From MozillaWiki
Jump to: navigation, search

« previous meeting | index | next week »

Places Meeting 4pm 11.29.06

[4:00pm] sspitzer: myk / dietrich / thunder / sayrer / et al:  meeting?
[4:01pm] dietrich: yep
[4:01pm] thunder: hello!
[4:02pm] sayrer left the chat room. (Ping timeout)
[4:02pm] thunder: I am ready.
[4:02pm] dietrich: should we go around the table, give status?
[4:02pm] sspitzer: I put up our raw notes, from the summit, here:
[4:03pm] sspitzer: (OK, I'll start with status)
[4:03pm] dietrich: thanks for doing that
[4:03pm] thunder: sure
[4:03pm] sspitzer: last week I made a little more progress on
[4:04pm] sspitzer: which is to get the Edit menu / context menu command enabling working with the history-on-places-sidebar
[4:04pm] dietrich: what's the problem w/ the controller code?
[4:05pm] sspitzer: the current problem with the controller code ( is how it does command enabling, vs how the edit menu overlay does it.
[4:06pm] sspitzer: I'm trying to wed the two, but I think I may need a "places edit menu overlay", that dan and I can share (for his bookmarks-on-places sidebar work)
[4:06pm] sspitzer: an example of the problems I have run into is this:
[4:07pm] sspitzer: for commands, the places controller only handles commands that are prefixed with placesCmd_
[4:07pm] sspitzer: see PC_supportsCommand()
[4:07pm] sspitzer: [except for cmd_undo and cmd_redo, but that is another story]
[4:08pm] sspitzer: so I've started to add support for generic commands (like cmd_delete and cmd_selectAll) to the places controller, but it's getting messy.
[4:10pm] sspitzer: the second part of that bug is the C++ changes to handle "delete of history by date", but I think I can re-use the existing nsNavHistoryExpire.cpp code (with a few modifications to generalize it).
[4:10pm] sspitzer: instead of rewriting the queries, as I have started to do.
[4:11pm] dietrich: cool for the re-use
[4:11pm] Archaeopteryx left the chat room. (Quit: ende)
[4:12pm] sspitzer: So, to wrap up my status, I'm working on bug #359462 (but it has a back seat to some Fx 2.0.0.x vista stuff, see bug #354226 for details on that, if you are curious)
[4:12pm]  sspitzer yields the floor to the gentleman from oregon
[4:13pm] thunder:                          
[4:13pm]  dietrich steps forward, trips and crashes into the podium
[4:13pm]  thunder applauds
[4:13pm] dietrich: the first set of schema changes are done
[4:13pm]  myk gasps
[4:13pm] dietrich: renaming etc
[4:13pm] thunder: woo
[4:13pm] dietrich: need to check in once tree is open
[4:14pm] dietrich: --enable-places-bookmarks patch is also done
[4:14pm] dietrich: i'll check that in asap once tree reopens
[4:14pm] dietrich: next step is to make the bookmark schema changes
[4:14pm] dietrich: which is more hairy, should be interesting 
[4:15pm] dietrich: also, as you guys saw, got the framework up for testing history and annos
[4:15pm] dietrich: a few basic tests in there
[4:15pm] dietrich: everyone feel free to add more 
[4:15pm] thunder: yea, kudos on that
[4:15pm] dietrich: i felt quite exposed making all those schema changes with no automated way of catching regressions
[4:16pm] thunder: nod
[4:16pm] dietrich: so now we have a place to start
[4:16pm] dietrich: final status note: i'm leaving on 12/16, until 1/1
[4:16pm] dietrich: w/ little or no net access
[4:16pm] thunder: ok
[4:16pm] thunder: sspitzer: what are your vacation plans?
[4:17pm] sspitzer: dietrich, I have a crazy suggestion regarding all your schema changes
[4:19pm] sspitzer: I was going to suggest leaving some bread crumb comments around any code you haven't tested, and then if, while debugging or writing unit tests, if we happen to touch that code, we can vouch it is good and remove the "XXX test me" comments
[4:20pm] sspitzer: are you worried about all of it, or just parts of the changes that you haven't been able to test yet?
[4:20pm] dietrich: sspitzer: ok, that sounds good, will do
[4:20pm] dietrich: sspitzer: more worried about regressions that won't immediately be noticeable
[4:21pm] dietrich: like if i missed changing a fieldname in a query somewhere
[4:21pm] dietrich: broader unit testing will expose stuff like that, so i'm going to continue to add more tests
[4:22pm] sspitzer: dietrich:  I'm ok with lots of "XXX test me" comments, noisy #ifdef DEBUG printfs, what ever means necessary.  but your patch looked good.
[4:23pm] dietrich: ok, i'll comment aggressively 
[4:23pm]  dietrich passes the stick to thunder
[4:24pm] thunder: woo
[4:24pm] thunder: let's see--
[4:24pm] thunder: I worked on
[4:24pm] thunder: with seth, and that should be all ready
[4:25pm] thunder: I am going to add a test case for the url serialization/deserialization
[4:25pm] thunder: I think I'll try to make a generic one, so we don't need a test for each option that can be serialized
[4:26pm] thunder: I've also been working on
[4:26pm] thunder: with lots of help from seth
[4:26pm]  thunder hugs sspitzer
[4:26pm] thunder: and that is going well
[4:26pm] thunder: I have a basic sidebar now which shows bookmarks from the database
[4:26pm] thunder: but nothing else works yet
[4:27pm] dietrich: cool
[4:27pm] thunder: I hope we don't step on each other
[4:27pm] thunder: when you do the bookmark schema changes
[4:27pm] dietrich: no, i'm 1000 miles away, easy
[4:27pm] dietrich:                          
[4:27pm] thunder:                          
[4:27pm] dietrich: yeah, we'll have to coordinate landings, etc
[4:27pm] thunder: I'm sure it'll be fine, but yeah
[4:27pm] dietrich: i'll look at the patches and we can coordinate
[4:28pm] thunder: cool
[4:28pm] dietrich: sspitzer: is the bookmark service built if --enable-places?
[4:28pm] thunder: also, my plan is to have my work fall under --enable-places
[4:28pm] dietrich: yeah that's what i'm curious about
[4:28pm] thunder: instead of the rdf sidebar
[4:29pm] thunder: the service still gets built, currently
[4:29pm] dietrich: if that's the plan, then it looks like --enable-places-bookmarks is vestigial
[4:29pm] thunder: though presumably that could be disabled
[4:29pm] dietrich: just for exhuming the old places ui?
[4:29pm] thunder: yeah
[4:29pm] thunder: it might be useful for some time
[4:29pm] thunder: though eventually it should go away imo
[4:30pm] dietrich: that begs the question (that vlad already asked, iirc): why add it at all, if we're not going to develop in that direction
[4:31pm] thunder: hmm
[4:31pm] thunder: I'm unsure
[4:31pm] thunder: it might be useful to be able to enable the old ui for comparison
[4:32pm] sspitzer: dietrich:  the RDF bookmarks service is currently build under --enable-places, yes
[4:32pm] sayrer joined the chat room.
[4:32pm] thunder: but whether that's worth landing that patch now is not clear to me
[4:32pm] dietrich: so, i guess one problem w/ doing the bookmark work under --enable-places is that it's hard to do piecemeal implementation
[4:33pm] dietrich: eg: if places becomes enabled by default for the history stuff
[4:33pm] thunder: I probably won't have time to land my bookmark patch before the alpha freeze
[4:33pm] sspitzer: dietrich:  good point, should dan's bookmark work be behind MOZ_PLACES_BOOKMARKS
[4:33pm] thunder: and when I do, it will be roughly equivalent to the history one in functionality
[4:34pm] sspitzer: that seems reasonable, to keep MOZ_PLACES -> history, MOZ_PLACES_BOOKMARKS, bookmarks
[4:34pm] dietrich: sspitzer: i'm having a hard time seeing how it could be just under --enable-places
[4:34pm] thunder: I don't think it's worth it
[4:34pm] thunder: hm
[4:34pm] dietrich: because it can't exist simultaneously w/ old bookmarks ui/backend
[4:34pm] dietrich: we want to incrementally land places by default
[4:34pm] dietrich: first history, then bookmarks
[4:35pm] thunder: you can't have --enable-places-bookmarks without --enable-places
[4:35pm] dietrich: right
[4:35pm] sspitzer: right, so what if he just puts his changes behind MOZ_PLACES_BOOKMARKS
[4:35pm] sspitzer: for what he is working on right now, that would just be some build changes to a couple files
[4:35pm] thunder: I could, but I'd have to gut that stuff first
[4:35pm] thunder: since I don't want half of it
[4:35pm] dietrich: thunder: gutting in --enable-places-bookmarks is a-ok
[4:36pm] dietrich: gut away
[4:36pm] thunder: ok
[4:36pm] thunder:                          
[4:36pm] dietrich: we're definitely in gutting mode
[4:36pm] thunder: then, I really do need your patch to land
[4:36pm] dietrich: i'll land it after the tree reopens
[4:36pm] thunder: so I can do places bookmarks builds
[4:36pm] sspitzer: at least for his bookmarks sidebar, it shouldn't be too ugly.
[4:37pm] dietrich: thunder: u have a patch for the sidebar?
[4:37pm] thunder: not yet
[4:37pm] sspitzer: thunder:  I think you can change your MOZ_PLACES to MOZ_PLACES_BOOKMARKS (in the old and add MOZ_PLACES_BOOKMARKS around your new sidebar in the components/places/content/ (I think) and it should do the right thing.
[4:37pm] thunder: I'd like to get some basic functionality going first
[4:38pm] thunder: right now the only thing that works is showing the bookmarks
[4:38pm] sspitzer: dietrich:  to re-iterate, I think we do want your --enable-moz-places-bookmarks patch
[4:38pm] dietrich: sspitzer: yeah, i thought it might not be necessary, but it is
[4:38pm] thunder: sspitzer: ah, right
[4:39pm] thunder: that makes sense
[4:39pm] thunder: okay
[4:39pm] sspitzer: so that we can enable history-on-places first, before bookmarks (including dan's sidebar) are ready.
[4:39pm] thunder: I'm sold on this plan
[4:39pm] dietrich: we don't want all-or-nothing places. incremental places enabling is key.
[4:39pm] thunder: nod
[4:39pm] sspitzer: agreed.
[4:39pm]  myk likes incremental
[4:40pm] sspitzer: about unit testing, see your inboxen...I now get errors!
[4:40pm] dietrich: ok good 
[4:40pm] thunder: haha
[4:40pm] thunder: ok
[4:40pm] sspitzer: I am not sure why it was passing before.
[4:40pm] sspitzer: it would be nice if tests had a "verbose" mode
[4:40pm] sspitzer: do you think sayrer would be the one to beg for that RFE?
[4:40pm] thunder: oh, about vacation-
[4:41pm] dietrich: yeah, the xpcshell output leaves much to be desired
[4:41pm] dietrich: yes, sayrer is the new test architect, and all RFEs and bugs should be assigned to him
[4:41pm] dietrich:                          
[4:41pm] thunder: I'll be taking a could of days around christmas, and a couple more around mlk jr day (in jan.)
[4:41pm] sayrer: man
[4:41pm] dietrich: hehe
[4:41pm] sayrer: I'll talk with robcee about it
[4:41pm] thunder: haha
[4:42pm] sayrer: definitely file a bug
[4:43pm] dietrich: sspitzer: another change i'm going to make is to only run the hist/anno tests for --enable-places
[4:43pm] dietrich: and add the bookmark/livemark tests when --enable-places-bookmarks
[4:44pm] sspitzer: sayrer:  will do
[4:45pm] sspitzer: sayrer:  about filing a bug on the RFE for being verbose
[4:45pm] sspitzer: dietrich:  that sounds good.
[4:45pm] thunder: yeah
[4:46pm] sayrer: sspitzer: I've had problems interpreting the places tests as well, fwiw
[4:47pm] dietrich: it'd be nice to have output of the xpcshell tests prettified like the mochikit tests
[4:47pm] dietrich: but w/ ascii art instead of html 
[4:49pm] sspitzer: sayrer:  let me fwd you a couple of emails about the unit tests
[4:52pm] thunder: are we done-like?
[4:53pm] dietrich: i think that's a yes
[4:53pm] dietrich: wait, one more thing
[4:53pm] dietrich: were u guys in the gran paradiso mtg yesterday?
[4:54pm] dietrich: history on places is obviously not going in this alpha
[4:54pm] dietrich: so shooting for next in, likely jan 07
[4:54pm] thunder: ah
[4:54pm] thunder: I missed that. okay.
[4:55pm] thunder: was there a concern for the few missing features, or lack of testing?
[4:55pm] thunder: or what?
[4:55pm] dietrich: yeah, this alpha is basically frozen cept for a few specific blockers
[4:56pm] thunder: okay, that sounds reasonable
[4:56pm] dietrich: i wanted to get the schema changes in but didn't really get it done in time for adequate nightly testing before freeze
[4:56pm] thunder: jan 7 it is
[4:56pm] thunder: 07, too
[4:56pm] dietrich: and then seth has a couple of open issues still i think?
[4:56pm] thunder: yes
[4:57pm] dietrich: sspitzer: where are we wrt to feature parity w/ old history?
[4:57pm] dietrich: eg: has anyone run the history tests in litmus against hist-on-places to see where things stand?
[4:59pm] sspitzer: dietrich:  yes, I have some issues, mostly covered by that one bug #359462)
[4:59pm] sspitzer: command issues (including the Edit menu) and delete-by-date
[4:59pm] sspitzer: I'll make sure to run over the tests in litmus after that to make sure we have parity
[5:00pm] sspitzer: once #359462 is fixed, I would not be ashamed to enable it.  (famous last words?)
[5:00pm] dietrich: hehe ok
[5:00pm] thunder:                          
[5:00pm] dietrich: so, i think there's 2 major milestones coming up
[5:00pm] dietrich: 1. fx3 feature PRD next week - we can do hard estimation and planning at that point
[5:01pm] dietrich: we should all make an effort to be involved re places there
[5:01pm] dietrich: and 2. the next alpha (sometime in jan i think)
[5:01pm] dietrich: which hist-on-places should be in
[5:02pm] dietrich: not sure what my point is there, other than to say we should keep those things in mind 
[5:02pm] thunder: by jan, we should have the bookmarks sidebar as well
[5:02pm] thunder: yeah
[5:02pm] thunder: do we want to meet to talk about the fx3 PRD?
[5:03pm] thunder: on the phone sometime, since it's more high-bandwidth
[5:03pm] dietrich: sure. in yesterday's meeting they said beard et al will have it out next week
[5:03pm] dietrich: yeah, phone mtg would be good for that
[5:03pm] thunder: ok
[5:04pm] dietrich: ok, end o' meeting?
[5:06pm] thunder: aye aye, cap'n
[5:06pm] sspitzer: end o meeting for me, dietrich:  thanks for the --enable-places-bookmarks, schema cleanup, and the unit test stuff with sayrer, that gets us going big time
[5:07pm] dietrich: woohoo yeah, and kudos guys on getting sidebars happenin so fast
[5:07pm] dietrich: adios everyone
[5:09pm] thunder: bye!