1. Feature Requests review

Review the ReleasePlan topic.

Any features not "Merged To Core" decide on disposition. The entire list needs a review prior to release, but the below items are still listed as under construction:

AllowGuestsToUseRESTAsDefault Remove rest from the default list of AuthScripts GeorgeClark UnderConstruction
ConfigurableCookieNamesAndPaths Make cookie path and names configurable PaulHarvey UnderConstruction
CustomNewUserTemplates Enhance register script to specify a templatetopic param instead of hard-coded 'NewUserTemplate' PaulHarvey UnderConstruction
MoveCPANLibToSeparateContrib Move foswiki core lib/CPAN into its own CpanContrib SvenDowideit UnderConstruction
RedesignLoggerAPI Modernize the Log API and add LogDispatchContrib to core GeorgeClark UnderConstruction
WysiwygFormFields Making Formfields WYSIWYG capable SvenDowideit, PaulHarvey UnderConstruction

2.1 Release Meetings and Process

  • Does this timing work for the majority of the developers

For those present it appears to work fine. Some significant developers were missing.

  • Is twice-monthly a good schedule

Seems to be fine for now.

  • Target dates:
    • Feature Freeze:
    • Branch from trunk:
    • Alpha builds:
    • String freeze:
    • Beta start:
    • Release target:

Dates deferred until we get the list of release blocking tasks down to a much smaller number.

2. Tasks Review - Release Blockers

Review the ReleasePlanDiscussion02x00 topic.

This topic has lists of tasks, This will serve as an initial list of Blockers. Please set any tasks that should be fixed to Urgent, Minor release, and Released In 1.2.

  • Item12150 FoswikiCache - Duplicate - Crash in SQLite cache code. No action needed
  • Item12915 FoswikiCache - Being Worked On - The PageCache::genVariationKey() fails to consider the action (view or rest) when calculating a cache key. GeorgeClark agreed to resolve this one.
  • Item11498 FoswikiRender - Proposal Required - Implement a new whitelist mechanism to help sanitize the script zone. This is a new feature. no work done. Defer to 2.0
  • Item10149 SEARCH - Confirmed - versions query w/SEARCH, RCS store = hang No resolution ... need a search guru to help
  • Item11650 SlideShowPlugin - Duplicate - trunk SlideShowPlugin over-rides click on jqUIDialogLink styled buttons No action needed
  • Item12816 WebCreateNewTopic - Being Worked On - WebTopicCreator on trunk has the "Allow non-wikiword" button reverse. MichaelDaum agreed to take action to fix this one

  • Item11624 - Waiting for Feedback - reprev comment and number persists in TOPICINFO long after the event. No action needed It was a change that was fully reverted.
  • Item11895 - Duplicate - Stray broken html tag in AdminGroup No action needed. This was due to the ICON quoting changes.
  • Item12888 - Being Worked On - Unit tests are failing Need someone to recreate who understands I18N
  • Item12887 Configure - Confirmed - Configure loops on some versions of perl GeorgeClark will attempt to address, with help from JanKrueger
All Locale / I18N related issues will be deferred to 2.0 or a TBD I18N focused release. This includes:
  • Item12115 I18N - Waiting for Feedback - Compare revisions crashes when header compression is enabled Locale issue. Defer to a future release.
  • Item11449 FoswikiRender - Proposal Required - Topic name beginning with lower case letters cannot be reached.
  • Item11953 I18N - Confirmed - {Site}{Locale} breaks editing of SitePreferences
  • Item5437 I18N - Waiting for Feedback - UTF-8 fixes for Foswiki 2.0 (was Foswiki 1.1 but deferred, was Foswiki 1.0 but deferred, was T4.2 but deferred)
This next one, because it causes topic corruption needs to be resolved: The hot fix breaks a lot of unit tests.
  • Item12705 I18N, WysiwygPlugin - Confirmed - Verbatim blocks are corrupted on save by Wysiwyg editor if $Foswiki::cfg{Site}{Locale} = 'utf-8'; (hotfix available)

Performance issues:
  • Item11658 EditTablePlugin - Confirmed - EDITTABLE does not work in multi-line syntax. This plugin is deprecated. Replaced by EditRowPlugin. Not a blocker
  • Item10647 FoswikiPrefs - Waiting for Feedback - Preference variable that refers to itself hangs Foswiki Long standing issue, deferred multiple times due to difficulty of the fix. Defer again.
Other long standing issues:
  • Item12477 - Confirmed - Spurious entries being added to .changes Needs someone familiar with Store to fix. This should be a blocker
Configure changes: At the end of the meeting JanKrueger agreed to open a task to identify issues he has with the new configure UI.

Next meeting:

Next agenda we pick up with the list of tasks that have commits. figure out how to decide if we need to revert. Then move on into the MinorReleaseBlockers.


Conversation with #foswiki-release at Mon 02 Jun 2014 08:15:26 AM EDT

(08:57:39 AM) andreli [] entered the room.
(09:01:56 AM) andreli: fairly quiet here?
(09:02:23 AM) gac410: yeah I just sent a reminder to #foswiki. The original email said the meeting would be in that channel.
(09:04:57 AM) JulianLevens [] entered the room.
(09:05:05 AM) ***MichaelDaum just sent an email to people pinging them again
(09:05:30 AM) gac410: Okay thanks. We'll wait a couple more minutes and get started.
(09:05:50 AM) MichaelDaum:
(09:06:52 AM) gac410: I'm logging this locally. Rather than try to do a running log in the topic, I'll update it after the meeting. If that's okay. Unless someone wants to volunteer to take minutes.
(09:07:34 AM) MichaelDaum: next time
(09:09:18 AM) gac410: To keep things moving I think we should not get bogged down on any one issue. Lets keep this meeting as a review - and unless a quick decision can be made, table things for more in depth discussion in the main channel.
(09:09:51 AM) gac410: We decide what needs to be done here, and not necessarily solve the worlds problems :)
(09:10:44 AM) MichaelDaum: we should note who is taking actions on certain tasks to get things moving on
(09:11:41 AM) gac410: Yes. In the logs, and then make a note in each task. Not sure if "being worked on" is right, since that might change to waiting for feedback.
(09:12:12 AM) gac410: How is everyone else for time. It's early morning for me, so I don't mind waiting to see if others show.
(09:12:45 AM) JulianLevens: I'm OK for time, just at $work and may get distracted
(09:13:07 AM) MichaelDaum: I've got time as well
(09:13:47 AM) gac410: Well, lets do some easy stuff. Starting with ReleasePlan.
(09:14:16 AM) gac410: 1) AllowGuestToUseRESTasDefault. I set that to merged. I don't think anything major was left.
(09:15:23 AM) MichaelDaum: I'll have to review the NatEdit rest handlers as they might need better security flags
(09:15:45 AM) MichaelDaum: a couple of non-standard plugins still need updating
(09:16:10 AM) MichaelDaum: they will get a separate task so that we can close the release-blocking ones asap
(09:16:17 AM) gac410: Okay. But the feature itself is complete. It's polishing and debugging.
(09:16:29 AM) andreli: So do I understand right, that a not updated plugin using rest might become a security problem for foswiki, no?
(09:16:46 AM) gac410: The defaults have changed, so no,
(09:17:04 AM) MichaelDaum: andreli, the default settings are more restrictive now
(09:17:10 AM) gac410: The risk is that with new more secure defaults, plugins might stop working. Admin knob will override
(09:18:04 AM) gac410: CustomNewUserTemplates: I set that to defer to 2.0. There was no work done at all.
(09:18:11 AM) tsnfoo [] entered the room.
(09:18:11 AM) MichaelDaum: +1
(09:18:48 AM) gac410: RedesignLoggerAPI was my responsibility. It's basically done, but needs some more review. IIRC the incomplete part was in LogDispatchContrib which we decided to not include in core.
(09:19:33 AM) gac410: And the work remaining was on the playback side. eachEventSince() processing new enhanced log formats.
(09:20:21 AM) MichaelDaum: so can we switch to waiting-for-release?
(09:20:30 AM) MichaelDaum: if not please outline the remaining work there
(09:20:56 AM) gac410: Okay. I'lll take that responsibility.
(09:20:56 AM) gac410: ConfigurableCookieNamesAndPaths. Another one with no work.
(09:21:06 AM) gac410: I think we defer that to 2.0
(09:21:27 AM) MichaelDaum: there are some checkins to trunk as per
(09:21:51 AM) MichaelDaum: but these are against non-core plugins
(09:22:16 AM) gac410: yeah. I think we can just ignore that.
(09:22:20 AM) MichaelDaum: btw why does list this item
(09:22:47 AM) gac410: I think the configureable path was useful to the https redirect & cookie paths.
(09:22:53 AM) MichaelDaum: it seems the item is only about HttpsRedirectPlugin
(09:23:16 AM) gac410: Lets remove that task then. and defer.
(09:23:30 AM) MichaelDaum: ... decouple it from core
(09:23:33 AM) gac410: Yes.
(09:23:41 AM) MichaelDaum: sec ... will do
(09:24:15 AM) gac410: MoveCPANLibToSeparateContrib and WysiwygFormFields are the two biggies.
(09:25:58 AM) gac410: WysiwygFormFields really needs Sven feedback. I don't understand the work, and have no idea how operational it is.
(09:26:18 AM) gac410: Maybe pharvey has knowledge too.
(09:26:56 AM) MichaelDaum: reading up on wysiwyg formfields
(09:27:27 AM) andreli: It would sad to see Foswiki 1.2 delayed by wysiwyg related tasks, as NatEdit is coming to core
(09:27:57 AM) tsnfoo: Yeah.
(09:28:24 AM) MichaelDaum: wysiwyg formfields are done per ... anybody able to review the changesets there in how far work is done or not?
(09:28:33 AM) jast [] entered the room.
(09:28:35 AM) tsnfoo: I've wanted wysiwyg formfields for years, but I want 1.2 more :-)
(09:28:41 AM) gac410: I don't think it has to delay anything regardless. More a decision about whether it can exist in the distro as committed, or if we need to revert it when we fork trunk.
(09:29:06 AM) jast: sorry for being late. the floodgates of the ticket system suddenly opened wide.
(09:29:25 AM) gac410: Trunk appears pretty stable so I think it's unlikely that we need to revert WysiwygFormFields
(09:29:52 AM) MichaelDaum: the proposal refers to a new plugin WysiwygFormfieldPlugin ... which I don't see in core manifest
(09:31:18 AM) gac410: Okay. yeah, it exists in svn. So we need to get Sven or other feedback from anyone who was involved with it. And decide whether it gets pulled into 1.2, or left as a standalone plugin.
(09:33:15 AM) MichaelDaum: I skipped over the changes briefly
(09:34:06 AM) MichaelDaum: about 2/3 are on WysiwygFormfieldPlugin. the rest is improvements of TinyMCEPlugin 's javascript integration
(09:34:17 AM) MichaelDaum: which have general value in itself
(09:34:49 AM) MichaelDaum: it seems the bulk work on the actual proposal are kept in WysiwygFormfieldPlugin ... which hasn't been released yet, nor added to the core.
(09:35:49 AM) gac410: Well wysiwyg seems pretty stable. So lets leave the work as is, We can ping Sven for feedback. or defer the feature for 2.0. We have a lot of other stuff to get done for 1.2.
(09:35:49 AM) MichaelDaum: changes to the TinyMCEPlugin js all deal with adding the ability to instantiate two or more wysiwyggers on the same page ... which wasn't possible before
(09:37:23 AM) MichaelDaum: yes. we have to. last checkin to this feature was on Jun 22, 2012 ... 2 years ago.
(09:37:43 AM) MichaelDaum: pretty dead in my book
(09:38:38 AM) gac410: So last feature is the "MoveCPANLibToSeparateContrib" The work remaining on that one is building an easily releasable and updateable CpanContrib
(09:39:22 AM) gac410: I did some work on that, trying to figure out how to use cpanm to populate a lib/CPAN/lib directory late in the release process. But ran out of steam.
(09:40:29 AM) gac410: Minimum goal in my mind is a manually maintained lib/CPAN/lib that we can release as as separate package. And we need to update the install docs. That and testing.
(09:42:14 AM) gac410: To me the real valuable input has to come from people experienced with hosting sites. Since that's where prebuilt cpan libs are most valuable.
(09:42:27 AM) MichaelDaum: Oliver did some work on it during the last summit
(09:42:41 AM) MichaelDaum: Sven is still assigned to the proposal.
(09:43:14 AM) gac410: Sven mainly wanted cpan GONE so his debian pkgs would be clean. :)
(09:43:33 AM) MichaelDaum: Oh he is more than free to do that on a packaging level.
(09:43:58 AM) MichaelDaum: by just skipping them from tools/pkg/debian/foswiki-core.install
(09:44:48 AM) MichaelDaum: question is: what's the situation for other envs that have less luck
(09:45:14 AM) gac410: I think that between Sven & Oliver the feature is mostly complete. It's really a packaging issue now. Do we make the CpanContrib a default extension, so the out of box is complete. Or leave it out and doc how to add it for the few that really need it.
(09:45:32 AM) MichaelDaum: I prefer the latter
(09:45:41 AM) MichaelDaum: with CpanContrib being mentioned in the core docs
(09:46:24 AM) gac410: I also prefer to just install the deps so I agree. In that case this is "Merged to core" and we can deal with packaging issues when we start the beta process
(09:47:00 AM) MichaelDaum: Oliver is listed as maintainer of CpanContrib
(09:47:15 AM) MichaelDaum: so things are fine afacs
(09:47:54 AM) MichaelDaum: we need to make sure that actually is published
(09:48:03 AM) gac410: CDot did undo a bit. Originally ALL of lib/CPAN was gone. He added back in CGI::Session since configure won't run without it. I think the goal was that bin/configure should run on most installations without dependencies.
(09:48:33 AM) jast: well, configure... that's still a big can of worms anyway :-)
(09:49:02 AM) gac410: Agreed. I think it will become an item on "BuildingARelease" To refresh, build and release CpanContrib along with the other extensions.
(09:49:54 AM) jast: the way I see it, configure still requires a few fairly invasive changes, and in that process it may be a low-hanging fruit to eliminate the dependency on CGI::Session
(09:50:40 AM) gac410: CGI is actually non-default in latest perl as well.
(09:50:44 AM) MichaelDaum: are there any compatibility issues to be noted on ... e.g. installing it on foswiki < 1.2?
(09:50:54 AM) MichaelDaum: and clashing with lib/CPAN already there?
(09:51:08 AM) jast: okay, getting rid of CGI is... more effort...
(09:51:10 AM) gac410: I don't think so. It will just overlay lib/CPAN/lib. There were no structural changes.
(09:52:15 AM) MichaelDaum: what's the ModificationPolicy for CpanContrib?
(09:52:26 AM) gac410: So installing lib/CPAN/lib would just refresh the "last resort" cpan path on an existing 1.1.x or 1.0.x system. The risk with cpan is that a new version of what we install is incompatible with a dependency found in the system libs.
(09:52:30 AM) jast: following release process IMO
(09:52:34 AM) gac410: +1
(09:52:38 AM) MichaelDaum: k
(09:52:54 AM) MichaelDaum: so it is a half-core plugin
(09:52:56 AM) jast: I don't think we want people randomly adding or removing stuff
(09:53:20 AM) MichaelDaum: i.e. it won't update between releases
(09:53:45 AM) jast: yeah, seems fine to me that way
(09:53:46 AM) gac410: hm. Well we do update other extensions between releases. Esp. for critical bugs or security work.
(09:54:24 AM) MichaelDaum: ok let's leave it like it is now. we'll see.
(09:55:13 AM) gac410: For extensions that need cpan, it needs to be considered if adding to CpanContrib or distributing lib/CPAN/lib with the extension (which may NOT be what the customer wants) makes more sense.
(09:55:33 AM) gac410: But lets not solve that here :D
(09:56:52 AM) gac410: Okay . It's been almost an hour. Next on agenda was a quick straw poll. How do the date/time work out for everyone that's here anyway smile Do we continue with 1st & 3rd Monday, 1300Z or would another day/time be better.
(09:57:11 AM) MichaelDaum: time is okay for me
(09:57:16 AM) jast: I don't think there's a better time for me
(09:57:25 AM) JulianLevens: Ok for me
(09:58:27 AM) gac410: Okay. We'll stay on this schedule for now. And of course impromptu release discussions in #foswiki can happen any time.
(09:59:02 AM) ***MichaelDaum marking MoveCPANLibToSeparateContrib MergedToCore
(09:59:12 AM) gac410: Okay great. Thanks!
(10:00:01 AM) MichaelDaum: 3. Tasks Review - Normal and Low priority ?
(10:00:11 AM) gac410: Do we want to try to set some target dates for beta, etc. Or just keep moving for a while.
(10:00:39 AM) gac410: Well we didn't do 2. ReleaseBlockers from the ReleasePlanDiscussion02x00 topic.
(10:01:13 AM) MichaelDaum: right. erm.
(10:01:37 AM) MichaelDaum: a target date really depends on people finishing
(10:01:51 AM) gac410: yeah agreed
(10:02:32 AM) MichaelDaum: so let's review them to get a feeling how much work is still infront
(10:03:30 AM) gac410: So for everyone, the list is at
(10:04:38 AM) gac410: Under "unfinished features" Covering just the ones that are in an open state. I think Item12475 can be closed.
(10:05:08 AM) MichaelDaum: Item12150 is a duplicate on an already closed bug
(10:05:13 AM) gac410: MichaelDaum and I need to hash out Item12915. I'm convinced there is a bug there somewhere.
(10:05:21 AM) MichaelDaum: agreed
(10:05:51 AM) gac410: I think I can even propose a fix. If you want I'll give it a try.
(10:05:53 AM) MichaelDaum: I didn't fidn time to follow your recipe to repro the cache collision. but I am pretty sure you are right.
(10:06:09 AM) Lynnwood [] entered the room.
(10:06:16 AM) MichaelDaum: the fix is adding the script name to the variation key
(10:06:34 AM) MichaelDaum: it will only be view or rest, but thats exactly the way you produced the clash
(10:06:57 AM) gac410: yeah. It looked pretty straight forward. I was going to add the script prefix, since for a single user, view / viewauth should resolve to the same cache.
(10:07:13 AM) MichaelDaum: definitely
(10:07:46 AM) MichaelDaum: no other script will ever scan the page cache
(10:08:05 AM) MichaelDaum: atm
(10:08:10 AM) jast: gah... I was just looking at configure in various versions of internet explorer to see if there are any more serious issues... and it turns out you can't use configure if your system time differs from the server's by too much
(10:08:18 AM) gac410: Item11498 is a feature without a proposal. No work. We can defer it. IMO
(10:08:34 AM) gac410: jast. Yeah. that's true. :(
(10:08:43 AM) jast: and I can't change the system time on those machines
(10:09:11 AM) jast: in the meantime I found a syntax error in the template that generates the associated error message, so currently on trunk it shows an exception instead
(10:09:12 AM) gac410: IIRC Sven ran into that. It caused a bit of strife
(10:09:27 AM) gac410: open a task please :)
(10:10:31 AM) gac410: Item10149 - rcs hang with search on "versions" only work was a doc statement to "not do that" ... I'm not a search guru. Not sure what to make of it.
(10:11:06 AM) MichaelDaum: ... all it takes is a copy-paste into a feature proposal ... but that's not going to happen within 1.2.0, will it?
(10:11:26 AM) gac410: No. I'd say defer it for 2.0
(10:11:43 AM) gac410: That was pharvey effort when he was being ambitious. :)
(10:12:07 AM) MichaelDaum: okay flagging it accordingly
(10:13:57 AM) timlegge [] entered the room.
(10:14:05 AM) jast: Item12816 seems straightforward, except for maybe forward-porting the extra JS validation from 1.1.x
(10:14:22 AM) jast: seems realistic and sensible for 1.2
(10:14:51 AM) gac410: Actually the js is badly broken. The WebCreateNewTopic doesn't work at all like 1.1.x
(10:15:31 AM) jast: is there anything that's better in the trunk version? otherwise maybe we can just nuke it
(10:15:37 AM) gac410: I think that one is a real blocker. But I could not figure out why the js was non-functional. It doesn't do the wikiword name validatino, etc
(10:15:52 AM) jast: hum
(10:16:14 AM) gac410: I think it was Arthur doing restructuring of the topics. Including Included Includes. etc. I get totally lost.
(10:16:43 AM) jast: is this pattern-specific stuff?
(10:17:02 AM) gac410: Yeah I think so.
(10:17:15 AM) MichaelDaum: "Pick from list" doesn't work anymore ... should be an autocompletion imho
(10:17:37 AM) MichaelDaum: what else is broken about it?
(10:17:50 AM) gac410: Yeah, And when you visit it the first time, it pre-fills the new topic name with WebCreateNewTopic
(10:18:50 AM) gac410: As you type the name, it's supposed to be wikifying it, removing invalid characters, etc. and showing the correct name to the right.
(10:19:57 AM) MichaelDaum: I'll take actions on this one
(10:20:06 AM) gac410: Great thank you!
(10:20:24 AM) gac410: I tried one time to trace / debug the js and got completely lost.
(10:20:28 AM) jast: I played with turning the parent selector into a select2, but the code/markup I've got isn't particularly nice
(10:20:31 AM) jast: just an idea, though
(10:21:04 AM) gac410: As long as it just works out of the box I'll be happy ;)
(10:21:06 AM) MichaelDaum: there's a parent-topic-picker in natedit that may be reused
(10:23:14 AM) jast: I can tell you why the JS doesn't work in the current one... it tries to bind using ID selectors, but the input fields don't have ID attributes
(10:23:20 AM) jast: bind events
(10:23:50 AM) gac410: Excellent. :)
(10:23:51 AM) MichaelDaum: ah okay
(10:24:16 AM) jast: man, I haven't looked at pattern in a long time
(10:24:18 AM) MichaelDaum: please go ahead fixing, jast
(10:24:39 AM) gac410: So I think we can move on. Let's not fix stuff here.
(10:24:48 AM) MichaelDaum: yes please move on
(10:24:57 AM) jast: and now I see a hierarchy of #patternScreen #patternPageShadow #patternPage #patternWrapper #patternOuter #patternFloatWrap #patternMain etc.
(10:25:06 AM) jast: right
(10:25:14 AM) MichaelDaum: btw does anybody slim down the item lists in ?
(10:25:36 AM) MichaelDaum: jast, yea :(
(10:25:41 AM) gac410: I was going to but didn't want to edit collide.
(10:26:01 AM) gac410: I'll edit later when creating the notes for this discussion.
(10:26:25 AM) jast: so, I think from that category that leaves Item11650, SlideShowPlugin
(10:26:39 AM) gac410: That can be closed I think
(10:26:48 AM) MichaelDaum: Item11650 is (1) a duplicate and (2) not applying to the current version of SlideShowPlugin anymore
(10:26:53 AM) jast: cool
(10:26:56 AM) gac410: And dups a closed task.
(10:27:40 AM) gac410: Next category. stalled feature. fully reverted. We can just drop that one.
(10:28:13 AM) gac410: Item11895 was the single/double icon quote debacle.
(10:28:22 AM) gac410: We can close it.
(10:28:34 AM) ***MichaelDaum wonders how much trash we will have to move off 2.0'ers sholders to see the light of real work.
(10:29:01 AM) jast: the proper long-term fix is to define a new macro syntax... that's foswiki 2.0 or later, though :)
(10:29:22 AM) MichaelDaum: ... or educate people to properly escape their quotes
(10:29:29 AM) jast: you can't always escape quotes correctly
(10:29:45 AM) gac410: Item12888 - Crawford needs someone to recreate. He doesn't have the issue. I have the issue and get lost in char encoding.
(10:30:22 AM) jast: is the issue documented? the Item claims all things listed there are fixed
(10:30:23 AM) MichaelDaum: all test on 12888 have a green tick ... which ones still fail?
(10:30:52 AM) gac410: We get it in the email every few hours. Hang on
(10:31:22 AM) MichaelDaum:
F: TranslatorTests::testROUNDTRIP_Item9973_cp1251 * F: TranslatorTests::testTML2HTML_Item9973_cp1251
(10:31:23 AM) jast: oh, translator tests, then
(10:31:34 AM) gac410: * F: TranslatorTests::testROUNDTRIP_Item9973_cp1251 * F: TranslatorTests::testTML2HTML_Item9973_cp1251
(10:32:25 AM) gac410: yeah. A character gets "changed" not necessarily corrupted. The fixes for the verbatim block corruption caused this test to fail.
(10:32:31 AM) gac410: iirc anyway
(10:32:50 AM) MichaelDaum: too bad CDot isn't here to comment
(10:33:28 AM) gac410: Yeah probably best to table this one. I don't understand it.
(10:34:21 AM) gac410: is a bug in perl that Jast's changes triggered. There is a foreach (regex) that the regex position pointer fails to move on perl 5.8.8
(10:35:00 AM) gac410: When I use perlbrew to run bin/configure, it runs my system out of memory and if I don't pay attention, crashes it
(10:35:46 AM) gac410: So i can probably fix it since I can recreate it. I'll take the task
(10:37:02 AM) gac410: ... needs someone I18N knowledgable. Been sitting out there for a couple of years. Defer to 2.0 for the I18N work?
(10:40:15 AM) jast: Item12887: do we know that's the cause? one of the execution branches in the loop updates pos() manually, there may be another bug in there
(10:40:52 AM) jast: Item12115 only happens with Locale enabled. we have pretty much given up on Locale...
(10:41:13 AM) gac410: I'm pretty sure. I used print statements to show the position and it only fails on 5.8.8.
(10:41:25 AM) jast: right...
(10:41:51 AM) jast: we can probably put a manual pos() update in the other branch of the if block to fix it
(10:42:51 AM) gac410: I was just going to change the for loop to not use the foreach match type processing. But if you have an idea how to resolve it, add some feedback and I'll test it later.
(10:43:18 AM) jast: possibly something as simple as: pos($_[1]) = $+[0];
(10:43:29 AM) andreli left the room (quit: Quit: Page closed).
(10:43:50 AM) gac410: okay. lets keep going though. starting to lose people.
(10:44:10 AM) MichaelDaum: can't repro Item12115
(10:44:26 AM) jast: again, Locale only
(10:44:28 AM) MichaelDaum: hard to repro anything that requires UseLocale switched on
(10:44:47 AM) jast: let's not spend too much time on Locale stuff
(10:44:49 AM) MichaelDaum: as far as I see switching on UseLocale immediately crashes my foswiki
(10:45:02 AM) jast: personally I think we need to make a Big Decision(tm) on locale support
(10:45:16 AM) MichaelDaum: .... or taintedness support
(10:45:58 AM) gac410: I think we decided that all I18N / Locale stuff would defer to 2.0. Maybe even next major release would be heavily focused on all that stuff. utf8, locale, etc.
(10:46:09 AM) MichaelDaum: I'll leave HttpCompress on for a couple of days on daily use. See if it crashes even though UseLocale is disabled.
(10:46:59 AM) gac410: So maybe the decision for 1.2 is we need to Release note and have a config checker warn that Locale is not recommended. / experimental support. Which will be a real bummer for any sites successfully using it.
(10:47:34 AM) gac410: really concerns me.
(10:48:42 AM) gac410: The hotfix broke a lot of tests.
(10:49:02 AM) ***MichaelDaum switched to "Answered"
(10:49:28 AM) MichaelDaum: yea 12705 is a constant pita
(10:49:38 AM) MichaelDaum: even for any new 1.1.9 deployment
(10:50:24 AM) MichaelDaum: 1.1.9 without patching lib/Foswiki/Plugins/WysiwygPlugin/ is pretty useless for german speakers
(10:50:47 AM) jast: one of our guys here has strong feelings about nuking WysiwygPlugin from orbit... with space lasers... ;)
(10:50:56 AM) jast: in any case, WysiwygPlugin is very, very fragile
(10:51:32 AM) ***gac410 has the battle scars. he swears to never say "that sounds like a simple change" again.
(10:51:33 AM) MichaelDaum: jast, tell him to contribute a better alternative ... and then use strong words ;)
(10:52:58 AM) MichaelDaum: again: it needs CDot's help to nail this one.
(10:53:02 AM) jast: first have to come up with an alternative... it's something that's on our agenda, but we are fairly likely to solve it in ways not terribly useful for most Foswikis
(10:53:16 AM) MichaelDaum: jast, sounds familiar
(10:53:24 AM) jast: just like your NatEdit approach solves it, but in ways not applicable to all setups
(10:53:58 AM) MichaelDaum: no it doesn't solve encoding problems
(10:54:09 AM) jast: well, you don't use WysiwygPlugin with it, do you :)
(10:54:24 AM) MichaelDaum: NatEdit embeds TinyMCE
(10:54:41 AM) MichaelDaum: which in turn uses WysiwygPlugin
(10:54:43 AM) jast: oh, okay... then I'm misinformed about NatEdit
(10:54:57 AM) gac410: We are approaching 2 hours. Do you want to stop the task review here and maybe talk about anything else on the agenda. last major point is Migration to git.
(10:54:57 AM) jast: anyway, back to the list
(10:55:21 AM) gac410: Or spend a bit more time on the tasks list. See if we can complete ReleasePlanDiscussion02x00 doc.
(10:56:06 AM) MichaelDaum: yea, walking tru all items in full depth is probably too much for the first release meeting.
(10:56:12 AM) jast: maybe continue the task list next time? it's hard to stay focused
(10:56:15 AM) MichaelDaum: the more as some people are missing out today
(10:56:25 AM) gac410: So all tasks mentioning i18n are tabled.
(10:56:30 AM) jast: +1
(10:56:51 AM) MichaelDaum: +1
(10:57:08 AM) gac410: is for a non-core extension ... so we can drop that from our list.
(10:57:43 AM) MichaelDaum: let's move on
(10:58:00 AM) gac410: is a very long standing issue, loop on recursive pref settings. We can probably defer. iirc fix was felt to be very difficult
(10:58:40 AM) MichaelDaum: besides %SET being a known makro now
(10:58:42 AM) gac410: . spurious entries in .changes files. probably needs Crawford. It's a store issue.
(11:00:07 AM) gac410: comes up again. And since EditTablePlugin is deprecated. :P
(11:00:59 AM) gac410: The next catetory: Code-quality issues that perhaps do not impact users that much .... every time I enable that macro it breaks the topic rendering, and I have no idea why.
(11:02:57 AM) MichaelDaum: yea tried my luck to fix it
(11:03:42 AM) gac410: Anyway Item11521 is already No Action, Item11908 is waiting for release, Item9983. I think that one is fixed.
(11:04:17 AM) gac410: Item11683 is also no action. So that category is done I think.
(11:05:33 AM) gac410: Which brings us to the big category I added. Tasks with commits that are still "being worked on". We really need everyone to look at tasks they've got open and decide what the disposition should be.
(11:05:56 AM) gac410: I don't think we should spend time heere going through all 17 item by item.
(11:07:21 AM) gac410: And that brings us to the end of the tasks list. The last 3 are commits for tasks marked for 2.0, so should be lumped into the last batch. Maybe if everyone could grab one or two over the next two weeks to review. That would really be helpful.
(11:08:15 AM) gac410: This is probably THE MOST PAINFUL part of the release process. I really need to review all tasks that show up in the svn commit log, and make sure they are in an appropriate state for a new release. Should they be documented, etc.
(11:08:42 AM) MichaelDaum: the code quality list didn't return any results. so the twisty was broken.
(11:09:09 AM) gac410: Yeah. Something breaks the twisty. I just could not figure out what.
(11:09:32 AM) MichaelDaum: %TWISTY% %SEARCH{footer="$percntENDTWISTY$percnt" }% .... -> breakage
(11:10:11 AM) gac410: hm.... Why just that one item, and not all the rest.
(11:10:44 AM) gac410: Anyway, that is a bit of a sidetrack :)
(11:11:46 AM) gac410: It's been 2 hours. Lets call the official release meeting closed, but feel free to hang around and figure out tasks. smile
(11:12:03 AM) gac410: Next meeting. 2 weeks. 6/16, 1300Z
(11:12:34 AM) MichaelDaum: gac410, thanks.
(11:13:35 AM) gac410: Next agenda we pick up with the list of tasks that have commits. figure out how to decide if we need to revert. Then move on into the MinorReleaseBlockers Does that sound fair?
(11:15:03 AM) MichaelDaum: yes sure :)
(11:16:02 AM) JulianLevens: Ok, I might even contribute a bit
(11:16:23 AM) gac410: I'll get the irc log and minutes updated. Yes please! Contributions needed to get this thing out for sure!

Topic revision: r6 - 27 Jul 2015, GeorgeClark - This page was cached on 05 Jun 2020 - 08:08.

The copyright of the content on this website is held by the contributing authors, except where stated elsewhere. See Copyright Statement. Creative Commons License    Legal Imprint    Privacy Policy