0N/AAPACHE 2.0 STATUS: -*-text-*-
189N/ALast modified at [$Date: 2002/03/22 15:22:24 $]
0N/A 2.0.34 : in development
180N/A 2.0.33 : tagged March 6, 2002.
0N/A 2.0.32 : released Feburary 16, 2002.
180N/A 2.0.31 : rolled Feburary 1, 2002. not released.
0N/A 2.0.30 : tagged January 8, 2002. not rolled.
0N/A 2.0.29 : tagged November 27, 2001. not rolled.
0N/A 2.0.28 : released November 13, 2001
0N/A 2.0.27 : rolled November 6, 2001
0N/A 2.0.26 : tagged October 16, 2001. not rolled.
0N/A 2.0.25 : rolled August 29, 2001
0N/A 2.0.24 : rolled August 18, 2001
0N/A 2.0.23 : rolled August 9, 2001
0N/A 2.0.22 : rolled July 29, 2001
0N/A 2.0.21 : rolled July 20, 2001
0N/A 2.0.20 : rolled July 8, 2001
180N/A 2.0.19 : rolled June 27, 2001
180N/A 2.0.18 : rolled May 18, 2001
180N/A 2.0.17 : rolled April 17, 2001
0N/A 2.0.16 : rolled April 4, 2001
0N/A 2.0.15 : rolled March 21, 2001
5N/A 2.0.14 : rolled March 7, 2001
5N/A 2.0a9 : released December 12, 2000
0N/A 2.0a8 : released November 20, 2000
8N/A 2.0a7 : released October 8, 2000
0N/A 2.0a6 : released August 18, 2000
0N/A 2.0a5 : released August 4, 2000
0N/A 2.0a4 : released June 7, 2000
0N/A 2.0a3 : released April 28, 2000
0N/A 2.0a2 : released March 31, 2000
0N/A 2.0a1 : released March 10, 2000
0N/APlease consult the following STATUS files for information
27N/ACURRENT RELEASE NOTES:
27N/A * 34 status: Let's get all API changes and showstoppers in this one.
0N/AFINAL RELEASE SHOWSTOPPERS:
0N/A * If any request gets to the core handler, without a flag that this
0N/A end of the ap_process_request_internal() processing. This provides
0N/A authors of older modules better compatibility, while still improving
0N/A the security and robustness of 2.0.
0N/A Status: still need to decide where this goes, OtherBill comments...
0N/A Message-ID: <065701c14526$495203b0$96c0b0d0@roweclan.net>
0N/A we need to look at halting this in the 'default handler' case,
27N/A and that implies pushing the 'handler election' into the request
27N/A internal processing phase from the run request phase.
27N/A Jim asks: would a stopgap be something bogus like adding another
27N/A flag to request_rec ala eos_sent and before we OK, if not set
27N/A Jeff says: reviewing the original message and the one
27N/A follow-up (also from OtherBill) it looks like OtherBill had a
27N/A good handle on the problem, though I wonder why not just put a
27N/A be in request_rec; a better place is core_request_config)
27N/A gregames says: can this happen somehow without a broken module
27N/A being involved? If not, why waste cycles trying to defend against
32N/A potential broken modules? It seems futile. Please vote.
32N/A not a showstopper: gregames
27N/A * API changes planned for 2.0 that should happen before the
27N/A * Free lists for bucket allocation
27N/A * We do not properly substitute the prefix-variables in the configuration
27N/A scripts or generated-configs. (
i.e. if sysconfdir is etc,
27N/A Aaron says: This is not a showstopper, these problems have existed
27N/A for as long as I can remember. It would be nice to fix
27N/A them but they are not new.
27N/A Not a showstopper: Jeff, Aaron, BillS
27N/A * Should we always build binaries statically unless otherwise
27N/A Message-ID: <20020129210006.B23512@Lithium.MeepZor.Com>
27N/A * If the parent process dies, should the remaining child processes
27N/A "gracefully" self-terminate. Or maybe we should make it a runtime
27N/A option, or have a concept of 2 parent processes (one being a
27N/A See: Message-ID: <3C58232C.FE91F19F@Golux.Com>
27N/A Self-destruct: Ken, Martin
189N/A Not self-destruct: BrianP, Ian, Cliff, BillS
27N/A Make it runtime configurable: Aaron, Jim, Justin
27N/A Have 2 parents: +1: Jim
27N/A +0: Martin (while standing by, could it do
27N/A * Make the worker MPM the default MPM for threaded Unix boxes.
27N/A +1: Justin, Ian, Cliff
27N/A -0: Aaron (premature decision, needs more discussion), Lars
27N/ARELEASE NON-SHOWSTOPPERS BUT WOULD BE REAL NICE TO WRAP THESE UP:
27N/A * Rewrite core_output_filter. It is nearly impossible to support
27N/A it with predictable results as it is implemented now.
27N/A * Convert all instances of the old apr_lock_t type to the new
27N/A Status: Aaron has converted all but the perchild MPM to
27N/A use the new lock API. Since perchild has been put
27N/A on the back burner, this is no longer a showstopper.
27N/A Aaron will patch perchild as soon as it becomes
27N/A * With AP_MODE_EXHAUSTIVE in the core, it is finally clear to me
27N/A how the Perchild MPM should be re-written. It hasn't worked
27N/A correctly since filters were added because it wasn't possible to
27N/A get the content that had already been written and the socket at
27N/A the same time. This mode lets us do that, so the MPM can be
27N/A * htpasswd blindly processes the file you give it, and does no
27N/A sanity checking before totally corrupting whatever file it was
27N/A you thought you had. It should check the input file and bail
27N/A if it finds non-comment lines that do not contain exactly 1
27N/A Message-ID: <20020217150457.A31632@clove.org>
27N/A * Can a static httpd be built reliably?
27N/A Message-ID: <20020207142751.T31582@clove.org>
27N/A * [Ken] Test suite failures:
27N/A o worker is also failing some of the 'cgi' subtests
27N/A Justin says: "Worker should be fine and passes httpd-test here.
27N/A If you can provide evidence that it can be reproduced
27N/A outside of httpd-test, then it's a showstopper. I
27N/A think it's a perl or a httpd-test problem."
27N/A Not a showstopper: Justin
27N/A * Usage of APR_BRIGADE_NORMALIZE in core_input_filter should be
27N/A removed if possible.
27N/A Message-ID: <Pine.LNX.4.33.0201202232430.318-100000@deepthought.cs.virginia.edu>
189N/A * There is a bug in how we sort some hooks, at least the pre-config
189N/A hook. The first time we call the hooks, they are in the correct
189N/A order, but the second time, we don't sort them correctly. Currently,
27N/A back when this is fixed. rbb
27N/A Justin says: "Is this really a showstopper? This has been here
189N/A forever. What's wrong? Does this have to do with
27N/A Not a showstopper: Justin, BrianP, trawick, gregames
142N/A administrator to order filters, beyond the order of filename (mime)
27N/A filename extensions. At minimum, some sort of +-[0-10] syntax seems
27N/A like the quickest fix for a 2.0 gold release.
27N/A Justin says: "Could we delay this for a point release or 2.1?"
27N/A Not a showstopper: justin, wrowe, trawick, stoddard, Jim, Ian, Aaron,
27N/A * Get perchild to work on platforms other than Linux. This
27N/A descriptors between vhost child groups. An API was proposed
27N/A Message-ID: <20020111115006.K1529@clove.org>
27N/A * CGI single-byte reads
142N/A BrianP suggests that this is caused by the ap_scan_script_header_err()
27N/A routine, which will do single-byte reads until it finds the end
27N/A of the header, at which point it constructs a pipe-bucket (buffered)
27N/A Proposed solution in:
27N/A Message-ID: <3C36ADAF.60601@cnet.com>
27N/A * Try to get libtool inter-library dependency code working on AIX.
27N/A Message-ID: <cm3n10lx555.fsf@rdu163-40-092.nc.rr.com>
27N/A Justin says: If we get it working on AIX, we can enable this
27N/A on all platforms and clean up our build system
27N/A Jeff says: I thought I tested a patch for you sometime in
27N/A January that you were going to commit within a few
27N/A * Handling of %2f in URIs. Currently both 1.3 and 2.0
27N/A completely disallow %2f in the request URI path (see
27N/A ap_unescape_url() in
util.c). It's permitted and passed
27N/A through in the query string, however. Roy says the
27N/A original reason for disallowing it, from five years ago,
27N/A was to protect CGI scripts that applied PATH_INFO to
27N/A a filesystem location and which might be tricked by
27N/A ..%2f..%2f(...). We *should* allow path-info of the
27N/A Since we've revamped a lot of our processing of path
27N/A segments, it would be nice to allow this, or at least
27N/A allow it conditionally with a directive.
27N/A * FreeBSD, threads, and worker MPM. All seems to work fine
27N/A if you only have one worker process with many threads. Add
27N/A a second worker process and the accept lock seems to be
27N/A lost. This might be an APR issue with how it deals with
27N/A the child_init hook (
i.e. the fcntl lock needs to be resynced).
27N/A More examination and analysis is required.
27N/A Status: This has also been reported on Cygwin.
27N/A Message-ID: <3C2CC514.8EF3BED1@wapme-systems.de> (cygnus)
27N/A Justin says: So, FreeBSD-CURRENT and Cywin have the same
27N/A problem. Yum. If another platform has this
27N/A with worker, this becomes a showstopper.
27N/A Aaron says: I spent some time disecting this and have come to
27N/A the conclusion that it is not a problem in the worker MPM
189N/A (or at least, it is not isolated to a problem in worker).
189N/A I'll list some of the problems I'm seeing in case someone
189N/A else wants to pick up where I've left off:
189N/A - Delivery of just about any signal to one of the child
27N/A processes will send it into an infinite loop as well.
27N/A - Even though the parent is spinning out of control,
27N/A at first the child or children will appear to work
189N/A properly. At times it is possible to get it into a state,
27N/A however, where a request will hang until another concurrent
142N/A request "kicks" the first, at which point the second will
27N/A hang. My theory is that this has to do with the
27N/A pthread_cond_*() implementation in FreeBSD, but it's still
27N/A possible that it is in APR.
27N/A Justin adds: Oh, FreeBSD threads are implemented entirely with
27N/A select()/poll()/longjmp(). Welcome to the nightmare.
27N/A So, that means a ktrace output also has the thread
27N/A scheduling internals in it (since it is all the same to
27N/A the kernel). Which makes it hard to distinguish between
27N/A our select() calls and their select() calls.
27N/A *bangs head on wall repeatedly* But, some of the libc_r
27N/A files have a DBG_MSG #define. This is moderately helpful
27N/A when used with -DNO_DETACH. The kernel scheduler isn't
27N/A waking up the threads on a select(). Yum. And, I bet
0N/A those decrementing select calls have to do with the
0N/A scheduler. Time to brush up on our OS fundamentals.
189N/A * There is increasing demand from module writers for an API
189N/A that will allow them to control the server � la apachectl.
0N/A Reasons include sole-function servers that need to die if
0N/A an external dependency (
e.g., a database) fails, et cetera.
0N/A Perhaps something in the (ever more abused) scoreboard?
0N/A rbb: I don't believe the scoreboard is the correct mechanism
27N/A for this. We already have a pipe that goes between parent
27N/A and child for graceful shutdown events, along with an API that
27N/A can be used to send a message down that pipe. In threaded MPMs,
27N/A it is easy enough to make that one pipe be used for graceful
27N/A and graceless events, and it is also easy to open that pipe
27N/A to both parent and child for writing. Then we just need to
27N/A figure out how to do graceless on non-threaded MPMs.
27N/A * Allow the DocumentRoot directive within <Location > scopes? This
0N/A and in-your-face.) DocumentRoot unset would be accepted [and would
0N/A not permit content to be served, only virtual resources such as
0N/A server-info or server-status.
0N/A This proposed change would _not_ depricate Alias.
0N/A * Win32: Rotatelogs sometimes is not terminated when Apache
0N/A goes down hard. FirstBill was looking at possibly tracking the
0N/A child's-child processes in the parent process.
0N/A OtherBill asks, wasn't this fixed?
0N/A stoddard: Not fixed. Shared scoreboard might offer a good
0N/A way for the parent to keep track of 'other child' processes
0N/A and whack them if the child goes down.
0N/A * Win32: Add a simple hold console open patch (wait for close or
0N/A the ESC key, with a nice message) if the server died a bad
0N/A death (non-zero exit code) in console mode.
0N/A Resolution: bring forward same ugly hacks from 1.3.13-.20
0N/A * Port of mod_ssl to Apache 2.0:
0N/A remaining work includes:
0N/A (2) Enabling SSL extentions
0N/A (3) Trying to seperate the https filter logic from mod_ssl -
0N/A This is to facilitate other modules that wish to use the https
0N/A filter or the mod_ssl logic or both as required.
0N/A * Performance & Debug: Eliminate most (and perhaps all) of the
0N/A light weight memory management functions that allow freeing
0N/A memory (putting it back into a memory pool) when it is no
0N/A longer needed. Enabling simple debugging features like guard
0N/A bands, double free detection, etc. would be cool but certainly
0N/A not a hard requirement.
0N/A Status: Cliff started to implement this using SMS as has
0N/A been discussed at length for months, but since
0N/A SMS is not being used anywhere else in the server,
0N/A several people expressed the opinion that we should
0N/A get rid of it entirely, meaning that the buckets
0N/A need their own memory management (free list) functions.
0N/A Cliff will implement that this weekend so we at least
0N/A * Eliminate unnecessary creation of pipes in mod_cgid
0N/A * Document mod_file_cache.
0N/A * Platforms that do not support fork (primarily Win32 and AS/400)
0N/A Architect start-up code that avoids initializing all the modules
0N/A in the parent process on platforms that do not support fork.
0N/A would eliminate some code in the Win32 branch that essentially
0N/A duplicates what is in APR.
0N/A * There are still a number of places in the code where we are
0N/A losing error status (
i.e. throwing away the error returned by a
0N/A system call and replacing it with a generic error code)
0N/A * Mass vhosting version of suEXEC.
0N/A the dbmmanage employs the first-matched dbm format. This is not
0N/A necessarily the library that Apache was built with. Aught to
0N/A rewrite dbmmanage upon installation to bin/ with the proper library
0N/A for predictable mod_auth_dbm administration.
0N/A Questions; htdbm exists, time to kill dbmmanage, or does it remain
0N/A useful as a perl dbm management example? If we keep it,
0N/A do we address the issue above?
0N/A * Integrate mod_dav.
0N/A Some additional items remaining:
0N/A - case_preserved_filename stuff
0N/A (use the new canonical name stuff?)
0N/A - find a new home for ap_text(_header)
0N/A - is it possible to remove the DAV: namespace stuff from util_xml?
0N/A * ap_core_translate() and its use by mod_mmap_static and mod_file_cache
0N/A are a bit wonky. The function should probably be exposed as a utility
0N/A function (such as ap_translate_url2fs() or ap_validate_fs_url() or
0N/A something). Another approach would be a new hook phase after
0N/A "translate" which would allow the module to munge what the
0N/A translation has decided to do.
0N/A Status: Greg +1 (volunteers), Ryan +1
0N/A calls ap_fixup_virutal_hosts(), ap_fini_vhost_config(), and
0N/A ap_sort_hooks() [to reduce the logic in main()]
0N/A * read the config tree just once, and process N times (as necessary)
0N/A * (possibly) use UUIDs in mod_unique_id
and/or mod_usertrack
0N/A * (possibly) port the bug fix for PR 6942 (segv when LoadModule is put
0N/A into a VirtualHost container) to 2.0.
0N/A * callers of ap_run_create_request() should check the return value
0N/A for failure (Doug volunteers)
0N/A * Win32: Get Apache working on Windows 95/98. The following work
0N/A (at least) needs to be done:
0N/A - Document warning that OSR2 is required (for Crypt functions, in
0N/A rand.c, at least.) This could be resolved with an SSL library, or
0N/A randomization in APR itself.
0N/A actually works) and add in a splash of Win9x service code.
289N/A * In order to use a DSO version of mod_ssl we have to link with
0N/A -lssl and -lcrypto. A workaround is in place right now where the
0N/A entire EXTRA_LIBS macro is being appended to the objects list, but
289N/A this is a hack. We should either revamp the APACHE_CHECK_SSL_TOOLKIT
0N/A autoconf function or come up with some other autoconf checks to
0N/A search for libssl and libcrypto and properly add them to mod_ssl's
0N/A * Fix the worker MPM to use POD to kill child processes instead
0N/A of ap_os_killpg, regardless of how they should die. (Ryan Bloom)
0N/A * mod_cache: handle cache_control: no_cache "field_name" to enable
0N/A cacheing the response w/o header "field_name"
0N/A See RFC2616 section 14.9.1
0N/A * Scoreboard structures could be changed in the future such that
0N/A proper alignment is not maintained, leading to segfaults on
0N/A some systems. Cliff posted a patch to deal with this issue but
0N/A later recanted. See this message to dev@apr.apache.org:
0N/APRs that have been suspended forever waiting for someone to
0N/Aput them into 'the next release':
0N/A missing call to "setlocale();"
0N/A * PR#78: mod_include
0N/A Additional status for XBitHack directive
0N/A Mod_proxy doesn't allow change of error pages
0N/A Modified PATH environemnt variable is not passed, instead
0N/A Proxy doesn't deliver documents if not connected
0N/A proxy converts ~name to %7Ename when name starts with a dot (.)
0N/A * PR#537: mod_access
0N/A mod_access syntax allows hosts that should be restricted
0N/A * PR#557: mod_auth-any
0N/A ~UserHome directories are not honored in absolute pathname
0N/A requests (.htaccess)
0N/A Proxy FTP Authentication Fails
0N/A * PR#623: mod_include
0N/A A smarter "Last Modified" value for SSI documents (see PR number 600)
222N/A Request of "Options SymLinksIfGroupMatch"
222N/A Proxy doesn't do links right for OpenVMS files through ftp:
189N/A imap should read <MAP><AREA>*</MAP> too!
189N/A RLimitCPU and RLimitMEM don't apply to all children like they should
189N/A Uses cwd before filling it in, doesn't use syslog
189N/A it is useful to allow specifiction that root-owned symlinks
189N/A should always be followed
189N/A Controlling Access to Remote Proxies would be nice...
189N/A Adding authentication "on the fly" through the proxy module
189N/A request_config field in request_rec is moderately bogus
189N/A DoS attacks involving memory consumption
189N/A Logging of virtual server to error_log as well
189N/A ProxyRemote make a dead cycle.
189N/A * PR#1117: mod_auth-any
0N/A suexec does not parse arguments to #exec cmd
0N/A * PR#1145: mod_include
0N/A Allow for Last-Modified: without resorting to XBitHack
27N/A improvements to child spawning API
27N/A * PR#1166: mod_proxy
27N/A ``nph-'' not honored (no buffering) for ProxyRemote mapping
27N/A Apache cannot handle continuation line in headers
27N/A setlogin() is not called, causing problems with
e.g. identd
5N/A regerror() exists, use it
there is no way to keep per-connection per-module state
Add frame-safe anchor attribute to mod_autoindex links
CGI scripts running as Apache user: security (suexec etc.)
add allow,
deny/deny,allow warning to mod_access
Need to know "hit-rate" on proxy cache
Selective url-encode of log fields (or maybe a pseudo
I make mod_headers to modify request headers as well as
No HTTP_X_FORWARDED_FOR set...
ProxyRemote proxy requests fail authentication by firewall
mod_rewrite forms REQUEST_URI different than mod_cgi does
mod_headers should allow mod_log_config-style formats in
mod_proxy to support persistent conns?
patches to mod_include to allow for file tests
Suggestion for improving authentication modules and core source
code, problem with 401 and ErrorDocument
listing of proxy cache content
Allow modules to set user:group for execution.
adding auth_why to conn_rec
pipelined connections are not logged correctly
mod_rewrite doesn't pass Proxy Throughput on internal subrequests
HTTP Server Rebuild Line Needs Changing for the better
mod_status always displays 256 possible connection slots
Make online documentation search link back to my installation
Can not POST to ErrorDocument - Apache/1.3b6
Status module averages are for entire uptime
suexec for general access of user content?
Proposal for TimeZone directive
/server-info doesn't check for the virtual host to list the info
problem specifying ndbm library for build ?with autoconfigure
A small addition to
rotatelogs.c to improve program functionality.
AllowOverride FileInfo is too coarse
TimeOut applies to output of CGI scripts
<IfDenied> directive wanted
CGI's for general use still have to be run as another user
Cache file names in Proxy module
mailto tags and bundling bug report script
Support for System Resource Controller
When will Apache support P3P? Any Plans?
Propose that Apache recommend $UNIQUE_ID for all "session id"
suggestion: power up your Include directive :)
cannot limit some HTTP methods
No module specific data hook for per-connection data
* PR#3191: mod_negotiation
no way to set global quality-of-source (qs) coneg values
Accessing URL through proxy server corrupts data.
Some anonymous FTP URLs ask for authentication
New ErrorDocumentMatch directive
Need to be able to override shebang line to make CGI scripts
"Files" and "FilesMatch" regexp does not recognize bang as
Please allow CGI env variables (QUERY_STRING, ...) to be logged
Suggestion for better handling of Last-modified headers
mod_cgi prevents handling of OPTIONS requests
[PATCH] install as win32 service with domain account
Status: Cannot accept password-as-arg, we should prompt the
user when -k install/-k config with a user argument.
AllowOverride should have a 'CheckNone' and 'AllowNone' argument
Other bugs that need fixing:
* MaxRequestsPerChild measures connections, not requests.
Until someone has a better way, we'll probably just rename it
"MaxConnectionsPerChild".
* 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.
* SIGSEGV on Linux (glibc 2.1.2) isn't caught properly by a
sigwaiting thread. We need to work around this, perhaps unless
there is hope soon for a fixed glibc.
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 :-)
Other features that need writing:
* Finish infrastructure in core for async MPMs
* TODO in source -- just do an egrep on "TODO" and see what's there
* Jon Travis's <jtravis@covalent.net> patch to deal with thread-safe
issues with inet_ntoa. See message <20001201163220.A12827@covalent.net>
Status: This is being set aside until the IPv6 work is finished
so that we know exactly what is required.
* Martin Sojka <msojka@gmx.de>'s patch to add error reporting for failed
htpasswd actions due to a full /tmp volume (other programs may have
* Mike Abbott's <mja@trudge.engr.sgi.com> patches to improve
Status: These were written for 1.3, and are awaiting a port to
* Jim Winstead's <jimw@trainedmonkey.com> patch to add CookieDomain and
other small mod_usertrack features
* Dan Rench's <drench@xnet.com> patch to add allow the errmsg and timefmt
of SSI's to be modified in the config file. Patch is available in
* Which MPMs will be included with Apache 2.0?