrowserMatch), this behavior will be
45e9809aff7304721fddb95654901b32195c9c7avboxsync disabled for non-GET requests.
45e9809aff7304721fddb95654901b32195c9c7avboxsync Status: Greg volunteers
45e9809aff7304721fddb95654901b32195c9c7avboxsync MsgId: <20010227104646.E2297@lyra.org>
45e9809aff7304721fddb95654901b32195c9c7avboxsync MsgId: <3A9C0097.9C83F07C@Golux.Com>
45e9809aff7304721fddb95654901b32195c9c7avboxsync * Usability: Sanitize the MPM config directives. MaxClients in
45e9809aff7304721fddb95654901b32195c9c7avboxsync the threaded MPM is totally misleading now as it has little to
45e9809aff7304721fddb95654901b32195c9c7avboxsync do with limiting the number of clients (it limits the number
45e9809aff7304721fddb95654901b32195c9c7avboxsync of child processes). Bill proposed nomenclature change to
45e9809aff7304721fddb95654901b32195c9c7avboxsync something like "StartWorkers, MaxWorkers, etc." that could
45e9809aff7304721fddb95654901b32195c9c7avboxsync apply to most all the MPMs (with some notable exceptions).
45e9809aff7304721fddb95654901b32195c9c7avboxsync Bill would be happy with changing MaxClients to MaxServers
45e9809aff7304721fddb95654901b32195c9c7avboxsync to make it agree with the operation of the StartServers
45e9809aff7304721fddb95654901b32195c9c7avboxsyncRELEASE NON-SHOWSTOPPERS BUT WOULD BE REAL NICE TO WRAP THESE UP:
45e9809aff7304721fddb95654901b32195c9c7avboxsync * Port of mod_ssl to Apache 2.0:
45e9809aff7304721fddb95654901b32195c9c7avboxsync The current porting state is summarized in modules/ssl/README. The next
45e9809aff7304721fddb95654901b32195c9c7avboxsync step is to figure out how the old three configuration contexts (global,
45e9809aff7304721fddb95654901b32195c9c7avboxsync per-server, per-directory) can be ported to Apache 2.0 (especially the
45e9809aff7304721fddb95654901b32195c9c7avboxsync global context which has to survive server restarts RSE still does not
45e9809aff7304721fddb95654901b32195c9c7avboxsync know how to port). Then the remaining source files (which depend on the
45e9809aff7304721fddb95654901b32195c9c7avboxsync configuration contexts) which are still tagged with "-" in
45e9809aff7304721fddb95654901b32195c9c7avboxsync One more thing to figure out is how and when mod_ssl can provide the
45e9809aff7304721fddb95654901b32195c9c7avboxsync interactive pass phrase dialog (in Apache 1.3 it used the first init
45e9809aff7304721fddb95654901b32195c9c7avboxsync round [where tty was still not detached] and skipped the second init
45e9809aff7304721fddb95654901b32195c9c7avboxsync round [where it was already detached]). Apache 2.0 requires (or already
45e9809aff7304721fddb95654901b32195c9c7avboxsync has?) an official hook where such tty-dialogs can be performed.
45e9809aff7304721fddb95654901b32195c9c7avboxsync RSE is on holiday until June 11th, 2001. After this he starts
modules/ssl/. Do whatever you think is appropriate to get it
#if 0...endif wrapped to not make trouble for you.
malloc/free calls in the bucket brigade code. Need some
for this. First step is to s/malloc/apr_sms_malloc/g, etc.
"Apache" layout from config.layout, and each variable settable
* Combine log_child and piped_log_spawn. Clean up http_log.c.
* Win32: Migrate the MPM over to use APR thread/process calls. This
losing error status (i.e. throwing away the error returned by a
* All DBMs suffer from confusion in support/dbmmanage (perl script) since
for predictable mod_auth_db/dbm administration.
* Explore use of a post-config hook for the code in http_main.c which
* (possibly) use UUIDs in mod_unique_id and/or mod_usertrack
* shift stuff to mod_core.h
process (e.g., mutex init failure), other children (cgid, at least)
Using NIS passwd.byname dbm files with AuthDBMUserFile
setlogin() is not called, causing problems with e.g. identd
add allow,deny/deny,allow warning to mod_access
A small addition to rotatelogs.c to improve program functionality.
Feedback/Comment on APACI
Inclusion of RPM spec file in CVS/distributions
No way to change ReadmeName/HeaderName suffixes.
MIME types for MNG and JNG files need adding to mime.types and
the mime.types and magic files
* orig_ct in the byterange/multipart handling may not be
obsolete directives in core.html to the MPM documentation.
* Revise manual/stopping.html and the last part of
manual/misc/perf-tuning.html to take account of the MPMs.