Release Meeting: 01 May 2017

(09:00:47 AM) gac410: Hi all ... Ready to begin?
(09:01:14 AM) cdot [~crawford@foswiki/developer/cdot] entered the room.
(09:05:28 AM) MichaelDaum: Hi there
(09:05:35 AM) gac410: Hi MichaelDaum
(09:06:20 AM) MichaelDaum: yea it is a holiday over here today
(09:06:57 AM) gac410: Ah... 1st of May. Forgot about that.
(09:07:37 AM) gac410: First up, Release Blockers.
(09:08:16 AM) gac410: https://foswiki.org/Tasks/Item14372 - UI::Draggable cannot be enabled in the config. The setting is missing from the Config.spec.
(09:09:03 AM) vrurg: Hi
(09:09:30 AM) MichaelDaum: gac410, I don't think Item14372 is any urgent
(09:09:32 AM) gac410: I didn't get from your comments on the task, if it was intentional or accidental. Should the task add it back, or were you telling the reporter to add it back themselves.
(09:09:55 AM) MichaelDaum: "The only thing you might probably be missing is the initializers"
(09:10:32 AM) gac410: Every other jquery plugin component has an enable setting. UI::Draggable was removed
(09:10:46 AM) MichaelDaum: this is all a %JQREQUIRE{"ui::draggable"}% will load in addition to jquery-ui itself
(09:11:18 AM) gac410: No it won't That's the point. It's disabled in the config by omission.
(09:12:52 AM) MichaelDaum: other than that I don't know how I could be clearer than what I've said on the task item
(09:13:48 AM) MichaelDaum: I wonder which error message Lynnwood is actually talking about
(09:13:55 AM) ***MichaelDaum asking him on the task
(09:13:58 AM) gac410: Did you remove the following from Config.spec intentionally or accidentially:
(09:14:00 AM) gac410: $Foswiki::cfg{JQueryPlugin}{Plugins}{'UI::Draggable'}{Enabled} = 1;
(09:14:25 AM) gac410: The code does NOT load and is not available for JQREQUIRE unless it's enabled in the config.
(09:15:00 AM) MichaelDaum: I seriously dont know what I did and why
(09:15:11 AM) MichaelDaum: too many things going on in the meantime
(09:15:40 AM) gac410: Anyway. For all the JQuery plugins ... that one, and only that one, is missing from Config.spec. So it's simple enought to add it back.
(09:15:55 AM) MichaelDaum: xactly
(09:16:22 AM) MichaelDaum: yet still to flag this task item URGENT there seems to be missing more
(09:16:24 AM) gac410: And my question I'm glad to do it, is *should* it be added back. I think the answer is yes, but I didn't want to cross you.
(09:16:45 AM) gac410: Mainly, because devs mostly ignore the Normals, and Lows never get seen :D
(09:16:52 AM) MichaelDaum: no problems. just go ahead ... as I already suggested on the task item :/
(09:17:41 AM) gac410: Okay. I wasn't sure if you were just telling Lynnwood to add it back locally.
(09:17:47 AM) gac410: I'll take care of it.
(09:18:22 AM) gac410: Okay ... next urgent: https://foswiki.org/Tasks/Item14349 EditRowPluigin doesn't work on IE 11 ... I don't have Edge to test if it's just 11 that's broken.
(09:18:45 AM) cdot: me neither
(09:18:59 AM) gac410: I did try to fix it, but got the UI all messed up. The issue is Button is illegal inside <a ...
(09:19:52 AM) gac410: and the docs do seem to support that we are in the wrong on this one. it's just that other browsers seem to be more accepting.
(09:19:55 AM) MichaelDaum: <button onClick="location.href = '$url'> thats bad
(09:20:08 AM) MichaelDaum: better is <a href="$url" class="foswikiButton">...</a>
(09:20:58 AM) gac410: Y, But in this case, I think the javascript is searching the DOM for elements, and it gets toasted by the change.
(09:21:29 AM) cdot: probably a simple change, but without IE not easy to test.
(09:22:15 AM) gac410: I'd be just happy to eliminate the <button element, which should fix ie, and not break it on firefox. My attempts were less than satisfactory.
(09:23:24 AM) MichaelDaum: a button elem is actually more intended to be used inside a form ... not for navigation
(09:23:47 AM) gac410: I know someone with a Windows 10 system. So I could probably test if I could figure out the fix.
(09:24:37 AM) MichaelDaum: ie edge should be just fine with anchors for navigation, without any button elem inside, yet styled as if it was a button using the foswikiButton class
(09:25:13 AM) cdot: just having a look at the code. AFAICT the JS doesn't care whether it's a button or an anchor.
(09:26:02 AM) gac410: All I know is I tried to change the html, I posted my patch to the task, and it completely breaks the button.
(09:26:11 AM) cdot: y, sure.
(09:26:28 AM) cdot: anyway, I don't see it as a release blocker. Only one browser out of many.
(09:27:02 AM) MichaelDaum: win10 is free for 90 days tech eval'ing: https://www.microsoft.com/en-in/evalcenter/evaluate-windows-10-enterprise
(09:27:27 AM) cdot: Edge market share standing at 1.61%
(09:27:57 AM) MichaelDaum: or have a vm from here: https://developer.microsoft.com/en-us/microsoft-edge/tools/vms/
(09:28:10 AM) Lynnwood [~Lynnwood@foswiki/developer/lynnwood] entered the room.
(09:28:22 AM) MichaelDaum: ready to be plugged in to a virtualbox
(09:28:37 AM) ***cdot isn't going to spend any time on it, sorry.
(09:31:17 AM) gac410: [off] I expect we should probably build 2.1.4 sometime in the next week or so - when some news is reported on a sanitization issue
(09:31:53 AM) gac410: cdot what's market share on IE 11 That's the one we know for sure is broken,
(09:32:09 AM) cdot: 4.44%
(09:33:29 AM) gac410: Okay. If I build 2.1.4 I'll just add a release note that EditTablePlugin does not work with IE 11
(09:35:59 AM) cdot: EditRowPlugin
(09:36:41 AM) gac410: Ah yeah.. sorry. Not enough coffee yet.
(09:38:15 AM) gac410: CDot ... Your propsal https://foswiki.org/Development/ImproveSupportForComments is still "Under Investigation" But the task https://foswiki.org/Tasks/Item13905 is "Ready to Merge"
(09:40:26 AM) gac410: If we can flip the proposal to Ready to merge to be consistent ?
(09:40:43 AM) cdot: yes, sure
(09:41:08 AM) cdot: the checkin is on a branch. Nothing in master yet.
(09:41:31 AM) gac410: Unless anyone sees any issues, I'll try to get the few proposals merged. Right. Nothing merged yet. But I guess they are "Ready" I'm just making sure
(09:41:38 AM) ***cdot really could have made use of this last week, would have saved a lot of pissing about
(09:42:44 AM) gac410: The proposal said - I'm investigatiing - The branch said "go for it" and RM was confused
(09:44:13 AM) gac410: So regarding features. Are the few from Crawford, and others enough to move to Release02x02? Or should we push 2.2 out another 4 months, maybe to September?
(09:45:00 AM) gac410: The https://foswiki.org/Development/ReleasePlan has a lot of work proposed for 2.2 thats all still needing development.
(09:46:27 AM) gac410: Julian, Michael and myself have the bulk of the proposed but incomplete work.
(09:47:02 AM) gac410: cdot has delivered all his proposals - though they are not merged yet.
(09:47:55 AM) cdot: https://foswiki.org/Tasks/Item14349 - there's a fix in there for you.
(09:48:16 AM) MichaelDaum: september sounds realistic
(09:48:36 AM) gac410: Thanks cdot. I'll give it a try later.
(09:49:33 AM) gac410: Okay. Julian not here to comment. September okay for me - I really do need to work on some of my proposals.
(09:50:25 AM) gac410: Lets set September 1st as our next "Feature freeze" date. Try to get code "Ready to Merge" before 9/1 ... and maybe we can have 2.2 out by the end of September.
(09:50:55 AM) gac410: cdot I'll try to get your work all merged by the end of this week. Hopefully sooner.
(09:51:07 AM) cdot: I'd like to see a 2.3, if only to catch some of the non-critical bugs I already fixed and posted for "next minor release"
(09:51:18 AM) gac410: 2.2
(09:51:26 AM) cdot: 2.schmoo
(09:51:35 AM) gac410: :D
(09:52:20 AM) cdot: I like incremental dev much more than sitting on releases waiting for checkins.
(09:52:52 AM) gac410: It would really be helpful if we would all try to bugfix into the Release02x02 branch. I'll cherry pick them for you cdot. and then merge them back into master.
(09:56:47 AM) gac410: cdot what would you do differently to make it more incremental?
(09:56:57 AM) cdot: I thought master *was* the 2.2 branch. It's the next release, innit?
(09:57:46 AM) gac410: Yes. master is 2.2 and will be the next "Feature" release. Bugfixes go into the Release*x* branches and merge into master.
(09:59:01 AM) gac410: This command: git cherry master Release02x01 -v will show you all the bugfixes that are missing from master release
(10:00:00 AM) gac410: On the other hand, flip that - git cherry Release02x01 master ... will show you all the commits that are not in the release.
(10:01:10 AM) gac410: Unfortunately anything that was cherry-picked from master -> Release02x01 and then merged back ... ends up duplicated, so the list is bigger than it needs to be.
(10:03:25 AM) gac410: Really it's up to consensus of the team. Do we move ahead with a 2.2 with the existing features or do we hold off for more "meat & potatoes"
(10:03:29 AM) cdot: ic. well, I only check in to master if I think something is critical, and therefore has to go into a patch. Otherwise, it's release material - don't want to risk "breaking" a patch.
(10:05:18 AM) gac410: The cleanest flow from making git commands most helpful, and avoiding duplicate commits but with different hash identifiesrs. is to always fix bugs on the Release ... so it ends up in the next patch. And merge into Master.
(10:05:28 AM) gac410: That way the sha1 hashes never change.
(10:07:00 AM) gac410: Need to know where a fix ended up? git tag --contains 9311e8673b5b or git branch --contains 9311e8673b5b
(10:07:17 AM) gac410: These only work if the sha1's don't change.
(10:14:58 AM) gac410: Anyway. Release 2.2. If we were to go ahead with 2.2 now, and leave the backlog for 2.3, there would be 8 features: Some very minor
(10:14:58 AM) gac410: CanonicalScriptURL, AdditiveACLs, QueryParsingIntoFoswikIRequest, RemoveHomegrownNet, MaketextSplitElements, AllMacrosInTemplateTopics, ImproveBulkRegistration and TopicComments
(10:18:43 AM) gac410: The current active developers has gotten quite small, and everyone is pretty busy with other demands. Even if we go ahead with 2.2, is there enough bandwidth to deal with the bugs and other work needed to get the release polished up?
(10:19:12 AM) gac410: With everyone's attention elsewhere, it just feels better to me to delay all this yet again.
(10:20:40 AM) gac410: Anyway, That's all I have. Thanks for the fix to try cdot, I'll get it tested. And I'll get your features merged asap.
(10:21:01 AM) gac410: Unless anyone has anything else, lets adjourn.
(10:21:07 AM) gac410: Thanks everyone.
(10:21:19 AM) cdot: Oh one thing....
(10:21:26 AM) gac410: okay
(10:21:46 AM) cdot: I noticed a lot of Config.spec don't say "CHECK="undefok"" even though it is
(10:22:02 AM) cdot: so get lots of spurious reports in Configure
(10:22:19 AM) cdot: just wondered what the consensus was re: changing the default
(10:22:36 AM) cdot: better to get the warning, perhaps?
(10:22:53 AM) gac410: doesn't that add the "undefined" checkbox to all the config variables which tends to be confusing?
(10:23:55 AM) gac410: There is one bug related in that area. If you try to "uninstall" an extension, save is blocked, because the deleted element becomes undefined which is not okay.
(10:25:02 AM) gac410: That's the only error that will actually prevent a save iirc. I have only run into it once or twice, but it's annoying when it happens.
(10:26:30 AM) gac410: I do seem to remember some real difficulty though with alternate defaults for undefok ... so I would make any changes in that area with great trepidation.
(10:26:50 AM) gac410: I can't remember the details but iirc it was ugly.
(10:29:41 AM) gac410: Anyway. I need to get moving. Physical Therapy appointment. (Though wondering if they could provide mental health therapy at times) :D Thanks everyone.
(10:30:16 AM) vrurg: Thanks gac410
(10:30:32 AM) gac410: cdot, After I merge a branch, I'll push a delete, to try to clean up the git tree a bit.
(10:32:14 AM) gac410: and vrurg, at some point we need to figure out how to go about merging the master work. There is unfortunately quite a bit. Easies to make it bite size pieces is to git branch master into managable chunks by commit hash, then merge small pieces.