#jquerymobile-dev_20110517.txt

[07:21:07] <toddparker> Hu all - I've created a doc for us to fill out an agenda for today
[07:21:08] <toddparker> https://docs.google.com/document/d/1b3X8L0gMxA0KwNfoRitb44upkRjVxdX34bqxrAWVBLg/edit#
[07:21:19] <toddparker> stevenblack - i added an item ont he git repo size
[07:21:27] <toddparker> thanks for the research
[07:38:25] <toddparker> stevenblack - is there a ticket for the self referential page issue?
[07:38:35] <toddparker> pls add to good doc
[07:38:48] <StevenBlack> I'll file one.
[07:41:21] <toddparker> i think there may be one, just can't remember the #
[07:43:30] <toddparker> stevenblack - looking at your comment re: commando mode in the doc
[07:43:47] <toddparker> mini does a great job rendering the page and can do all the PE dom manipulations fine
[07:44:35] <toddparker> the various facets of ajax nav are tripping it up IMHO so i don't think we need to try to add the complexity you're suggesting of supporting a light enhancement
[07:44:43] <toddparker> would make the code very complex
[07:44:52] <StevenBlack> ... except when it times-out, in which case it bonks entirely.  So the idea is, minimize the probability of timeout.
[07:45:18] <toddparker> i think that is the result of the ajax loading
[07:45:19] <StevenBlack> Actually, some easy gets here.  Button markup would just no-op, for example.
[07:45:30] <toddparker> but it renders buttons fine
[07:45:47] <toddparker> the real test is to shut off ajax nav on your local copy and browse with mini
[07:45:48] <StevenBlack> Buttons yes.  Fake CSS buttons no.
[07:46:03] <toddparker> i'll do that this mornign to prototype what i'm suggesting
[07:46:10] <StevenBlack> k
[07:50:29] <toddparker> i just remembered that you cant test that way because you can't point to a local IP to test opera mini
[08:06:54] <toddparker> uploading a copy to my server to test
[08:07:18] <StevenBlack> ya, the Opera Mini problem :-)
[08:13:03] <toddparker> I uploaded the git latest with ajax links and forms turned OFF here for testing:
[08:13:03] <toddparker> http://filamentgroup.com/clients/om/
[08:14:18] <StevenBlack> Occurs to me that we have several agenda items that potentially lead to long, long IRC conversations...
[08:14:58] <toddparker> get ready to type
[08:15:10] <toddparker> the nav is just a check in status ont he various parts
[08:15:20] <toddparker> same with ajax imho
[08:15:29] <toddparker> scott is looking for ways to exclude those browsers
[08:15:37] <toddparker> we can talk about your suggestion too
[08:16:18] <StevenBlack> I'm considering removing commando-mode.  Not because it's not a great idea, but because I have low probability of being able to sell it.
[08:16:52] <toddparker> might be good to consider for post 1.0
[08:17:16] <toddparker> i just see that as a lot of complex work to try and figure out how each widget would work in a different way
[08:17:42] <toddparker> btw - opera mini is perfect with ajax off. fast, no blank page or scroll issues
[08:18:14] <toddparker> you even get horizontal page transitions....must be baked into opera!
[08:19:23] <StevenBlack> Love it when a plan comes together.
[08:21:27] <StevenBlack> Is Ghyslain aware of the regular meeting?
[08:35:32] <StevenBlack> Ticket for same-page navigation targets.  ?m1643
[08:35:40] <StevenBlack> ?m1643
[08:35:42] <bot-t> [#1643] Enhancement Request: Navigate to the same page container (open) - https://github.com/jquery/jquery-mobile/issues/1643
[08:37:19] <toddparker> thanks bot-t
[08:37:20] * bot-t sighs
[08:37:40] <StevenBlack> bot-t: when you are AWOL, we miss you.
[08:37:40] <bot-t> StevenBlack, Is that a question?
[08:37:53] <kinblas> StevenBlack: 1643 is your fancy way of saying I want to changePage to the current page? :-)
[08:37:53] <bot-t> (15 hours 1 min ago) <StevenBlack> tell kinblas hi! reworded it to "concern" :-) It's mainly about semantics. Options aren't options anymore once we've augmented them with defaults. They are settings. Secondly, it's weird to pass-in say a 3-key option object to a function and, as a byproduct of the function call, your option object turns into, say, an 8-key object.
[08:38:34] <kinblas> the code doesn't mutate the opt object passed in
[08:38:44] <kinblas> it creates a new one filled with defaults and "overlays" the opts passed in
[08:38:51] <kinblas> I'm just re-using the opts variable name
[08:39:36] <kinblas> that's why I was asking if its just a semantics/naming thing you are proposing
[08:39:43] <kinblas> I'm all for making the code readable
[08:39:55] <kinblas> btw, the code I landed is just where I was at 2 weeks ago
[08:40:06] <kinblas> I'm going to make more sweeping changes today
[08:40:21] <kinblas> like separating page loading from "changePage"
[08:40:28] <kinblas> such that it can be called on its own
[08:40:43] <kinblas> changePage will just be about showing a currently in-memory page
[08:41:03] <kinblas> if changePage is passed a URL of some sort, it will fire off a loadPage and then bail
[08:41:11] <toddparker> kin - great, so you pushed a bunch of changes to the navigation branch?
[08:41:18] <kinblas> after setting a callback on the deferred
[08:41:24] <kinblas> loadPage returns
[08:41:34] <kinblas> this will simplify the code and allow us to put hooks at sensible points
[08:41:52] <toddparker> nice
[08:41:59] <toddparker> (trying ot get scott on irc...)
[08:43:02] <StevenBlack> kinblas: Understood.
[08:43:54] <kinblas> StevenBlack: so back to 1643
[08:43:56] <StevenBlack> The defaults - options - settings thing is well-covered in some of the great plugin authoring articles.  1 sec and I'll get a link or two.
[08:44:06] <kinblas> if folks are updating a container repeatedly
[08:44:19] <kinblas> why aren't they doing it in a container in the same page?
[08:44:36] <kinblas> why do they have to "changePage" to the same container? Which probably won't look nice transition wise
[08:44:42] <kinblas> Or is it just to get the hash to update?
[08:45:26] <scottjehl> hey
[08:45:52] * kinblas turns of the clicky sound on his blackberry
[08:45:54] <StevenBlack> kinblas: lots of reasons.  The tactile nature of mobile means a swipe is a-propos, for example.  But there's a bigger more general reason.
[08:46:20] <StevenBlack> Great things happen when structures are self referrent.
[08:46:24] <StevenBlack> (always)
[08:46:29] <scottjehl> kin: HAH
[08:47:07] <kinblas> StevenBlack: I guess what I'm trying to understand in your swipe example is why isn't the page you are swiping to a "page" itself?
[08:47:14] <kinblas> *OR*
[08:47:27] <kinblas> if it's just swiping between content in an existing "page"
[08:47:36] <kinblas> why does changePage() have to be involved?
[08:47:39] <StevenBlack> I need a blackboard, not a bulletin board :-)
[08:48:08] * kinblas checks out the whiteboard functionality in connect
[08:48:51] <kinblas> dougneiner: did you ever do a pull request for that performance tweak? If not I can just land a patch and credit you in the checkin comments.
[08:49:52] <StevenBlack> Well, on pure principle, looking at it at the street-gutter level, the fact that changePage() hard-codes the denial is contemptible.  I think we can probably agree on that.  Just sayin'...
[08:50:37] <kinblas> I wouldn't say contemptible, it's just tryin' to stop something the original code didn't handle
[08:51:07] * kinblas notes his vocabulary is expanding due to StevenBlack
[08:51:08] <kinblas> :-)
[08:51:28] <StevenBlack> :-)    changePage() currently assumes that we want, in all conditions, to do only A -> B transitions.
[08:51:42] <kinblas> right, hence the name
[08:52:23] <StevenBlack> This leads to practical problems, like A -> B -> C -> D is four different containers in the DOM.
[08:52:39] <StevenBlack> Practical problem #2 is no id attribute can be repeated.
[08:53:05] <StevenBlack> THEREFORE some pretty slick and generic template things are out of the question.
[08:53:30] <kinblas> because the templates use ids?
[08:54:03] <StevenBlack> Because smart jQuery uses IDs.
[08:54:32] <kinblas> I need to pause this conversation while I pick up someone for an interview downstairs. :-)
[08:54:46] <kinblas> I'll be back in a bit
[08:55:01] <StevenBlack> Cool.  No problem I just want to encourage thinking outside-the-box while we still can.
[08:59:36] <toddparker> I
[08:59:50] <toddparker> I've been filling out the agenda. Lots to review today: https://docs.google.com/document/d/1b3X8L0gMxA0KwNfoRitb44upkRjVxdX34bqxrAWVBLg/edit#
[09:00:30] <StevenBlack> I've come to see, building data-driven jQuery Mobile apps, since data has structure, it cries-out for structurally consistent display.  We want to do that WITHIN jQuery Mobile's page-centric abstractions.  Especially on smaller devices.  Flipping a page in a data-driven site is often to display another record fleshed on the same skeleton.  It's optically a new page.  But behind the scenes, we're constrained by current de
[09:00:30] <StevenBlack> from reusing the current container.
[09:10:47] * kinblas is back
[09:11:19] <kinblas> StevenBlack: So "flipping" between pages ... why use a single page to do all that versus really flipping between 2 different pages?
[09:11:40] <kinblas> fyi, I'm not questioning architecture/implementation just trying to understand the use case
[09:13:18] <StevenBlack> Because you need to build a whole new page to re-display everything even though only a few elements vary from page to page.
[09:13:59] <StevenBlack> And NOW you need to test to see if the page is in cache because back and forth, you don't want to dupe pages.
[09:15:23] <StevenBlack> Many possible targets === code to track these targets.  ID's now out of the question too.
[09:16:13] <kinblas> So a couple of questions
[09:16:21] <StevenBlack> Tangentialy, if we *really* isolate transitions -- see the agenda -- then we could do this slickly circumventing changePage altogether.
[09:17:39] <StevenBlack> Good exposé of this here, BTW: https://developers.jivesoftware.com/community/blogs/engineering/2011/04/15/jivin-with-jquery-mobile-inter-page-navigation
[09:20:27] <toddparker> now that we have someone from jive helping out, maybe this is a topic he can weigh in on
[09:20:34] <toddparker> (offline now)
[09:21:05] <toddparker> Restore user zooming capability - https://github.com/jquery/jquery-mobile/issues/1645
[09:21:36] <toddparker> Scott and I want to get rid of our zoom meta tags. when he mentioned this at mobilism, people were very happy about the idea
[09:21:44] * kinblas reads the jive article
[09:21:57] <StevenBlack> It might be wise to defer this item until Ghyslain is available.  That would unclutter today's agenda, and give the topic its due separately.
[09:22:05] <kinblas> heh they used an URL scheme like I started in the pagecontainer branch
[09:22:29] <toddparker> really? wonder if they were inspired or just has the same idea
[09:22:42] <toddparker> yeah, we have a full boat today
[09:22:55] <toddparker> in case there is any confusion out there in the world: jQuery is about mobile *web* sites and apps
[09:22:56] <kinblas> The scheme was just to not use top-level URLs in the hash but use name/value pairs
[09:23:07] <kinblas> that way arbitrary pageCotainers could load content independently
[09:23:14] <kinblas> and have data associated with each URL
[09:23:29] <toddparker> we're making some changes like the auto back button, user zoom, etc. to better align ourselves with web
[09:23:42] <toddparker> i'll shut up now :)
[09:24:53] <kinblas> StevenBlack: Trust me when I say I get what you guys are describing here. What I'm trying to understand is what folks expect ... transitioning from A to A doesn't really work well unless you transition out, and then transition in
[09:25:37] <kinblas> Also, this sounds very advanced, in terms of users that will be able to take advantage of all the hooks we're going to provide to allow someone to build something like this
[09:26:10] <toddparker> is A > A just a case where you want to update parts or the whole current page and call the page-level enhancements on the new markup?
[09:27:10] <toddparker> i've been saying that having essentially a refresh/enhance method that you can call on any target (page, form, selector) would make life easier
[09:27:28] <toddparker> that is the kind of simple, generic tool i'd be in favor of
[09:28:53] * kinblas agrees with toddparker on the enhance method
[09:28:57] <StevenBlack> toddparker: +1
[09:29:11] <toddparker> woo hoo, i suggested a technical thingie that makes sense
[09:29:11] * kinblas needs to learn to use +1 to save typing
[09:29:32] <kinblas> heh
[09:29:38] <kinblas> there's an issue on that
[09:29:45] <toddparker> steven, does that single method cover the basics of what you're asking for?
[09:30:16] <kinblas> toddparker: its more than that
[09:30:24] <kinblas> it requires hooks in several places
[09:30:34] <toddparker> sure
[09:30:43] <toddparker> I see this issue, but closed: https://github.com/jquery/jquery-mobile/issues/435
[09:31:10] <kinblas> to interpret/set the hash appropriately, place to process what gets loaded off the server to potentially genreate HTML from data, etc
[09:31:14] <toddparker> I guess stevenblacks's ticket is the best one to use anyway: https://github.com/jquery/jquery-mobile/issues/1643
[09:32:21] <toddparker> should i add this to navigation re-factor checklist?
[09:32:50] <toddparker> ah, already there it the agenda, just added ticket link
[09:35:54] <StevenBlack> Ya, the scenario is, you have a bunch of JSON, and you want to allow the user to 1) step horizontally through individual records, and/or 2) vertically between parents and child collections, and the kicker is you want to do it cleanly.  There may be no URL changing at all, in fact.  Swipe -> Transition out -> generate markup -> Transition in.  Same container.
[09:36:39] <toddparker> sure, makes sense. so you might want to use a JS template engine to build a page, enhance it, then show
[09:38:29] <StevenBlack> In fact, let me nail it down further: no URL change.  You've got JSON, it's now a matter if navigating and displaying the data you've got.
[09:38:43] <StevenBlack> if = of
[09:40:19] <StevenBlack> Um, take that back.  No change to the "core" URL but being able to use a chunk of the URL for data passing between pages (see agenda), that would be sweet.
[09:44:27] <StevenBlack> =======
[09:51:04] <StevenBlack> Mind=blown if you think of a generic changeContainer() whereby now data-role="page" and its contents data-role="content" certainly, but perhaps "header" and "footer" too, have multiple cardinality, with just one displayed at any given moment.
[09:51:06] <StevenBlack> In short, a page with multiple data-role="content" containers, one visible, the others cached.
[09:52:20] * StevenBlack wonders if you can css-3 transition just part of a page.
[10:21:07] <scottjehl> of course - add our classes to any element you want
[10:55:40] <kinblas> hmmm I can't join #jquery-meeting
[10:55:54] <kinblas> is it just me?
[10:56:15] <_nickel> kinblas: I think it is
[10:56:21] <kinblas> grr
[10:56:28] <kinblas> ok re-connecting then
[10:57:10] <kinblas> ok so what does this mean?
[10:57:11] <kinblas> Cannot join channel (+r) - you need to be identified with services
[10:59:47] <toddparker> hey all
[11:00:02] <toddparker> let me unmoderate that channel for ya Kin
[11:01:09] <kinblas> toddparker: let me know when you've done that
[11:01:11] <toddparker> can you see if i unmoderated right over on the other chan
[11:01:16] <toddparker> did it
[11:01:27] <kinblas> grr no dice
[11:01:42] <toddparker> hrm
[11:01:52] <kinblas> I'm not sure if it's a firewall thing or what ... they had to make an exception for me
[11:02:11] <kinblas> hmmm but I can get here
[11:02:19] <kinblas> so is there something special about that room?
[11:02:29] <toddparker> nope
[11:02:35] <toddparker> public, unmoderated
[11:02:39] <toddparker> back to the web ui?
[11:02:40] <kinblas> did you register your nick with freenode?
[11:02:45] <toddparker> me?
[11:02:49] <toddparker> think so
[11:03:30] <kinblas> what's the URL?
[11:04:52] <toddparker> for what?
[11:05:01] <toddparker> jquery-mobile
[11:05:21] <gseguin_> Hi Todd, are we meeting today?
[11:05:44] <toddparker> yes, on #jquery-meeting
[11:06:44] <gseguin_> I get  a: "Cannot join channel (+r) - you need to be identified with services"
[11:06:52] <scottjehl> same
[11:07:41] <kinblas> wow
[11:07:47] <kinblas> even the web interface denies me
[11:07:49] <kinblas> wtf?
[11:07:56] <toddparker> so odd
[11:08:03] <toddparker> because you were both in last week
[11:08:07] <kinblas> yeah
[11:08:16] <toddparker> i'm an irc n00b
[11:08:24] <toddparker> any ideas on how to get you in?
[11:08:24] <gseguin_> same
[11:08:28] <toddparker> hrm
[11:09:36] <gseguin_> why not meeting on #jquerymobile-dev since we seem to all be here
[11:09:48] <toddparker> kinblas and gseguin_ are you registered with nickserv?
[11:09:58] <toddparker> scott's in now that he did that
[11:10:15] <gseguin_> no, let me find out how to and do it
[11:10:17] <kinblas> toddparker: nope
[11:10:19] <kinblas> hang on doing it
[11:10:23] <toddparker> gseguin_ that's just where we do all jquery meetings
[11:10:35] <scottjehl>  guys you need to /msg NickServ  REGISTER password [email]
[11:10:40] <toddparker> ...can you register??
[11:10:54] <scottjehl> then check email, paste the followup
[11:12:32] <gseguin_> alright, I'm in
[11:13:56] * kinblas waits patiently for email from freenode.net
[11:14:21] <toddparker> spam filtered?
[11:14:39] <kinblas> nope its empty
[11:15:36] <toddparker> hrm
[11:15:53] <kinblas> so someone changed the policy on the room eh?
[11:15:54] <toddparker> maybe try again? different email?
[11:17:46] <toddparker> dunno
[11:17:54] <toddparker> tell me how to change it back!
[11:18:29] <kinblas_> -r? :-)
[11:18:38] <toddparker> ?
[11:18:46] <toddparker> looks like you're named
[11:18:57] <toddparker> you still can't join?
[11:19:23] <toddparker> you all set?
[11:19:28] <kinblas> it tells me I need to follow instructions on an email
[11:19:32] <kinblas> which I haven't gotten yet
[11:19:42] <toddparker> something is wrong
[11:19:52] <toddparker> the other guys got theirs right away
[11:25:44] <toddparker> hey paul_irish
[11:27:25] <toddparker> hold on kin talking to paul on skype
[11:28:55] <toddparker> kinblas - try getting in now
[11:28:56] <paul_irish> _kinblas: try now?
[11:29:24] <paul_irish> \o/
[11:29:27] <_kinblas> paul_irish: what did you do? Remove -r?
[12:49:42] <toddparker> hey all
[12:49:58] <scottjehl> HEY
[12:50:50] <toddparker> stop yellin'
[12:51:01] <toddparker> anyone want to look at this: https://github.com/jquery/jquery-mobile/issues/1512#comment_1192308
[12:51:45] <kinblas> toddparker: Ah hah, on Android it is only available with no scale!
[12:51:55] <scottjehl> BOO
[12:52:05] <scottjehl> not worth it
[12:52:05] <kinblas> heh, I chuckled when I re-read that
[12:52:16] <toddparker> WUT?
[12:52:18] <kinblas> zoom == no fixed pos
[12:52:36] <scottjehl> these are decisions we mortals should not be left to make
[12:52:47] <kinblas> just as an FYI
[12:52:50] <kinblas> you have to use this:
[12:52:52] <kinblas> <meta name=”viewport” content=” width=device-width, user-scalable=no">
[12:52:59] <toddparker> so if you shut off the min/max scale=1, toolbars fail?
[12:53:02] <scottjehl> yeah no
[12:53:05] <toddparker> not helpful
[12:53:09] <kinblas> user-scalable=no is not equivalent to min/max=1
[12:53:16] <kinblas> at least in their interpretation
[12:53:17] <toddparker> right
[12:53:19] <toddparker> = NO
[12:53:39] <scottjehl> we're fine without fixed positioning almost all the time anyway
[12:53:48] <toddparker> so a further qualification for fixed toolbars
[12:53:57] <scottjehl> there's a feature test
[12:54:04] <scottjehl> if we ever really want to go that route
[12:54:09] <toddparker> for zoom?
[12:54:15] <_nickel> toddparker: scottjehl: kinblas: late to the party, but I'm glad to hear that disabling ajax by default on those platforms will fix a lot of issues
[12:54:48] <scottjehl> yeah, it's reactive based on what we know, but better than doing nothing
[12:55:28] <toddparker> simple fix too
[13:06:44] <toddparker> go team!
[14:45:44] <toddparker> kinblas - ideas on why you can't seem to blur out of an input in iOS?
[14:45:45] <toddparker> https://github.com/jquery/jquery-mobile/issues/1647
[14:46:11] <kinblas> nope
[14:46:21] <kinblas> we're not killing any events
[14:49:46] <kinblas> I can repro it on a page with no jquerymobile
[15:07:47] <_nickel> kinblas: scottjehl: I realize that you guys haven't had much time to look at it but we've got a pretty good set of failing tests
[15:07:59] <_nickel> good sized set
[15:08:53] <kinblas> _nickel: strange, I checked the tests on Friday
[15:09:00] <kinblas> aside from the usual navigation and that checkbox one
[15:09:02] <kinblas> others were passing
[15:09:09] <scottjehl> yeah just noticed
[15:09:11] <scottjehl> one thing
[15:09:26] <scottjehl> I think the jqmData et al. are failing due to changes in attr and prop
[15:09:31] <scottjehl> we need to get on that
[15:09:51] <kinblas> ugh
[15:12:16] <toddparker> is that with 6.1?
[15:12:27] <toddparker> thought that was backtracked a bit
[15:12:48] <scottjehl> MonkeyPatch, MonkeyPoop
[15:12:52] <toddparker> oh _nickel, we miss you!
[15:13:54] <kinblas> did someone just change 100+ files?
[15:13:55] <kinblas> :-)
[15:14:08] <toddparker> blame it on scottjehl
[15:14:21] <toddparker> ...hey, you wanted him back in helping right?
[15:14:30] <toddparker> now that's productivity!
[15:14:32] <kinblas> heh
[16:03:21] <_nickel> toddparker: yah I'm not enjoying being out of the loop at all
[16:03:40] <_nickel> I'm working as hard as I can to get stuff off my plate at work
[16:03:51] <_nickel> but we are completely innundated with projects :/
[16:20:04] <_nickel> at least the test failures are consistent between the two browsers :D