Remove obsolete TODOs for mod_ssl: No, we don't - it was removed in r90511. DH keys are changed for every connection, SSL_OP_SINGLE_DH_USE is applied since mod_ssl's initial commit (r88988). We no longer have our own CRL callback (delegated to OpenSSL as of r1165056), so this is effectively moot. ssl_engine_pphrase.c needs to be simplified, not blown up further (see also https://issues.apache.org/bugzilla/show_bug.cgi?id=24031, which few [if any] people really seem to miss)

Address a todo listed in https://mail-archives.apache.org/mod_mbox/httpd-dev/200205.mbox/%3CPine.LNX.4.33.0205292300380.27841-100000%40mako.covalent.net%3E "init functions should return status code rather than ssl_die()" For diagnostic purposes, ssl_die() is still there, but instead of abruptly exit(1)ing, it will return APR_EGENERAL to the ssl_init_* callers in ssl_engine_init.c, and these will propagate the status back to ssl_init_Module.

Add another URL for the "removing the first connection filter in chain" bug. The other site seems dead.

Note a TODO for Lua 5.2.0 compatibility.

Note some things that should be considered, but not shortly before a release

Fix typo which was already fixed in 2.4.x.

Wish came true: CacheEnable works within LocationMatch, giving us mod_cache regexes for no additional cost.

Partial cleanup of trunk STATUS.

Fix assertion failure during very high load by preventing race condition between appending to the timeout queues and adding to the pollset. We don't add additional locking calls but only extend the present calls to include the apr_pollset_add. Therefore this hopefully should not cause too much performance regression. Add some comments Replace two AP_DEBUG_ASSERTS with better error handling

Found a bug in mpm_event, investigating

And we are at 2.3.16-dev

Getting ready to tag 2.3.15-dev

Point to list of undocumented modules in the docs subdirectory

remove solved Windows AcceptFilter blocker add mod_lua blocker

- start definitive list of modules not to be included in 2.4 - mod_fcgid not a blocker - install of beta on www.a.o done thanks to pquerna/infra

GA plan

two more blockers

remove two obsolete or done items

And we are at 2.3.15-dev

Get ready to tag 2.3.14

One more showstopper gone

And roll on to 2.3.14

Get ready to tag as 2.3.13-beta

The example configuration looks good enough to me

Use the new "ap_expr" expression parser. The old parser can still be used by setting the new directive SSILegacyExprParser

fill in conn_rec::current_thread for WinNT, NetWare, and OS/2 MPMs

one ap_expr item done

One I'm signing up to finish

Some STATUS thoughts from the 2.4.0 barcamp session

and bump to 2.3.13

Ready to tag 2.3.12-beta

1.3? nah

GA plan

Note release date

And roll onto 2.3.12

Ready to tag 2.3.11 as beta!

no docs should only be a blocker for GA

'votes'

We are going for beta

revert r1026746: If an unknown Content-* header is received for a PUT request, we must not ignore it but reply with 501 per RFC 2616 9.6. PR: 42978 Vetoed by Roy T. Fielding: http://mail-archives.apache.org/mod_mbox/httpd-dev/201102.mbox/%3C9A5A8B62-ED16-4BD3-97DA-8475026BB81D@gbiv.com%3E

Add support to set variables with the 'Define' directive. Change ap_resolve_env() to look for variables defined in this way. This allows to use the variables in the config using the ${VAR} syntax known from envvar interpolation

minor note... bunch of old, old stuff lives here

note release

And roll on 2.3.11...

Get ready to tag 2.3.10-alpha

Update with blocker notes and indicate that 2.3.10 was never released

Further trivial updates mod_disk_cache -> mod_cache_disk. No functional change.

Roll on .10

Get ready for tagging

Add a thing to do.

Update status of ap_expr and Limit

Fix hook sorting for the pre_config hook. Move modules/http/config2.m4 back to modules/http/config.m4, since this was only done as work-around in r88887 for the broken hook-sorting.

More whitespace

Whitespace commit

cleaning up / replacing ap_expr should be a blocker for a beta

timeline

If an unknown Content-* header is received for a PUT request, we must not ignore it but reply with 501 per RFC 2616 9.6. PR: 42978

If a malformed Content-Range header is received for a PUT request, we must not use the supplied content per RFC 2616 14.16. Send 400 response instead of ignoring the Content-Range. PR: 49825

Remind "Contributors looking for a mission" that the docs always need a lot of work.

Rename MaxRequestsPerChild to MaxConnectionsPerChild, which describes more accurately what the directive does. The old name still works but logs a warning.

Make the ssl expression parser thread-safe. It now requires bison instead of yacc. Also change the make file magic so that the real source file name is embedded in the debug info. The generated files have been created with flex 2.5.35/bison 2.4.1. The two 'no previous prototype' warnings are supposed to be fixed with the next flex version.

- add blocker for GA: review example config - remove old votes: * "what to do if parent dies" went nowhere after many years * we now have event MPM as default

Add ErrorLogFormat directive for configuring the error log format, including additional information that is logged once per connection or request. Add error log IDs for connections and request to allow correlating error log lines and the corresponding access log entry.

Add a TODO.

Bump up for 2.3.9-dev

Get ready to tag

And we are at 2.3.8-dev...

Prepare for tagging of 2.3.7

note some work in progress

note prelim plan

Add some open issues

ap_note_auth_failure seems to be broken in 2.2.x, too. So this is hardly a blocker for 2.4.

Fix authorization by user or IP/ENV/... Note ap_note_auth_failure() breakage in STATUS

Remove two blockers from the list. Fixed in r954590, r956202, r956204

start a list of modules which should probably be in experimental

Change the eor bucket to use a pre-cleanup to run the log transaction hook. This fixes some segfaults.

roll on 2.3.7 development...

Prep for 2.3.6 tag

Review of RFC 2616 violation issues in Bugzilla, by Igor Galić.

Vote. Even if there is no Vote.

another blocker for beta

Note intent to TR June9

not sure why blocker

odd assortment of spelling fixes

Note unresolved cause of current crashes in the test suite.

enable-mods-shared works on Darwin, drop from "NICE TO WRAP UP" list

Add AAA issue

Divide showstoppers into sub-sections for alpha, beta, ga Please add issues!

vote

Let the nominations for the name begin

test svn update

roll on to 2.3.6-dev

prepare for 2.3.4 tag

roll on to 2.3.4-dev

hrmm

remove leader, perchild, and threadpool MPMs (other than non-English bits in common parts of the manual)

Roll on to 2.3.2.

Add conn_rec::current_thread.

Bump to 2.3.2.

Note that 2.3.0 was tagged, but not released

2.3.1 is now in development.

o cast some votes o FreeBSD 5.x is no longer supported in even security patches.

seems to work

just a test of commits list

Reflow ALL of 2.x trunk STATUS to the same indention (indent 4, with a hanging '*')

Stale Bugzilla link as noted by Christophe Jaillet

Add comments re worker shutdown troubles on Linux and use of ap_os_killpg() in worker.

mod_ssl stuff which has been done.

Project name changed in Bugzilla when we released 2.2

A more generic explanation

New Win32 9x efforts are a nonstarter.

Not Applicable anymore

2.1 -> 2.3 changes. Only track the 2.3/2.4 release history.

Make the release date formats non-assbackwards and more readily readable. Corrects a regression from 2.0.

Update STATUS to reflect the current state of 2.1, including the backwards non-ISO date format.

The event MPM now works on Solaris 10. It is fast too.

2.1.6 got released.

More cleanup, organization to make README easier to follow.

Rearrange some deck chairs, just reformatting, axing unmaintained 'Binaries' section.

Move on to 2.1.7-dev.

Add some comments

Onward to 2.1.6-dev!

Add a vote.

httpd-config would be a great thing to replace apxs with.

AllowEncodedSlashes fix was backported in 2.0.52. Remove this STATUS file entry.

* STATUS Update to reflect the current state of affairs.

Update the Solaris 10 & Event MPM issue.

Note 2.1.3 released as alpha.

replying to why "the edge connection filter cannot be removed" is a showstopper

Sync with httpd-2.0 changes.

--showstoppers;

Word Up. ap_discard_request has used buckets for a couple years.

2.1 will always have >=APR-1.0. No need to worry about this then.

Update the Worker MPM+FreeBSD note. It all works under FreeBSD 5.3/5-STABLE.

mod_filter in 2.1 allows better ordering of the filter chain.

Remove this note about a test suite failure from 2 years ago. The test suite passes for me.

PR 9457 is marked as fixed. (this is also about APR, not httpd...)

I beleive this works fine on Glibc 2.2+, and I don't care about supporting glibc 2.1.2 anymore.

Remove the entire suspended forever section. All of these are extremely old GNATs PRs. If any of them are still valid, I welcome a new section with the proper Bugzilla PR numbers.

Remove old statements about mod_proxy. It now supports both load balancing and persistent connections.

Solaris 10+Event MPM doesnt work. Note it as a non-showstopper.

proxy request body handling

Onward ho to 2.1.3...

nagoya isn't our Bugzilla server anymore.

At long last, promote mod_cache and friends out of experimental/. Also, take quick pass through docs files to remove experimental label for them. Reviewed by: Paul Querna, Bill Stoddard, Justin Erenkrantz

Vote. mod_cache is ready for prime time

* STATUS: Update a bit...

* STATUS: The tree is now open for business. Some housekeeping will be undertaken now. But, that shouldn't effect anyone else.

voted. Note on the Worker MPM vote..

Not that I'd *need* to pre-declare a 2.1 alpha, but let's aim for something that can start the march towards 2.2 by close of AC next week... If your large 'pet' feature that just *has* to be in 2.2 can't be included by next Friday, I vote tough: wait for 2.3. Just like the mod_auth changes have now waited almost two years to be included in a release! Put up or go home...

Hey look, the latest commit to mod_info closes a PR from 1998. In other news, it looks like 99% of the bugs in that part of the STATUS file are completely bogus now...

repository is opened back up. it turns out it's basically impossible to delete a branch in CVS without manual intervention -- which i don't want to do. we'll do it when we convert to SVN, at which point it will be trivial. in the meanwhile, it's not hurting anything to leave it as-is. here's a note to remind us to make the change when we convert. so, hey guys. how 'bout that SVN? :)

Sync comments with STATUS/2.0 PR: Obtained from: Submitted by: Reviewed by:

If you ignore a problem long enough, it'll be fixed by someone else. FreeBSD 5.2-RC with either libthr or libkse enabled (instead of the default libc_r) is confirmed to be operational with the worker MPM. Add notes to STATUS and README.platforms (as there is now an upgrade path that may be reasonable for those needing threads). [No soak tests run, but it serves pages - which is an improvement over before]

Reference the bug database in the STATUS file.

This doesn't go in here... it goes in the BACKPORT section of the STATUS file that's on the APACHE_2_0_BRANCH

fix the config parser to support <Foo>..</Foo> containers (no arguments in the opening tag) supported by httpd 1.3. Without this change mod_perl 2.0's <Perl> sections are broken. PR: Obtained from: Submitted by: "Philippe M. Chiasson" <gozer@cpan.org> Reviewed by: stas

Resolve the ap_get_client_block() showstopper by looking at APR_BRIGADE_LAST of what we get from ap_get_brigade and set a nugget for our next call to pick up on.

these 3 bugs were reported long time ago, but there are still there. The entries include links to the threads where these bugs were discussed PR: Obtained from: Submitted by: Reviewed by:

More votes. PR: Obtained from: Submitted by: Reviewed by:

- revert my patch (r1.155) since it produces regressions. - add a note to STATUS that we don't forget the thing.

note an issue that needs an ongoing fix

Not sure about the forest v.s. the trees - if anyone wanted to call out the real tall trees that would be good.

Hop hop i am a yop. all i like to do is hop from finger top to finger top. i hop from left to right and then, hop hop I hop right back again.

Adding my vote for mod_cache

My thoughts on mod_cache being non-experimental.

Note votes on moving mod_cache on up

(grabbed from archive) Check the processed file on validity. If a line is not empty and not a comment, it must contain at least one colon. Otherwise exit with error code 7. Submitted by: Thom May <thom@planetarytramp.net> (on 2002-07-02) Kris Verbeeck <Kris.Verbeeck@ubizen.com> (on 2002-10-22)

Updated to reflect the backport of the auth_LDAP SSL standardization patch

some votes

Start tracking the PRs for known RFC 2616 violations. (Note that the test cases in the PR are ridiculously hard to follow.) These tend to be primarily in caching/proxy, but some are in the core.

Update the status of mod_auth_ldap

and add it to the racial memory

well, it's voting time, isn't it?

Stuff done and more WIBNIFs.

State the obvious (to some).

Keep track of the changes. My guess (although we will see in practice) is that 'new stuff' will stay here in HEAD, while problems will probably linger in APACHE_2_0. History will be relative to the tree one is viewing.

Would be helpful to put it on the right line

Ferg

Update votes...

And a chicken in every pot

Big brains here say this after trying 4.7 out...

I withdraw this complaint.

Actually this is working correctly as best I can tell.

Improves the user friendliness of the CacheRoot processing over my last pass. This version avoids the pool allocations but doesn't avoid all of the runtime checks. It no longer terminates during post-config processing. An error is logged once per worker, indicating that the CacheRoot needs to be set. [Paul J. Reder]

note release date

Added my two cents worth of votes and updated the cache entry in the experimental status section.

record a few votes. I need to catch up on emails regarding modules/experimental/ so I understand what I'm voting on there. I would hate to see mod_cache & friends, mod_ext_filter, and mod_charset_lite (esp. for EBCDIC) suddenly disappear.

PR 8482 is resolved

Commit rbowen's votes at his request, gotta get that access issue cleared up for him in avail.

it's a darned good thing these things are in the status file or I'd NEVER be able to keep up

Votes

add some votes

Clarified the wording on backward compatibility for 2.0, and added my vote

Add in my 2c on the 2.1 and versioning schemes.

Cast some votes (my 0s are true 0s in that I'll abstain for now)

Other humble opinions

Submitted for your consideration. Please consider the question of a new versioning schema, based on Rich Bowen's interest in getting the website docs back to a reliable state, and reflecting the released versions of Apache.

more cache work todo.

cache updates

Ryan asked that these be updated. As a no-longer-voting member, and no longer a voulenteer to fix these things, they are staying put amoung the open-issues, no-name-attached

Get ready for a potential release this afternoon

Clean this up a bit. 1. A patch isn't an available patch without a reference to the thing, and if it applies to 2.0. 2. It's ain't a showstopper if we released with the issue. Cull out all the silly voting statistics. 3. Update some bits I know are done. Drop some 'discussion' and leave the meat of the issue. 4. Escalate the CGI piping issue. That one is really a serious issue.

--with-static-support really satisfied my arguments here. YES, I'd still like that as the default, but let's restart the discussion, 2/2 vote with no activity is obviously not a pressing issue for anybody. [Really, a one-month limit on closing votes or dropping them from status would be nice.]

Added an entry to track mod_auth_ldap in experimental. If anybody knows of other ldap module issues, please add them.

Update the status of 2.0.42.

Update after the tag (and roll).

Moving on...

Update status of 2.0.41.

Update after the tag.

--release_non_showstoppers_but_it_would_be_real_nice_to_wrap_these_up;

Binaries for Apache for NetWare 2.0.40 are available

add a couple more binaries

some binaries are up, etc

Bump after the tag.

Document the content-length filter performance problems

Thunderstorms every day for the next week

libaprutil apps work on AIX again the similar Tru64 symptom must have been caused by something else since it went away

not much point in having this in STATUS any more

obviously the current apachectl interface isn't weird after all

shouldn't be picky about where the darn things are installed

Cast some votes.

A vetoing I will go... A vetoing I will go.. hi ho the merrio .. a vetoing I will go. ;-0

vote on wrapper scripts Keeping apachectl simple seems desirable. The tasks listed for httpd.sh seem reasonable too.

After an install, the default config files are useful. After an upgrade they aren't. If the config directive's syntax changed, Apache will tell you which one changed, and the docs are a better place to go for details then the default config files. BTW, the syntax of a directive hasn't changed for a VERY long time. If the semantics changed, then nobody will ever know that they need to look in the default config files to determine that they semantics changed, because there is no error message. The argument that you want to be able to get the defaults back doesn't hold water, because 99.9% of all admins backup their files before modifying them. If you aren't backing them up, then that is your problem. The argument that you may not have a full config directory also doesn't hold water, because the assumption is that the conf/ directory is coming from a working server.

sysconfdir/examples is my current favorite

vote on generating default config files. Abstaining for now on where they should live. It would be work to move them. Is it worth it? dunno

I'm VERY reluctant to get into this argument, but... I do consider it useful to have an updated httpd-std.conf with a new version. Some docs-project people would like to have a directory to dump all sorts of sample configurations.

perhaps there is some magic place to copy foo-std.conf...

--clues

a few updates

somebody please tell me I don't know how to read C code anymore (I guess the pool for an Apache socket will grow on every read/write operation that would block.)

gregames fixed this in APR a while back, I believe.

Just one more showstopper on my list and I'm not sure we can do anything about this one in 2.0 due to required handler changes...

:( (and if somebody can figure out how to build APR today, my hat is off to them)

I've had these for over a week but finally got around to signing them.

LinuxPPC binaries PR: Obtained from: Submitted by: Reviewed by:

so much for my alphabetizing skillz. ;-/

hpux binaries are up, or at least will be if my upload to daedalus would ever finish. also re-alphabetized the unices (except the Mandrake-foo, AIX-foo, etc, which need to get their names formatted in the standard way)

Up and ready

Updated binary status for NetWare

Here's the rh70 version, for the heck of it.

OS/2 2.0.39 binary is up.

Darwin5.5, Solaris8/intel, and Linux2.4 (Redhat7.2) are on daedalus. I also removed Darwin 5.3 since it's unlikely someone will want binaries for this kernel.

Add back in the original volunteers that I probably shouldn't have removed on the last run through.

Update the volunteer list for binary distributions.

bump after the tag

Getting ready for 2.0.39. Here we go again

moving on. :)

Update before the tag

David Reid is testing on BSDi for me and then off we go....

On hold for now.

Day is done, gone the sun, From the hills, from the lake, From the skies. All is well, safely rest, God is nigh.

Is this a new bug? NO. Let's get a release out for the fixes we have.

This has magically gone away, and prefork's graceful now passes my torture test. Things I changed: 1) make extraclean and rebuild after cvs up (it was not *that* old though) 2) removed mod_pop3, mod_ldap, mod_auth_ldap, and mod_php4 Shrug. On to check worker while I'm at it.

ugggggghhhhh

Add a note on the overall problem with HEAD requests generating wrong headers. Again, I believe this is a showstopper, but no one seems to agree with me. Point at patch for 500/501 problem - I don't plan on testing or committing it.

These are not showstoppers. Let's get a release out the door that includes fixes for all the bugs that have been fixed since dot-36.

moving on.

Add STATUS note on a newly discovered RFC violation, and add an update on the other one we know about. I firmly believe any known violation of the appropriate RFCs is a showstopper. So, I kindly ask for votes so I may know which way the group feels. (Showstopper == 'you will not vote positively for a release until fixed.')

getting ready for 2.0.38-dev

showstoppers-- I can't reproduce this one any more. Probably the "make extraclean" took care of it. Sorry for the confusion.

hot'n'sunny in cville

showstoppers--; thanks!

If nobody is proposing another name on list (that doesn't elicit "ewww gross" comments), this is not a showstopper.

reverse proxy is working again showstoppers -- PR: Obtained from: Submitted by: Reviewed by:

Clarify note on ap_discard_request_body(). If we want to change it, we should do it before .37 - hence its remaining as a showstopper. Add note on the 304 showstopper (aka PR 9673 in BZ).

Ordered lists are a whole lot more useful. It's almost surprising, in retrospect, that someone decided to send a dump to the STATUS file without some thought to it. Lists are easy, finding owners is a bit more difficult. Perhaps we jettison these in 30 days if there is no Status: notes to correspond to specific suggestions.

Updating 2.0.37 status. Even though this is dragging on, I'm glad to see us ironing these issues out before the release, lest we have to deal with another month's worth of duplicates of the same three PR's like we have with the last two releases. :) Also demoting the server-pushed CGI thing to not-a-showstopper since there has been zero to virtually zero talk about it since I added it to the showstoppers list.

wrowe just committed a fix for this showstopper

trying to keep the days straight (time flies regardless)

Update on showstoppers.

demote the hpux problem since there's an easy workaround add more status updates on the two biggies

make a note about an issue listed as a showstopper even with no change, user can --disable-threads so I wonder if this should really be a showstopper; it would be nice to get feedback from the user on my patch

--funky_build_issues

proxy showstopper

This has been resolved, but it never should have been added to this file. This was an issue for APR/APR-Util, not httpd.

Outstanding veto: the metadata flag should be per-type not per-bucket.

IMHO, known HTTP protocol violations fall under showstopper.

added a suggestion on one of the showstoppers

items_to_fix--; cliff++; (Cliff rocks for picking up after my mess and fixing this.)

Removing mod_bucketeer from the output filters seems to be the cure. Why? So, this isn't a showstopper since no one will enable bucketeer in production.

showstoppers++; (groan) I need to step away from this for a half hour or so because I'm completely confused as to how we are even getting in this state.

some comments after testing "apxs -g" from a 2.0.36 binbuild

Thanks for verifying these are valid, Doug. Now this [much shorter] list should be living in our STATUS file. Question of module maps and file layout is already off to the list.

All rather stale. Any new/remaining issues should be moved to CHANGES in the present tense, as opposed to the "Future port to 2.0". Heh

Now a moot issue. Isn't it time to just drop "PRs that have been suspended forever waiting for someone to put them into 'the next release'" if they have no status?

okay then :)

Yet another reason the MMN needs to get bumped. If there are changes that anyone's been holding off on because they would affect the MMN, might as well get them in now so we minimize the number of releases with MMN changes. I believe there are more renames pending in APR that should be done sooner rather than later, for example.

A clean prefork build starts up for me on Darwin. Will try with worker later.

I'm holding off on the pool patches until they settle down a bit. Getting ready to tag PRE2. We already know there will be a PRE3, but this is just to help me in my RM duties later on.

Nicely done, gentlemen. :) showstoppers -= 4;

hot

showstoppers--;

vote++; (cheer) showstopper++; (groan) Cliff can revoke this as a showstopper, but I think this'll get fixed as soon as I (or someone else) figures out how to return HTTP errors from a filter.

showstoppers--;

showstoppers--; Thanks, Bill!

Mostly harmless.

remove a vote which was resolved PR: Obtained from: Submitted by: Reviewed by:

Updated the release showstoppers section to reflect the completion of the critical suexec fixes

I *think* I tracked down all the issues that have come up on list in the last week or so and remain unresolved or needing consensus. If I missed something, PLEASE BE SURE TO ADD IT HERE.

71 degrees F and sunny, with scattered clouds and an increased chance of a 2.0.37 release

Veto rescinded: <20020523141308.A22513@lyra.org>

Sunny in c-ville...

Add a note about worker problems

This was never resolved. I still think it's a valid optimization, but the veto was not rescinded.

Feeling more confident in the worker MPM now, so I'm with Brian on this one.

Comin' at ya from the WWDC with my thoughts

Changing *any* default behavior is a bad idea without _very_ careful consideration, IMO, as it can silently break existing configs. Principle of least astonishment applies. Besides, the security argument is bubkus. Security through obscurity at best, and hardly even that.

bleh... bad idea.

Added my vote on minimal server version reporting

vote:change default config so that we only report our version instead of everything PR: Obtained from: Submitted by: Reviewed by:

Move mod_deflate from experimental to filters to represent our belief in its stability.

Update to reflect the release.

mod_mem_cache/mod_cache requires APR_HAS_ATOMIC_* feature macros.

This looks complete.

Done

for 2.0.36: showstoppers = 0; for 2.0.37: showstoppers = 1;

Rewrite ap_byterange_filter so that it can work with data that does not have a predetermined C-L - such as data that passes through mod_include. Previously, these requests would generate 416 since when the byterange filter ran, r->clength would be 0. r->clength is only guaranteed to be valid after C-L filter is run, but we need C-L to run after us so that our data can have a proper C-L returned. So, we need to rearrange the code so that we can deal with this case. Highlights: - Remove r->boundary since it is possible to have this self-contained in boundary's ctx. (May require MMN bump?) - Remove call to parse_byteranges in ap_set_byterange since this would wrongly return -1 for dynamic responses. We have to wait until we see EOS to call parse_byteranges. - Move bound_head computation inside the num_parts == 2 check. - Change a NULL brigade check to APR_BRIGADE_EMPTY - Move the 416 error return to after we've run through all ranges and found none of them to be valid.

Keep rollin', rollin', rollin' Though the streams are swollen, Keep them dogies rollin', Rawhide!

Updating my vote on the worker MPM

Modify the worker MPM to not accept() new connections until there is an available worker thread. This prevents queued connections from starving for processing time while long-running connections were hogging all the available threads.

Added a vote on MPMs

Update release info.

Even though this is in APR, it is causing big problems in Apache.

The worker's fdqueue is now a queue again.

Some of these worker bugs may or may not be showstoppers, but I'll let Striker decide that. I hope to have the top two of these finished in the next 24 hours, but the third I may need some help with.

I just posted a few bug fixes that need to go into the worker MPM, so I'm -1 here until those are fixed. I also think we need to do a lot more load/soak/burn testing on worker before we switch it to be the default. (and maybe by that time we'll have a better candidate for the hybrid MP/MT MPM).

Brain dump and a vote

demote the Content-Range/416/SSI problem to non-showstopper. It's been happening for ages, there are no PRs I'm aware of, and the core dumps that happened as a side effect in 2.0.35 seem to be resolved.

Clarify the remaining problems with Range: headers. The seg faults went away when we changed ap_send_error_response to get rid of resource filters. This was triggered because we generate a lot of invalid 416 responses for SSI. That happens because the complete length of the entity body (response) isn't known accurately until after the C/L filter runs. That's C/L's job, after all. Flipping the order of the byterange and C/L filters fixes the invalid Content-Range headers and 416's. But then the Content-Length header contains the full length, rather than the sum of the returned ranges. That's wrong, but doesn't seem too hard to fix.

This one appears to have been fixed by Greg and Jeff in http_protocol.c rev 1.408. showstoppers--;

um, one too many negatives in that paragraph. :)

showstoppers--; showstoppers+=2; :-( ...but these might have already been fixed, I just never got a response on-list about them and wanted to note them somewhere where they wouldn't be forgotten

hot hot hot

Adds support for reading trailers on input by exporting get_mime_headers to ap_get_mime_headers and calling it in the appropriate place in ap_http_filter. showstoppers--;

Too many incomplete changes to the code base.

NetWare binaries have been posted

Time for sleep

LinuxPPC binaries uploaded... PR: Obtained from: Submitted by: Reviewed by:

problem already fixed in 2.0.36-dev

"FINAL" is ... well, too final.

2.0.36 showstoppers

Red Hat 7.0 and 7.2 builds for i686 are in place.

Well, soon I hope - but yes, I'm on it.

AIX binaries on the way (compile and link god willing)

FreeBSD 4.5 and Solaris8/i386 binaries available.

Just added

Volunteering for i386-unknown-freebsd4.5 (from daedalus).

Signing up for the NetWare binaries

Added this am

I port perchild last week.

We should note that this has gone GA.

I made this available yesterday.

I can take care of LinuxPPC (YDL v2.1)... PR: Obtained from: Submitted by: Reviewed by:

Reformat this to look pretty (blatently stolen from the 1.3 STATUS file). AFAIK none of these are available, so my pardons if they are available.

I can take care of sol8/intel and some old sk00l lee-nux.

Minor formatting plus add arch

Remove a patch that is no longer needed (bug was fixed a LONG time ago) Add a sign-up area for people to volunteer to create binaries for 2.0.35.

With a cool spin cast...

Note about ap_cache_msec2hex() and ap_cache_hex2msec() to keep me from forgetting to fix it...

And it only took me what, 10 months? :)

Removed the single-byte header reads from mod_cgid

BUCKET FREELISTS Add an allocator-passing mechanism throughout the bucket brigades API. From Apache's standpoint, the apr_bucket_alloc_t* used throughout a given connection is stored in the conn_rec by the create_connection hook. That means it's the MPM's job to optimize recycling of apr_bucket_alloc_t's -- the MPM must ensure that no two threads can ever use the same one at the same time, for instance.

Ready for .34 evaluation.

Changed mod_cgi to not do single-byte reads to consume the script headers

record a vote.

Reclasses and notes; cheers reverberate through the cybersphere.

Not a showstopper IMO. But posting a patch anyway...

Yep, it's cold today in Raleigh

Vote and remove from showstopper status the win32 install PRs. Those will be fixed in the installer after the Apache tarball is rolled.

comment on a "showstopper" which is apparantly dormant.

cold

One showstopper down...

the mod_ssl provided with Apache >= 2.0 no longer has an independent version number

things are looking up...

status on ap_rset_content_type() fix.

It's trying to rain in C'ville....

OtherBill took care of my mod_negotiation concern. Ken: My change to retry after connect() failures didn't affect the pipes.

Remove the pools API change from the showstoppers section.

FYI...

mention an outstanding scoreboard alignment issue (not a showstopper :) )

This was fixed last week but I didn't notice the bug report or the STATUS entry until today. The bug report had to do with recursively removing all CVS directories from a person's home directory. PR: 9993

mention a mod_negotiation+Win32 concern

comments on a showstopper

This is not a showstopper, and it's definately not new.

33 has come and gone. Add note about the fact that we aren't doing config substitution properly. This is a showstopper, IMHO.

Looks like we'll have a nice weekend in the Carolinas.

axe the entry on graceful restart problems with worker I was too stupid to read the code to determine that the accept mutex failure log messages were harmless and not indicative of a real problem. I'll try to understand the conditions where I'm seeing connections dropped.

Drop one of the bad worker MPM symtoms (dropped connections) Even though I can hit it pretty readily on my Linux box, I can't reproduce on AIX or Solaris and I don't think anybody else noticed it on Linux.

mod_cache tweak to make it HTTP/1.1 compliant

These are fixed...

This finishes the mod_dir/mod_negotiation bug. This final part of the solution ensures that we don't lose filters if they are added later than we expect. The problem could be seen if a connection filter was added after a request-based filter was added in the past. The problem was that the request-based filters pointed to the first filter in the connection record, so the new connection filter was never called. Now, all filters are put on their correct filter lists, and we are sure to always update all pointers when adding a filter.

Add stupid idea

All I can say is "Ouch" this is nasty.

Our favorite showstopper is getting reports from our fans.

How can reach our goals if we never have them? So, I ask, "Can we kill all showstoppers in a week?" Are there enough developers paying attention to go showstopper hunting?

Fix --sysconfdir and --localstatedir, which were being ignored. Fix --enable-layout to work. showstoppers-- Note: As far as we can tell, autoconf will not let us override our layout settings from paramers like --prefix, etc. This is unfortunate, but we will not let this get in the way of --enable-layout working. This will be made clear in the docs. PR: 9124, 9873, 9885, 9888

When we fix this, make sure to close 9963 as this is that problem.

mention some graceful restart breakage with worker

No releases until these two are addressed. It doesn't make sense to release a new build if our top two filed issues aren't addressed.

I just tried to do a graceful restart while pounding my box with worker MPM and no segfaults. Am I missing something or were these fixed with Jeff's latest fixes?

Don't set bytes_sent to be 0 when r->assbackwards since this screws up logging. The content length filter seems to track bytes_sent properly regardless of HTTP version. The HTTP header filter knows not to send C-L (or any headers) to this client. showstoppers--; (Also fix typo of my last name in CHANGES...)

All the work on the lock APIs probably deserves a little note, and besides the module authors will appreciate the note directing them to the new APIs. showstoppers-- Also, fast_redirect is lame, toss it.

I learned about AP_MODE_EXHAUSTIVE while reading some of the new filter code earlier today. With this mode, the Perchild MPM can finally be fixed to work with filters. I have changed a comment in the core to make it clear that this mode is required, but I have mentioned how dangerous this mode is. Also add a comment to STATUS about my plans. Hopefully I'll have some time this week to hack through the MPM.

I prefer to remove stuff :(

mention a new showstopper introduced after 2.0.32

Mention the -DNO_DETACH changes (to implement apr_proc_detach prototype change). This removes a non-showstopper "nice-to-have". :)

Remember AddOutputFilterByType.

There is no way, IMHO, that we could ever maintain fast redirect. Toss please.

Hmmm, tried committing this in a hurry this aftn, seems it hung up. Another unseasonably warm day in Chicagoland - short sleeve shirt weather.

Grammariffic! Obtained from: Cliff Woolley

htpasswd is not my friend anymore.

stop using APR_BRIGADE_NORMALIZE() (okay, making a copy of it is cheating perhaps, but the name of the game is to get APR_BRIGADE_NORMALIZE() out of the API)

Remove and demote two showstoppers: - The support binaries now have an option for static builds at configure-time. However, it is questioned whether httpd can even be built statically. In point of fact, this exact item has been addressed, but httpd is still a bit murky so add a "nice to have" item. - The worker MPM showstopper has been questioned on-list and no justification or further information was provided. No one else is able to reproduce this case, therefore it is now demoted to a "nice to have." It is not fair to us to have a showstopper when there is no reproduction or explanation present. When information is provided on how to independently reproduce this case, this issue can be promoted. But, no one else has reported problems with CGI and worker.

Ramblin' Man.

Added a list of known API changes happening between now and GA (Please add to this list if you know of anything that I've missed.)

vote on .32

Fix a nice-to-have non-showstopper -- change the semantics for how we create a scoreboard from the ScoreBoardFile directive.

Cliff weighs in. (It seems no one has time to update STATUS themselves!)

Fix ap_directory_walk() per-dir merge bug seen when no <Directory /> is present. showstoppers-- Kudos to Jeff for finding it. Kudos to BrianP for leading us in the right direction. Kudos to OtherBill for pointing out the right way to fix this.

Another +1 for .32 beta (from Ian)

Per <005b01c1b55f$dcd917a0$50381b09@sashimi>, FirstBill is +1 for beta.

.32 has been rolled and placed in /dist/httpd

Since we had 3 +1s for JRE_1 to become .32, so it is. Tarballs are on the way.

Take care of a couple quickies -- mod_cgid shutdown and signals in the worker MPM (one_process case).

time to go buy hershey's kisses

Good work Mr. RM :)

A final (hopefully) roll of JRE_1 has been made. *Please* test.

Just a heads up to let people know what I'm working on. I ran into some problems with the new POD code in the worker MPM, and although I'm almost finished I didn't want people rolling a tarball without knowing of these problems first.

Performance optimization: updated the worker MPM to recycle per-transaction pools instead of destroying them. Based on Ian's benchmark testing, this reduces CPU utilization by about 1% on Solaris.

accept() error handling should be OK now. I don't have a strong opinion on what we should do if the parent dies, now that we don't have a real life example of that any more. It was a little confusing at first, but OTOH it kept serving requests.

Windows should not hold up an otherwise excellent release.

it's cold outside

I say the input filtering change are as complete as it will be for 2.0, so remove the note in STATUS.

Point at the JRE_1 tag if anyone is interested. IMHO, v31 is dead.

hmmm... not so sure anymore

why are the skies cloudy all day?

Explain why I'm sitting on .32.

Allow statically linked support binaries with the new --enable-static-support flag, and enable this behavior in the binbuild script. Also add a new --enable-static-htdbm flag. BTW, this takes care of 7/8th of a showstopper.

bit foggy PR: Obtained from: Submitted by: Reviewed by:

This issue is completed.

Changing my vote. I no longer have a problem with this. At the time, lots of changes were being made to worker to fix bugs, and it seemed premature. Haven't seen much of that as of late.

Move the 31 + fixes note to the beginning of the section.

- We're not going to be doing an unmodified 2.0.31 release, so noting the AP_DEBUG_ASSERT fix isn't helpful. - The vote for perchild being required for GA is unanimous against. No matter what interpretation you take, it isn't going to be a showstopper.

Split out anything that calls for a vote to be in its own section.

Be nice if we recorded that this problem might be fixed. =)

Rearrange the titles a bit.

Point at Greg's daedalus patches.

Cite the message id not the URL.

De-tabify. No content changes.

Separate showstoppers into next-beta and final-release sections. Change a couple of my votes. Some minor formatting changes (it's not code, so just deal with it ;-).

Ummmm experience on Daedalus implies we could do better :)

Link to patch that incorporates all relevant bug fixes seen so far. (only conflict from HEAD is in core.c where rbb/FirstBill snuck in some hook changes...)

Add note about APR_BRIGADE_NORMALIZE usage in core_input_filter. Consensus has overruled the hook ordering showstopper and is now demoted accordingly.

Revision 1.79 was bogus.

re-read this, and withdrew my vote. I thought we were talking binbuild, but it seems more general.

That Train Don't Run

Note fixes so that DB1 on a FreeBSD box work with apr-util.

brrrrrrrr

We seem to have fixes in our tree for all of our cited problems. If we could take the v31 tree and apply those revisions that fixes bugs to make v32, I'd be game for beta.

segfaults--;

Updating my vote on 2.0.31 based on the segfaults on daedalus

Jeff seems to have fixed this. Go Jeff!

While the tide was turning twords my take... this number of bugs in the Winnt mpm mean I can't go against Unix getting a good beta. Now the Q, is unix a good beta :-?

We don't have weather in Irvine. We do have that chick who does the NFL "weather" for Fox Sports. 'Nuf said.

Last week it was spring, this week it's winter. What's wrong with this picture?

This change keeps the server from allowing multiple instances to bind to the same port. Previously, this was necessary, because the Windows MPM was binding to the socket in both the parent and child. Today's code passes the attached socket to the child from the parent, so we don't need to re-attach in the child.

JinH found a shared memory problem IMHO a beta killer

Being stuck in a basement, I have no idea what the weather is like. :-(

record some votes and a couple of comments.

it's just not my day today ;(

oops PR: Obtained from: Submitted by: Reviewed by:

put in Manoj's comments

no idea what it is like outside.. I'm in a overheated office

Brrr... it's frigid in Chicago today! Hope FirstBill doesn't mind if I speak for us both here.

Not being able to bind to a socket is a fatal error. This makes all MPMs treat it as such. We now print a message to the console, and return a non-zero status code.

Yes, beta ready

+1 for Beta Despite its shortcommings, it is IMHO *much* better than our last beta, which in my mind qualifies it for beta.

Compiles on Linux w/ kernel 2.2

time for a walk around the neighborhood

Agreed

Some votes

Added some notes and votes

Vote...

I fully agree with Cliff and Justin on this one. Given the fact that this only appears in maintainer-mode, I take back my dev@httpd list comments that this would be a showstopper.

If they are willing to run in maintainer mode, they should be willing to read the release notes and apply patches to deal with this sort of thing.

- Add note on the bad AP_DEBUG_ASSERT call. Is this a showstopper or can it be indicated in the 2.0.31 release notes? Voice your opinions. - The worker CGI test failure Ken sees seems to be a problem with perl or httpd-test as no one else seems to be able to reproduce it. So, let's try to demote this. - I have no earthly idea what rbb is talking about. Whatever it may be, I can't fathom that it'd stop a release.

Um, where did my vote go?

Okay with the configurability. Not okay with the two parents idea.

Hopefully, this makes it easier to vote and decide...

High of 79 F yesterday, 50 F today.

I was leaning toward the configurable flavor, but OtherBill convinced me we should be more concerned with keeping the parent rock solid than adding extra layers to protect the children from a broken parent.

Updated my vote on shutting down child processes if the parent dies

Linux 2.4 (MDK 8.0) is fine here as reported earlier. I'm fine with adjusting that script in a -beta tarball since it isn't a code change. However, since official binbuilds are contributed by committers only, this isn't a big deal as they can be smart enough to use the updated script anyway. And, binbuild has some other issues that will be addressed shortly. So, even if it doesn't happen, oh well.

We have two factions here, let's give 'em what they want.

Added my vote that these are not showstoppers. Demote entries until there is majority support for showstopper status.

v31 beta notes are now in the status PR: Obtained from: Submitted by: Reviewed by:

It took all day, but we rolled .31.

- Add note about EBCDIC. - ap_[r]getline rewrite has landed. I think it is time for this STATUS item to be removed. Is there anything more w.r.t. input filtering that needs to be addressed?

There are really two issues here: - Can we *ever* build a static binary? That *is* a showstopper, I agree. And, Aaron has volunteered to fix this soon-ish. - Separate from that, is the question whether we should we *always* build a static binary? That I do not believe I *is* a showstopper. So, let's make the distinction and call for a vote.

Record my quasi-evident thought formally.

Some STATUS notes from jag

Was 81 deg F here yesterday

Still bright and sunny

Bright and sunny

some things just aren't important when compared with those horrible whitespace problems :)

One down, my fault

No way.

Don't let me forget about the scoreboard changes. Perchild works, but only on some platforms. ** Vote if you will not support a GA until perchild is portable. ** Volunteer for fixing binbuild.

Question everything.

Record my vote on the proposal to kill all the child processes if the parent crashes

Vote time.

If the RedirectMatch target is an abs_path, make it an absoluteURI

fix redirects for directories. fixup_dir was munging the URI before other fixup hooks (such as fixup_redir in mod_alias) had a chance to redirect the original directory URI.

ooops, pin the tail on the correct donkey. mod_alias is being bad, dunno about mod_rewrite.

Another showstopper.. protocol violation, this time.

correct the broken URL

32-dev is now it PR: Obtained from: Submitted by: Reviewed by:

its tag time v31 is being tagged

Fix a type-d'oh

If we die we need to write an epitaph..

Remove more spmt_os2 MPM leftovers.

Added logic to the worker MPM to skip the poll before accept if there's only one listener socket (based on the implementation in the prefork MPM). This also enables us to do single-listener unserialized accept

fix the problem where a scoreboard init failure could leave mod_cgid stranded a pre_mpm hook can now return failures, so problems in ap_create_scoreboard percolate back to a place where Apache can exit cleanly

Added a note about potential performance improvements for the worker MPM

Been bad - forgot to keep my thoughts up to date.

remove one showstopper and add the next

showstoppers--; APR knows when to let go of those pesty kids now.

Remove mod_auth_db since we've received enough votes (Justin, Ian, and Lars) and ample warning has been posted to dev@httpd. mod_auth_dbm should be able to take over all functionality of mod_auth_db with the AuthDBMType directive.

Never got rolled, problems with daedalus' thundering herd poll (among others I can't remember) weren't resolved in time.

Keep track of this so we don't forget about it. It's not a showstopper, but it is definately irratating. I'll look into it if nobody beats me to it.

I looked at this again with a clear head and I'm much more comfortable with the current solution. Thank you please drive through.

This is a showstopper. Let's remember it.

Add message-ids to ease issue tracking

Scoreboard update comming soon.

another note on the mod_cgid leaky pipes

Record my suggestion. (I'm *not* volunteering, but I think that'd be the way to go. It'd require substantial changes to how we handle binary builds.)

the load spikes seem to be gone.

call the plumber (or maybe the pilot)

make note of load spike problem on daedalus.

mention a binbuild SNAFU

Add a few votes.

oh well.. there goes the logging directive

Brian's fix looks good! with Brian's fixed applied and running the same tests: I don't hit any segfaults with yesterday's HEAD and APR_POOL_DEBUG defined I don't hit any segfaults with current HEAD and APR_POOL_DEBUG *not defined* I do hit segfaults down in apr_pool debug code with current HEAD and APR_POOL_DEBUG *defined* PR: Obtained from: Submitted by: Reviewed by:

add back in the out-of-descriptor segfault for now...

Note some things that belonged here ... mentioned on list

Updated STATUS to cover the worker segfault fixes

mention a couple of segfaults with worker

Remove the link to mbox.apachelabs.org. (Amazingly, that link worked.) But, we cite everything via the Message-ID to be archive-neutral. (FWIW, www.apachelabs.org has a feature to search on Message-ID.)

Does Apache need a fast way to exclude log lines?

Welcome to 2.0.31

Update the FreeBSD saga...

Cold and dark here...

This issue has turned into a FAQ. Start tracking it in STATUS so we can point people to this if they ask about this again.

Note this for the record. I have no idea if/when I'll have the time to implement this.

2.0.29 isn't going to be rolled.

You know, I got an A in my "Principles of Operating Systems" class last quarter. Ain't worth a cent now.

Threads on FreeBSD are not my friend.

showstoppers--; This appears to be due to an API change in LWP 5.61. A fix has been committed to httpd-test. (Works on Linux with prefork and worker now.)

Add note about worker MPM weirdness observed on FreeBSD and Cygwin. Does anyone else see this?

showstoppers--;

if (votes_to_remove_auth_db >= 3) --showstoppers; Assuming there isn't anything else that mod_auth_db does that mod_auth_dbm doesn't do.

showstoppers--

change ap_unix_create_privileged_process() to allocate enough space in the newargs array to allow for a null terminator. This allows suexec and mod_cgid to work together reliably. Apparently, rounding up the size to a multiple of 8 bytes in apr_palloc hid the problem sometimes. At other times, suexec would complain about a null command name.

does anyone need/use ndbm? if so should we make it a apr_util_dbm DB type? PR: Obtained from: Submitted by: Reviewed by:

add a 'AuthDBMType' so that users can select what type of DBM they want to use to authenticate users. I think this removes a 'showstopper' (deprecating mod_auth_db)

Updated the release showstoppers list to include a couple of outstanding problems

Add ap_os_killpg as well as unixd_killpg and beosd_killpg. Move the worker MPM to use ap_os_killpg in place of unixd_killpg as this makes it more portable. Ryan notes that really we should be using the POD mechanism to control death of the child processes, so this is temporary until we have that change in place. Note added to STATUS to that effect.

mention an ugly problem with scoreboard creation

Pick up Brad Nicholes' old fix so that make_exports.awk can handle AP_HOOK invocations which span lines.

Showstoppers and something not to be forgotten..

some acts of stupidity are better left undescribed

mention an easy but important problem in case I forget about it

Oh, FreeBSD, you don't like having multiple children, do you?

Fix segfault when restarting worker MPM. We can not examine the POD as a normal listener. There was an additional bug than what Ian submitted to fix listen.c (i.e. check sa for null): make_pipe_of_death does not zero out lr->bind_addr (since it uses apr_palloc). So, on Solaris, the first time through, bind_addr was probably 0, but the second time, not (pools!). (This caused Aaron's analysis that the patch worked the first time, but not the second time.) Submitted by: Ian Holsman Reviewed by: Justin Erenkrantz, Aaron Bannert

Wow, it just got really windy outside, and warm!

Any release that goes out that doesn't have this fixed is getting vetoed by me as anything other than alpha. I'm sick of this - it's been a pain for a long time and just hit me hard when one of the other admins on my 2.0 site tried to send SIGHUP and the thing died - it confused the hell out of him. =) Let's start tracking this down.

Add a note about terminating the server. PR: Obtained from: Submitted by: Reviewed by:

People are asking for this..