Log for #mozilla: 2004-02-10
----------------------------
00:52 < fantasai> bernd?
00:52 < bernd> hi
00:53 * bz watches the tables task force gather
00:53 -!- MrMazda [~chatzilla@B2a-CommercialWay-Brooksville-24.ij.net] has quit [Quit: ChatZilla 0.9.59 [Mozilla rv:1.7a/20040207]]
00:54 <@bz> http://www.editions-eyrolles.com/livres/glazman/tests/listes/listes15.htm
00:55 < fantasai> bernd: in 229883, you cleaned up some of the bc functions
00:55 < fantasai> bernd: but I still don't know what to do about the calcdominantborder calls I make in 4510
00:55 <@bz> Does someone here have a system with not enough fonts?
00:56 < fantasai> bernd: the ones I marked //xxxfr why true?
00:56 < luser> probably will in a minute, my fresh debian system
00:56 < fantasai> bernd: I did that because I didn't know what the boolean arguments meant
00:56 <@bz> fantasai: bernd's patch documents them a bit better, too...
00:56 < fantasai> bz: yeah, I've been looking at it :)
00:56 < bsmedberg> bz: I get some ?? on my win32 box
00:57 <@bz> bsmedberg: ok
00:57 < fantasai> I remember finding a lot of problems with that test suite
00:57 <@bz> bsmedberg: see bug 14713
00:57 * fantasai wonders if glazou fixed them
00:57 <@bz> bsmedberg: opinions?
00:57 <@bz> fantasai: that test renders perfectly over here.... so I can't test the bug filed on what happens when it doesn't
00:58 < bsmedberg> bz: hrm... good question. This only affects ordered lists, right?
00:58 <@bz> yeah
00:58 < fantasai> bz: oh, I meant the suite in general, not that specific test
00:58 <@bz> with bizarre list-style-type
00:58 < bsmedberg> bz: other than ?, the only other fallback I can see would be arabic
00:58 <@bz> fantasai: I understood. ;)
00:58 <@bz> bsmedberg: exactly.
00:59 < fantasai> if it's the one I remember, that is...
00:59 <@bz> bsmedberg: the question is whether that's something worth attempting to do
00:59 < bsmedberg> bz: I'm not a spec-God, but I don't think our current solution is all that bad
00:59 <@bz> bsmedberg: it's not a spec question but a user-experience question. ;)
00:59 < bsmedberg> RESOLVED GOODENOUGH
00:59 * bz notes that the spec assumes all chars will render
00:59 < bernd> fantasai: I added the revised patch to the bug, but it still lacks testing that I didnt regress something
01:00 -!- Hendikins|DomesticatedWolfox is now known as Hendikins
01:01 < bernd> fantasai: but it is better documented what the function arguments are
01:01 -!- Wolf [~Wolf@clt56-112-096.carolina.rr.com] has quit [Quit: ChatZilla 0.9.58p [Mozilla rv:1.7a/20040131]]
01:02 < fantasai> bernd: I'm looking at the documentation for static BCCellBorder CalcDom..
01:02 < fantasai> bernd: your sentence, I mean
01:02 < fantasai> bernd: I can't make sense of it
01:02 < fantasai> "At the table edges..."
01:02 < fantasai> that one
01:02 <@bz> that's a review comment. ;)
01:03 -!- fire_fire [ffdsfds@FSH200159.resnet.ucsb.edu] has joined #mozilla
01:03 < fantasai> bz: heh
01:03 < fantasai> bz: except I need to make sense of it soon ;)
01:04 < fantasai> bernd, bz: btw, I have most of the changes to nsTablePainter done
01:04 < fantasai> I can put them up tonight
01:04 < bernd> fantasai: At the table edges the rules should be ignored as they are only inner borders. (the period is missing in the comment)
01:04 < fantasai> what rules?
01:04 < bernd> these html thingies
01:04 < fantasai> 'rule'?
01:05 < fantasai> the attribute?
01:05 < bernd> yep
01:05 < bernd> yeah
01:05 < fantasai> ok.. why is that handled here?
01:05 < fantasai> shouldn't it be in ua.css?
01:05 <@bz> fantasai: I meant, I am adding "make fantasai happy with comments" to my review comments
01:05 <@bz> not quite
01:05 <@bz> because there is no good way to handle it in ua.css....
01:06 < bernd> fantasai: it will not work the border depends on the position of the cell
01:06 <@bz> rules="cols" eg
01:06 < fantasai> ....
01:06 -!- lucifer_ [~lucifer@pcp02797998pcs.goosck01.sc.comcast.net] has joined #mozilla
01:06 < fantasai> I'm quite sure I wrote up the CSS for that years ago
01:06 * bernd says only one number 915
01:06 < lucifer_> is there a way to make moz read from stdin?
01:07 < fantasai> bernd: it should have nothing to do with 915
01:07 < fantasai> bernd: 915 deals with non-CSS behavior
01:07 < fantasai> bernd: this is CSS-describable behavior, if I'm not mistaken
01:07 * fantasai goes to dig up old comments
01:08 < bernd> fantasai: as the borders depend from the position of the cell in the cellmap the underlaying problem of bug 915 hits you
01:08 < bernd> hixie did a good write up of the architecture problem somewhere
01:08 < fantasai> bernd: why does it depend on the position of the cell in the cellmap?
01:09 < lindi-> lucifer_: date | mozilla /dev/stdin works here
01:09 -!- yason [~yason@addr-212-50-149-209.suomi.net] has quit [Ping timeout: 365 seconds]
01:09 < fantasai> http://bugzilla.mozilla.org/attachment.cgi?id=10217&action=view
01:09 < bernd> fantasai: how do you decide that a cell belongs to a column via the cellmap
01:09 <@bz> lucifer_: and do what with it?
01:10 < fantasai> bernd: how is "rules=cols" different from
01:10 < lucifer_> bz: open a page
01:10 < fantasai> col {border: 1px solid}
01:10 < fantasai> ?
01:10 <@bz> lucifer_: where?
01:10 <@bz> lucifer_: what if mozilla is already running?
01:10 < lucifer_> bz: yes
01:10 < lucifer_> lindi-: thats a good idea
01:10 <@bz> lucifer_: that wasn't a yes-or-no question.....
01:12 < lucifer_> bz: what if?
01:13 < ajschult> lucifer_: yes
01:13 < lucifer_> lindi-: doesn't show anything
01:13 < bernd> fantasai: do you have an understanding how much code we could remove with http://bugzilla.mozilla.org/attachment.cgi?id=10217&action=view ?
01:14 <@bz> lucifer_: what should mozilla do if it's already running and you do 'foo | mozilla'?
01:14 < fantasai> bernd: no
01:14 < lucifer_> bz: i dunno
01:14 < bernd> fantasai: > 500
01:14 < fantasai> bernd: I thought that we /did/ remove that code
01:15 < fantasai> bernd: Do you mean to say we've got a ton of rule/frame code tangled into BC because no-one paid attention to that attachment three years ago?!
01:15 <@bz> lucifer_: well, that's the problem with implementing it. Have to decide what to do in various cases like that.
01:15 < bernd> fantasai: exactly
01:15 < fantasai> aaahhh!!!!
01:16 < bernd> http://lxr.mozilla.org/seamonkey/source/content/html/style/src/nsHTMLStyleSheet.cpp#214
01:16 < fantasai> this is so .. exasperating
01:16 < lucifer_> bz: its a one time thing, its not for an app
01:16 < lucifer_> bz: i just want to pipe some dynamic text to mozilla
01:16 < bernd> especially: http://lxr.mozilla.org/seamonkey/source/content/html/style/src/nsHTMLStyleSheet.cpp#247
01:17 <@bz> lucifer_: mozilla 'data:text/html,whatever`
01:18 <@bz> er, forward quotes
01:18 -!- Sander [~abuse@203.97.90.154] has joined #mozilla
01:18 <@bz> can basically prepend 'data:text/html,' and use backquotes on the program output....
01:18 < fantasai> bernd: the one problem is with the cascade
01:18 < fantasai> bernd: the rules need to exist at the HTML preshint level
01:18 -!- Kovu [~chatzilla@pcp983794pcs.northw01.in.comcast.net] has quit [Ping timeout: 365 seconds]
01:19 < fantasai> and we don't have a CSS stylesheet for that
01:19 < fantasai> just that class
01:19 <@bz> how would you tell?
01:19 <@bz> wait.
01:19 <@bz> So I guess the point is that interaction with user rules is what worries us?
01:20 < fantasai> yeah
01:20 <@bz> We have several options
01:20 <@bz> 1) ignore the problem
01:20 < fantasai> if we had a preshint level, I wouldn't have gotten into so many arguments with hixie and dbaron over it, I'm sure
01:20 < bernd> sounds like a decent option
01:20 <@bz> It's already questionable that <b> is a ua-level thing while <font size="4"> is a preshint
01:21 <@bz> 2) Implement a stylesheet that mozilla loads like ua.css but cascades at the document level right before the attribute stylesheet
01:21 < fantasai> that's what I've been arguing the whole time
01:21 <@dbaron> Adding a preshint level would be easy.
01:21 * bernd would like to remove all of that rule handling
01:21 -!- yason [~yason@addr-212-50-149-209.suomi.net] has joined #mozilla
01:21 <@bz> Like dbaron said, #2 is pretty easy
01:21 < bernd> from the c++
01:21 <@bz> especially with bryner's nsStyleSet changes
01:21 < fantasai> I say we do #1 first
01:21 <@dbaron> bernd: but note that adding rules like those can slow things down
01:22 < fantasai> and then do #2
01:22 <@bz> We've been doing #1 for ....
01:22 <@dbaron> and it's also worth noting that the current rules code is much messier than it should be and broken in many many ways
01:22 < fantasai> dbaron: the rules would be much simpler with BC
01:22 * bz checks the filing date on bug 2942
01:22 <@dbaron> and is nowhere close to the preshint level
01:22 <@bz> 1999-02-04 21:06 PST
01:22 < fantasai> dbaron: those were written before BC
01:22 < fantasai> uhhh
01:22 < fantasai> before border collapse rewrite
01:23 <@dbaron> we've done BC for a long time
01:23 <@dbaron> see bug 3000
01:23 < fantasai> it was disabled for a good while while karnaze rewrote it
01:23 < bernd> it was enabled with bug arounf 41900
01:23 < fantasai> and /that/ was when I wrote that attachment
01:24 < fantasai> like, the first year I had a Bugzilla account...
01:24 < fantasai> ok, ok
01:24 < fantasai> so
01:24 < fantasai> we have several things to deal with
01:24 < fantasai> a) 4510
01:24 < fantasai> b) Bernd's BC reorg
01:25 < fantasai> c) getting all that rules complexity into some CSS
01:25 < bernd> the bc checkin was on Feb 19 2002
01:25 < fantasai> bernd, do you want to reopen 43178?
01:26 < bernd> fantasai: its not a reorg its just a little bit easier to read
01:27 < bernd> fantasai: http://bugzilla.mozilla.org/show_bug.cgi?id=43178#c1 :-)
01:27 < fantasai> bernd: you're the one who named the patch "reorg" ;)
01:27 < fantasai> bernd: *grin*
01:27 < bernd> it reorgs a little part of 2300 lines of code
01:27 < fantasai> yeah
01:28 < bernd> fantasai: does bug 4510 depend on the rules issue ?
01:29 < fantasai> bernd: it's affected by it
01:29 -!- lordpixel is now known as lp_sleep
01:29 < fantasai> bernd: because of the calcdominantborder calls
01:30 < fantasai> which you wanted to reorganize
01:30 < fantasai> and whose reorganization, I'm guessing, will now depend on the rules issue
01:30 < fantasai> 4510 could go in without the reorg
01:30 < fantasai> the calls it makes don't affect BC, just the data I'm stashing to use for background painting
01:31 -!- ssieb_home [~ssieb_hom@24.80.50.208] has quit [Quit: ChatZilla 0.9.59 [Mozilla rv:1.7a/20040201]]
01:31 -!- ajschult [~ajschult@martin.che.ncsu.edu] has quit [Quit: no reason]
01:31 < bernd> then go ahead with your patch, adapting your patch to the "reorg" is a nobrainer
01:31 < fantasai> ok
01:32 -!- ssieb_home [~ssieb_hom@24.80.50.208] has joined #mozilla
01:33 < fantasai> bernd: do you want to reopen 43178 or should I?
01:33 < bernd> dbaron: your margin testcase is good enough, the table should not overflow period
01:33 -!- Dauphin [~chatzilla@ip68-6-227-247.sd.sd.cox.net] has quit [Quit: ChatZilla 0.9.59 [Mozilla rv:1.7a/20040208]]
01:33 -!- kherron [~kherron@c-24-7-164-81.client.comcast.net] has quit [Quit: Leaving]
01:33 < bernd> fantasai: its your idea, so please reopen it
01:34 -!- Dauphin [~cls@ip68-6-227-247.sd.sd.cox.net] has joined #mozilla
01:34 < bernd> fantasai: do you know why the col group case does not work
01:35 -!- caillon_work [caillon@gw.office.mozilla.org] has quit [Quit: Faith in Chaos]
01:35 < fantasai> bernd: bernd@gmx.de?
01:35 < fantasai> bernd_mozilla, right?
01:35 < bernd> bernd_mozilla@gmx.de
01:35 < fantasai> bz: do you want to be CCed, too, while I'm at it?
01:36 * dbaron realizes we should probably make scrollbars switch sides on RTL documents
01:36 <@dbaron> I think I've argued against that in the past, though.
01:37 < fantasai> dbaron: yes, it's inconsistent
01:37 < fantasai> bernd: groups doesn't work because I didn't design for BC
01:38 <@dbaron> the problem is that for RTL documents, we want the bottom scrollbars to allow scrolling left but not right, and that doesn't fit with a vertical scrollbar on the right
01:38 < fantasai> ?
01:38 < fantasai> bernd: all the borders there are assigned on the cells
01:39 < fantasai> bernd: but with BC, we can just assign on <col>, <colgroup> etc.
01:40 < fantasai> dbaron: imagine an Arabic + English document
01:40 < fantasai> dbaron: the scrollbars would be on both sides
01:40 <@dbaron> fantasai: I just can't imagine any good UI for having a horizontal scrollbar that allows scrolling to both sides of the initial position.
01:41 <@dbaron> fantasai: the type of scrollbars would be determined by the direction of the root element
01:41 < fantasai> dbaron: if I have two very wide tables in my document
01:42 < fantasai> dbaron: one rtl
01:42 < fantasai> dbaron: one ltr
01:42 <@dbaron> fantasai: then you'll be no worse off than you are today
01:42 < fantasai> dbaron: do you mean to say you're just going to cut off one of them?
01:42 <@dbaron> fantasai: yes
01:43 <@bz> fantasai: yes, please cc me on anything involving code removal in layout. ;)
01:43 < fantasai> ^o^
01:44 * dbaron wonders what "^o^" is (other than looking a lot like a symbol for a bat)
01:44 <@bz> mouth wide open with surprise?
01:45 <@bz> eyes slitted up in a dangerous way?
01:45 < fantasai> bz: honestly, I'm not entirely sure what it pictographically represents
01:45 <@dbaron> it still looks like a bat to me
01:45 < fantasai> bz: I know it's usage, though
01:45 -!- basic [~basic@219.95.191.83] has joined #mozilla
01:45 * fantasai picked it up from some artistic friends
01:46 <@bz> looks like diethyl ester, actually
01:46 < fantasai> lol
01:46 <@bz> the "bird of hope"
01:46 < fantasai> it's derived from ^_^
01:46 * bz has fond memories of organic chemistry
01:46 * fantasai grins
01:46 < fantasai> I'm sure a lot of people don't share the sentiment
01:46 * dbaron thinks diethyl ester requires O rather than o
01:47 <@bz> dbaron: so.... would you be willing to remove the aPresContext argument of GetUniqueStyleData as part of this patch?
01:47 <@dbaron> bz: sure, if I don't need another round of revie
01:48 <@bz> dbaron: I think I can trust you to do it, yes.. ;)
01:49 < fantasai> hmm
01:49 * fantasai thinks the o is a nose
01:49 < fantasai> anyhow
01:50 * fantasai has to draft an email to what will no doubt be a very exasperated CSS WG
01:50 < bernd> fantasai: too late isnt it ?
01:51 < fantasai> bernd: better late than never, ne?
01:51 <@dbaron> nsIPresContext* PresContext() const { return mRuleNode->GetPresContext(); }
01:51 <@dbaron> on nsStyleContext?
01:51 <@dbaron> bz: ok with you?
01:51 <@bz> dbaron: sounds good to me.
01:52 <@bz> fantasai: what're you mailing about?
01:52 < fantasai> bz: tables
01:52 < fantasai> :)
01:52 * fantasai ran into a few things while working on 4510
01:52 <@bz> heh
01:53 <@dbaron> bz: new patch attached, if you want to see it
01:54 <@bz> dbaron: thanks
01:55 <@bz> dbaron: looks great
01:56 <@bz> argh
01:56 <@bz> the testcase in that rules bug doesn't provide the css for rules=groups....
01:56 < fantasai> bz: the CSS would need to be rewritten
01:56 < fantasai> bz: so it takes advantage of the BC model
01:56 < fantasai> bz: you can'd do groups without BC
01:56 -!- Hendy [~wolfox@DC-136-132.bpb.bigpond.com] has joined #mozilla
01:56 < fantasai> bz: it would be
01:57 <@bz> hmm
01:57 < fantasai> thead, tbody, tfoot, colgroup {border: 1px solid}
01:57 < fantasai> or somesuch
01:57 < fantasai> with a table[rules=groups] >
01:57 <@bz> with rules=cols, shouldn't the rules be on colgroup boundaries?
01:57 < fantasai> of course
01:57 <@bz> not on column boundaries?
01:57 <@bz> or am I confused?
01:57 < fantasai> confused, methinks :)
01:57 < fantasai> it's analogous to rows
01:57 * bz checks
01:57 < fantasai> yeah, that one would be changed to
01:58 < fantasai> table[rules=cols] col {border: 1px solid}
01:58 < fantasai> i think
01:58 <@bz> oh, I see
01:58 < fantasai> in any case, it needs a revision
01:58 <@bz> you can't do just col groups
01:58 <@bz> fantasai: there is no "col" in there....
01:58 <@bz> fantasai: is there?
01:58 < fantasai> hmm
01:58 <@bz> fantasai: nothing to match that CSS rule
01:59 < fantasai> yeah, I guess you'd have to do it the same as I did :)
01:59 <@bz> we stick in a tbody
01:59 <@bz> in HTML
01:59 <@bz> but....
01:59 -!- Tukon [~Tukon@adsl-64-161-201-181.dsl.snfc21.pacbell.net] has quit [Read error: Connection reset by peer]
01:59 <@bz> so does setting border on colgroup work?
01:59 <@bz> or rowgroup?
01:59 < fantasai> yep
01:59 < fantasai> oh, we do it better than any other browser i'm aware of
01:59 <@bz> s/rowgroup/tbody/
02:00 <@bz> so rules=groups should work, no?
02:00 < fantasai> Part C of my backgrounds test suite is a nice demonstration
02:00 < fantasai> IE chokes on it
02:00 < fantasai> yes
02:00 <@bz> although....
02:00 <@bz> <table>
02:00 -!- kerz [kerz@craptastic.org] has quit []
02:00 <@bz> <colgroup span="2" /><col /><colgroup span="2">
02:00 * bz thinks
02:00 <@bz> so
02:00 -!- Hendikins [~wolfox@DC-129-223.bpb.bigpond.com] has quit [Ping timeout: 365 seconds]
02:01 < fantasai> bz: not allowed
02:01 <@bz> it's not?
02:01 < fantasai> bz: no
02:01 < fantasai> http://www.w3.org/TR/html4/struct/tables.html#h-11.2.1
02:01 -!- Ganesh [~Ganesh@203.145.156.130] has quit [Ping timeout: 365 seconds]
02:01 <@bz> ah, indeed
02:01 <@bz> good
02:02 <@bz> So what should rules=groups do if there are no colgroups?
02:02 < fantasai> then there's no group for it to delineate, now is there?
02:03 < lindi-> lucifer_: well, it works here
02:03 <@bz> so
02:03 <@bz> table[rules=groups] > colgroup { border-left: 1px solid; border-right: 1px solid }
02:04 <@bz> table[rules=groups] > thead, table[rules=groups] > tfoot, table[rules=groups] > tbody { border-top: 1px solid; border-bottom: 1px solid }
02:04 * bz tries that
02:04 < bernd> bz: use myrules
02:04 <@bz> sure
02:04 * bz is using bernd's testcase anyway
02:05 < bernd> its fantasai's
02:05 < fantasai> :)
02:06 <@bz> why didn't people put this css in the testcase so we can see _how_ it doesn't work? ;)
02:06 <@bz> hrm
02:06 <@bz> it doesn't paint anything
02:06 <@bz> known issue?
02:06 <@bz> oh, wait
02:06 <@bz> border-collapse
02:06 <@bz> worksforme
02:06 <@bz> can we check it in? ;)
02:07 <@bz> Note that border with a given value setting the width in px is trivial
02:07 < fantasai> check what in?
02:08 < fantasai> and put what css in what testcase?
02:08 <@bz> the rules in http://bugzilla.mozilla.org/attachment.cgi?id=10217&action=view
02:08 * bz backs up
02:08 <@bz> I am looking at http://bugzilla.mozilla.org/attachment.cgi?id=10217&action=view
02:08 < fantasai> ok
02:08 <@bz> it has this nice comment
02:08 <@bz> /* Groups don't work */
02:08 < fantasai> oh
02:08 -!- dcmwai [~dcmwai@219.95.155.153] has quit [Quit: Client exiting]
02:08 <@bz> and no css that would demonstrate an attempt to make them work
02:09 < fantasai> bz: I'm not entirely sure why
02:09 < fantasai> bz: but the rest of the test case sets everything on the cells
02:09 <@bz> table[myrules=groups] { border-collapse: collapse }
02:09 <@bz> table[myrules=groups] > colgroup {
02:09 <@bz> border-left: 1px solid;
02:09 <@bz> border-right: 1px solid
02:09 <@bz> }
02:09 <@bz> table[myrules=groups] > thead,
02:09 <@bz> table[myrules=groups] > tfoot,
02:09 <@bz> table[myrules=groups] > tbody {
02:09 <@bz> border-top: 1px solid;
02:09 <@bz> border-bottom: 1px solid
02:09 <@bz> }
02:09 <@dbaron> doesn't rules=groups mean between and not outside the groups?
02:10 <@bz> not clear...
02:10 < fantasai> dbaron: that's why we have 'hidden'
02:10 <@bz> but that could be done too
02:10 -!- jst_ [~jst@adsl-64-175-237-19.dsl.sntc01.pacbell.net] has joined #mozilla
02:10 -!- mode/#mozilla [+o jst_] by killer
02:10 <@dbaron> fantasai: ah, right
02:10 < fantasai> :)
02:10 <@bz> yeah, this is between
02:10 <@dbaron> fantasai: but you need to make sure frame= overrides
02:10 < fantasai> dbaron: that part of the CSS is further down
02:10 <@dbaron> watch specificities
02:10 * bernd hears the smashing sound of http://lxr.mozilla.org/seamonkey/search?string=aIgnoreIfrules going away
02:10 < fantasai> dbaron: irrelevant
02:10 < fantasai> table {border: hidden}
02:11 < fantasai> none of the others set a border on the table
02:11 <@dbaron> frame=border doesn't?
02:11 < fantasai> and specificities don't determine how borders collapse
02:11 < fantasai> dbaron: if you have frame=border and rules=groups
02:11 * dbaron wonders how frame=border works, then
02:11 <@bz> this is the thing
02:11 < fantasai> then you have borders between groups
02:11 -!- Hendy is now known as Hendikins
02:11 < fantasai> and around the table
02:11 <@bz> I'm reading the HTML spec
02:12 <@dbaron> fantasai: but you've got to be careful of specificity and/or order with the border-style: hidden
02:12 <@bz> groups: Rules will appear between row groups (see THEAD, TFOOT, and TBODY) and column groups (see COLGROUP and COL) only.
02:12 <@dbaron> fantasai: so that the border from the frame=border beats it
02:12 < fantasai> dbaron: the only "hidden" you'd need is on <table>
02:12 < fantasai> dbaron: and frame=border would cause something solid to go on the table
02:12 * bz senses we need a testcase suite here and a clear spec of what the behavior is
02:12 <@dbaron> fantasai: I know.
02:13 <@dbaron> fantasai: but you have to have the frame=border rule after the rules=groups rule
02:13 <@dbaron> fantasai: and you have to be careful of not adding specificity in weird ways
02:13 -!- ajschult [~ajschult@user-37ka01u.dialup.mindspring.com] has joined #mozilla
02:14 <@dbaron> fantasai: e.g., table[frame=border] { border-style: solid; } table[rules=groups] { border-style: hidden; } wouldn't work
02:14 <@bz> hmm
02:14 < fantasai> dbaron: I wouldn't do that
02:14 <@dbaron> then again, you probably already know all that
02:15 < fantasai> hehehehe
02:15 -!- berkut [~chatzilla@c-24-20-211-90.client.comcast.net] has joined #mozilla
02:15 -!- jst [~jst@adsl-64-175-237-19.dsl.sntc01.pacbell.net] has quit [Ping timeout: 365 seconds]
02:15 <@bz> come to think of it....
02:15 <@bz> table[myrules=groups] > colgroup + colgroup {
02:15 <@bz> border-left: 1px solid;
02:15 <@bz> }
02:15 <@bz> That may be better
02:15 <@bz> doesn't put a left border on the first colgroup or a right border on the last one
02:15 < fantasai> nice
02:16 <@dbaron> bz: what about <colgroup/><col/><colgroup/>
02:16 <@dbaron> or is that illegal?
02:16 <@bz> dbaron: not valid. ;)
02:16 <@bz> dbaron: I asked that about 10 mins ago
02:16 <@dbaron> but it's legal in CSS?
02:16 <@dbaron> I guess that doesn't matter, though.
02:16 <@bz> right
02:16 * fantasai girins
02:16 < fantasai> s/gi/g/
02:17 < fantasai> happy perl syntax~
02:17 <@bz> so here is a question
02:17 * dbaron wonders if we'd be better off with scoped stylesheets for these things
02:17 <@bz> if I only have a thead, I shouldn't do rules, right? ;)
02:17 < fantasai> you wouldn't need to
02:17 < fantasai> dbaron: I don't know what exactly you mean, but this all needs to be overridable by the author's CSS
02:18 <@bz> btw, should we support stuff like <thead/><tbody/><tfoot/> ?
02:18 <@bz> note the ordering
02:18 < fantasai> dbaron: I filed a bug on that when we we'd hard-coded it like that
02:18 < fantasai> dbaron: wouldn't want to regress ;)
02:18 <@dbaron> fantasai: table[rules=groups] { -moz-stylesheet: url(rules-groups.css); }
02:18 < fantasai> dbaron: why?
02:18 <@dbaron> (XBL can already do that, but we could add something so you don't need to go through XBL, perhaps)
02:18 <@dbaron> performance
02:18 <@bz> yeah...
02:19 <@bz> if we have rules that set borders on td, that may hurt
02:19 <@bz> even with '>' combinators
02:19 <@dbaron> we also need to unbreak :root/:-moz-bound-element for scoped stylesheets
02:19 * bz can see us getting to scoped inline styles soon. ;)
02:19 < bernd> dbaron: so it would load the stylesheet only when rules=groups
02:19 <@dbaron> and probably remove :-moz-bound-element in favor of :root
02:19 <@dbaron> bernd: yes
02:20 <@bz> so _should_ we support <thead/><tbody/><tfoot/> ?
02:20 < fantasai> bz: iirc, it's illegal
02:20 <@bz> yes
02:20 <@bz> that's not my question
02:20 < bernd> fantasai: why its illegal?
02:20 <@bz> bernd: per html
02:20 <@dbaron> content model of (tr)+ ?
02:21 <@dbaron> not (tr)*
02:21 < fantasai> http://www.w3.org/TR/html4/struct/tables.html#h-11.2.1
02:21 <@bz> bernd: the order is <thead/><tfoot/><tbody/>+
02:21 < bernd> great than I can invalidate that nasty bc bug :-)
02:21 <@dbaron> oh, I was thinking of something else
02:21 <@bz> eh?
02:21 <@bz> thing is, we _do_ support the broken syntax to some extent
02:21 <@bz> going to some pains to do it, even
02:21 < bernd> bc borders fail when a empty rowgroup is inbetween
02:21 <@dbaron> empty rowgroups are illegal too
02:22 <@dbaron> that's what I was thinking the previous question was
02:22 <@dbaron> I think they should be legal, but...
02:22 * dbaron should probably go back to reviewing code
02:23 <@bz> and we want this to work with XHTML
02:23 <@bz> where tbody is optional....
02:23 -!- kerz [asdf@gw.office.mozilla.org] has joined #mozilla
02:23 <@bz> but TFOOT could still happen?
02:23 * bz cries
02:24 <@bz> Rather than require document authors to insert extraneous elements, XHTML has made the elements optional. User agents need to adapt to this accordingly.
02:24 * bz flips the XHTML spec authors the bird
02:25 <@bz> ok
02:25 * fantasai consoles bz
02:25 <@bz> so realistically, no one will use a tfoot without a tbody, right? ;)
02:25 <@bz> so....
02:25 < fantasai> bz: what's the content mode, exactly?
02:25 < bernd> bz: hixie will
02:25 < fantasai> *model
02:25 -!- berkut [~chatzilla@c-24-20-211-90.client.comcast.net] has quit [Quit: ChatZilla 0.9.59 [Mozilla rv:1.7a/20040208]]
02:25 < fantasai> bernd: lol
02:26 <@bz> table[myrules=groups] { border-collapse: collapse }
02:26 <@bz> table[myrules=groups] > colgroup + colgroup {
02:26 <@bz> border-left: 1px solid;
02:26 <@bz> }
02:26 <@bz> table[myrules=groups] > tfoot, table[myrules=groups] > tbody + tbody {
02:26 <@bz> border-top: 1px solid;
02:26 <@bz> }
02:26 <@bz> table[myrules=groups] > thead {
02:26 <@bz> border-bottom: 1px solid;
02:26 <@bz> }
02:26 <@bz> fantasai: content mode?
02:27 < fantasai> model
02:27 < fantasai> for XHTML <table>
02:27 <@bz> oh
02:27 <@bz> well....
02:27 < fantasai> btw, you don't need that second rule set
02:27 <@dbaron> as long as both thead and tfoot have borders, you don't need the border from the omitted tbbody
02:27 <@bz> I don't?
02:27 <@bz> you can have multiple tbodies.....
02:27 <@dbaron> not when they're omitted
02:27 <@bz> sure
02:28 <@bz> but I can't be sure it's omitted....
02:28 <@dbaron> why does it matter?
02:28 <@bz> I was trying to get the case when there is no thead or tbody and only a tfoot right
02:28 <@dbaron> as long as thead has a bottom border, tfoot has a top border, and tbody has both borders, you're OK for rules=groups
02:28 <@bz> it should have no border
02:28 <@bz> with those rules it has a top border
02:28 <@dbaron> why should the tfoot have no border?
02:28 <@bz> oh
02:28 <@bz> heh
02:28 <@bz> well, I was going on the "rules between groups" interpretation
02:29 <@bz> is it wrong?
02:29 < fantasai> no, it's right
02:29 <@dbaron> consider it an implied group when tbody is omitted
02:29 < fantasai> but, bz, if it's making things complicated
02:29 <@bz> ah, heh
02:29 < fantasai> you can use hidden on the table
02:29 <@bz> no, it's not
02:29 <@bz> those rules right there do it
02:29 < fantasai> and it will be just as correct per HTML
02:29 <@bz> assuming we "consider it an implied group when tbody is omitted"
02:29 <@bz> which we should, imo
02:29 <@bz> it makes sense to me
02:29 <@bz> ;)
02:29 * fantasai goes to dig up XHTML
02:29 <@bz> it's useless
02:29 <@bz> don't bother
02:30 <@bz> they basically say that any time you have a tag where start and end are both optional in HTML, the XHTML version the whole tag is optional
02:30 <@bz> and may not appear in the DOM
02:30 <@bz> oops
02:30 <@dbaron> bz: no, just tbody, not body, head, or html. :-)
02:30 <@bz> really?
02:30 <@bz> argh
02:30 <@dbaron> head is not optional in XHTML
02:31 <@bz> Also, some XHTML elements may or may not appear in the object tree because they are optional in the content model (e.g. the tbody element within table). This occurs because in HTML 4 some elements were permitted to be minimized such that their start and en
02:31 <@dbaron> what else had both tags omissible?
02:31 <@bz> d tags are both omitted (an SGML feature).
02:31 < fantasai> bz: that's a stupid content model
02:31 < fantasai> bz: they could have done better
02:31 <@bz> fantasai: YES. ;)
02:31 < fantasai> bz: if they'd just thought about it
02:31 <@bz> dbaron: nothing I can think of.....
02:31 * fantasai rearranges it
02:31 <@dbaron> fantasai: which is a stupid content model?
02:31 <@bz> fantasai: "if"
02:31 < fantasai> <table>
02:31 <@bz> ok
02:31 * dbaron would have prevented no-TBODY with thead/tfoot, at least
02:31 < fantasai> right now it's (caption?, (col*|colgroup*), thead?, tfoot?, (tbody+|tr+))
02:32 * bz goes to attach his rules to the bug
02:32 <@bz> yes
02:32 <@bz> that's correct
02:32 < fantasai> it should be (caption?, (col*|colgroup*), ((thead?, tfoot?, tbody+)|tr+))
02:32 <@bz> "* dbaron would have prevented no-TBODY with thead/tfoot, at least"
02:33 < bernd> bz: there is already a testcase list for bc http://bugzilla.mozilla.org/attachment.cgi?id=63256&action=view
02:33 < fantasai> bz: when you're finished having fun with CSS rules, post them in the bug? :)
02:33 <@bz> fantasai: so I will add these rules to http://bugzilla.mozilla.org/attachment.cgi?id=10217&action=view and attach to bug 43178 ?
02:34 * bz was too slow at getting bug numbers
02:34 < fantasai> yeah, go ahead
02:34 < fantasai> hmm, 2:32am
02:34 <@bz> we need testcases not for bc but for these attr values specifically
02:34 <@bz> in various combinations
02:34 < fantasai> I hope this excitement keeps me awake for much longer ^^;;
02:35 <@bz> heh
02:35 < fantasai> run a google search ;)
02:35 <@bz> well, I'll post testcase in a sec
02:35 -!- fire_fire [ffdsfds@FSH200159.resnet.ucsb.edu] has quit []
02:35 * bernd needs to go earn some money
02:35 < fantasai> wonder if I have some lying around..
02:35 -!- bernd [~chatzilla@pD9EB11FA.dip0.t-ipconnect.de] has quit [Quit: ChatZilla 0.9.59 [Mozilla rv:1.7a/20040202]]
02:37 < fantasai> don't see any
02:37 -!- lucifer_ [~lucifer@pcp02797998pcs.goosck01.sc.comcast.net] has quit [Quit: Client exiting]
02:39 <@bz> sorry; taking a min to clean up this stylesheet a bit
02:41 <@bz> should I roll "frame" in here too?
02:42 -!- jshin [~chatzilla@218.158.75.152] has joined #mozilla
02:43 < fantasai> bz: yeah
02:43 * fantasai finishes part 1 of the email
02:44 < fantasai> no multicolumn columns, thanks
02:44 <@bz> I'm gonna go with hidden borders on table, actually
02:44 < fantasai> bz: makes things cleaner :)
02:45 <@bz> yeah
02:45 -!- botbot [~pid-7897@bah.mozillazine.org] has quit [Quit: I was told to shutdown by kerz. :-(]
02:45 < jshin> bz: Hi, i've got a question about string...
02:45 * bz realized that '+' uses indexof, which is not quick
02:45 <@bz> hey jshin
02:45 < fantasai> bz: one of the other things I wish I'd followed up on more closely is altss switching
02:45 < jshin> bz: hi ! i'm wondering if repeated use of getter_Copies on a single nsXPIDLString leaks ..
02:45 -!- kravi [~kravi@202.144.86.147] has joined #mozilla
02:47 <@bz> so here is a question
02:47 <@bz> jshin: it shouldn't
02:47 <@bz> fantasai: if I have rules=cols, say
02:48 <@bz> actually, nevermind
02:48 * bz goes back to what he was doing
02:48 < jshin> bz: so, I can safely use patterns like : method1(getter_Copies(xpidlstr)); method2(getter_Copies(xpidlstr)), right?
02:48 <@bz> hmm
02:48 <@bz> no, this is a question
02:48 <@bz> jshi: yep
02:49 <@bz> say my user stylesheet has td { border-top: 1px solid red }
02:49 <@bz> and a page has table groups=cols
02:49 < jshin> bz: thx !
02:49 <@bz> should those cells have a red top border?
02:51 < fantasai> bz: probably
02:51 <@bz> ok
02:51 < fantasai> bz: but, I don't think it's important enough for you to avoid hidden
02:51 <@bz> well, no
02:51 <@bz> I was just asking whether groups=cols should explicitly set top/bottom borders. ;)
02:52 < fantasai> heh
02:52 < fantasai> I think it should do the minimum
02:52 <@bz> ok
02:52 <@bz> makes sense
02:52 < fantasai> of what it needs to do
02:52 <@bz> btw, in border-collapse mode
02:52 < fantasai> yes?
02:52 <@bz> a border of "none" will override things, right?
02:52 <@bz> or no?
02:52 < fantasai> no
02:52 < fantasai> "hidden"
02:52 < fantasai> will override tings
02:52 < fantasai> things
02:53 <@bz> ok
02:53 < fantasai> bz: are you having fun with this? :)
02:54 -!- dmose [~dmose@dsl081-050-187.sfo1.dsl.speakeasy.net] has joined #mozilla
02:54 -!- mode/#mozilla [+o dmose] by killer
02:54 -!- JustinArthur [~JustinArt@pool-141-152-69-120.roa.east.verizon.net] has quit [Quit: Tired.]
02:55 <@dmose> rheeeeet
02:55 < fantasai> greeeetings, sir
02:55 < Dauphin> hey dmose
02:55 <@Pavlov> rheet
02:55 <@dmose> how goes, folks?
02:56 * dmose is laying LDAP smack down
02:58 -!- bb [~bb@dsl-082-082-220-126.arcor-ip.net] has joined #mozilla
02:58 <@bz> fantasai: yes. ;)
02:58 * bz attaches the testcase
02:58 < fantasai> And three browsers vote for 'visibility: collapse'!
02:58 <@bz> we need better coverage, though
02:58 <@bz> dmose: heh
02:59 < fantasai> bz: there was one test case on microsoft's site, linked from the bc list
02:59 < fantasai> bz: it covered everything
02:59 -!- drepper [drepper@cpe-24-221-190-179.ca.sprintbbd.net] has quit [Remote host closed the connection]
02:59 < fantasai> but it was scripted
02:59 < fantasai> and the scripting wasn't working for me
02:59 < fantasai> 'twas at the top of the list
02:59 < fantasai> if you want to take a look
02:59 * bz scrolls up
02:59 -!- hewitt [~hewitt@adsl-64-175-36-75.dsl.pltn13.pacbell.net] has joined #mozilla
03:00 * bz downloads that testcase
03:00 <@bz> scripted is good. ;)
03:02 <@bz> oh, the scripting is IE-only
03:02 * bz kicks it
03:03 < fantasai> must be why it didn't work ;)
03:04 < blake> fx 0.8 is out.
03:04 * fantasai , not being much familiar with scripting, doesn't know how easy it would be to adapt it for Moz
03:04 <@bz> I already did
03:04 <@bz> replace-regexp is great. ;)
03:04 < fantasai> ;D
03:05 <@bz> some of these don't work...
03:05 * bz checks why
03:05 < Dauphin> ugh.
03:05 -!- ben [ben@gw.office.mozilla.org] has joined #mozilla
03:05 * Dauphin wonders when Ben forked these firebird installer scripts
03:05 <@bz> first off, our rules for "rows" show gaps between the cells
03:05 < Dauphin> speak of the devil
03:05 <@bz> I wonder why
03:05 <@Pavlov> mmm, fx
03:05 -!- mode/#mozilla [+o ben] by killer
03:05 -!- glazou [~daniel@dyn-195-242-103-97.ppp.tiscali.fr] has joined #mozilla
03:05 < glazou> bonjour
03:05 < fantasai> bz: put it up?
03:05 -!- jshin [~chatzilla@218.158.75.152] has quit [Remote host closed the connection]
03:05 < glazou> ROFL @ http://bugzilla.mozilla.org/show_bug.cgi?id=217808#c2
03:05 < Dauphin> Ben: how long ago did you fork the firebird installer scripts?
03:05 < fantasai> bonjour
03:05 <@bz> oh... duh
03:06 -!- bryner [bryner@gw.office.mozilla.org] has joined #mozilla
03:06 <@ben> Dauphin: when the files were created.
03:06 < glazou> hi bz fantasai
03:06 <@ben> busy with release. back later.
03:06 -!- mode/#mozilla [+o bryner] by killer
03:06 < gisburn> blake: http://ftp.mozilla.org/pub/mozilla.org/firebird/releases/0.8/ - where are the binaries ? :)
03:06 <@bz> hey glazou
03:06 <@Pavlov> gisburn: http://ftp.mozilla.org/pub/mozilla.org/firefox/releases/0.8
03:06 * Dauphin hrms
03:06 < gisburn> uhm uhm
03:07 < gisburn> WTF is firebox ?
03:07 < gisburn> don't tell me you renamed it again
03:07 <@Pavlov> (;
03:07 < gisburn> s/box/fox/
03:08 <@bz> hmm
03:08 <@bz> why would row borders not paint??
03:08 < glazou> hi Pavlov
03:08 <@Pavlov> :-)
03:08 < glazou> ben: congrats on 0.8 :-)
03:09 < fantasai> bz: are the borders dynamically added?
03:09 <@bz> oh, I see why
03:09 <@bz> yes
03:09 * fantasai would hate if that were the problem
03:09 < fantasai> it is?
03:09 <@bz> and we have bugs on not invalidating bc-stuff
03:09 < fantasai> Clearly why it needs a rewrite!
03:09 <@bz> if I switch from rules=none to rules=whatever it works
03:09 <@bz> since it reframes
03:09 <@bz> we have a bug on this already
03:09 <@bz> bernd owns it, iirc
03:10 < fantasai> *darkly* everything about BC indicates that it needs a rewrite
03:10 < fantasai> :)
03:10 * gisburn gets 404 at http://ftp.mozilla.org/pub/mozilla.org/firefox/releases/0.8
03:10 < ajschult> gisburn: it's a mirror.
03:11 <@Pavlov> it'll be up in a few mins
03:11 <@bz> yeah, dynamic stuff has issues
03:11 < fantasai> bz: views also have issues
03:12 < fantasai> views on tables have issues in general
03:12 <@Pavlov> ftp://mozilla.ussg.indiana.edu/pub/mozilla.org/firefox/releases/0.8/
03:12 <@bz> heh
03:12 < fantasai> bz: I'm getting some extremely wacky painting on one of my testcases
03:12 < fantasai> I know it's not my fault only because it does the same thing when I back out my changes ;)
03:13 -!- patrick [patrick@c-24-6-51-183.client.comcast.net] has left #mozilla []
03:13 -!- gabrielle is now known as hyatt
03:14 -!- JanSleep is now known as Jan
03:14 < glazou> oh :-)
03:14 < glazou> everybody's on board!
03:14 <@bz> <TABLE ID="oID_1" BORDER="" MYRULES="cols">
03:14 <@bz> how should that render?
03:15 <@bz> fantasai: heh
03:15 -!- harshal [~harshal@210.18.23.22.sify.net] has joined #mozilla
03:15 < hyatt> MYRULES?
03:15 < hyatt> wtf is MYRULES?
03:15 < basic> lol
03:15 < fantasai> hyatt: we're testing
03:15 <@bz> hyatt: a way to test a ua.css rules impl without it colliding with our C++ code
03:15 <@bz> so
03:15 < hyatt> ah
03:15 <@bz> should that have a left and right border on the table?
03:15 * hyatt was going to try to impl rules using border-collapsing etc.
03:15 <@bz> that's us
03:16 < fantasai> that's what we're going to do
03:16 <@bz> that's what we wanna do
03:16 < hyatt> oooh, that would be cool.
03:16 <@bz> and then remove some code that thinks it knows what it's doing
03:16 < hyatt> i have no frame/rules impl in safari yet
03:16 < hyatt> but i do have border collapsing supported
03:16 < fantasai> hyatt: we'd have had it three years ago, except karnaze wasn't paying attention
03:16 < fantasai> :P
03:16 <@bz> post-resolve callbacks my ass
03:16 <@bz> fantasai: so does "border" imply "frame=border"?
03:16 <@dbaron> bz: that's what I was saying about not being at the preshint level anyway
03:16 < hyatt> i assume it's ok to set border-collapse: collapse when someone starts specifying rules="rows" and what-not?
03:17 < gisburn> In which branch does firefox sit ?
03:17 < hyatt> that was what i was unsure of
03:17 <@bz> dbaron: heh
03:17 < hyatt> if it was ok to turn the table into a border-collapsing table when rules was specified
03:17 < fantasai> bz: hang on
03:17 <@bz> hyatt: that's our plan. We do it anyway. ;)
03:17 < fantasai> hyatt: http://bugzilla.mozilla.org/show_bug.cgi?id=43178
03:18 <@bz> table[frame=box],
03:18 <@bz> table[frame=border],
03:18 <@bz> table[border=""]:not([frame]) {
03:18 <@bz> border: thin outset;
03:18 <@bz> }
03:18 <@bz> seem reasonable?
03:19 < hyatt> mmm, won't that be less efficient than using mapped attributes
03:19 <@bz> perhaps
03:19 <@bz> for now I'm just testing
03:19 * fantasai studies bz's code and html.css
03:20 <@bz> that code is not working for some reason....
03:20 <@bz> that rule is not applying to my table
03:20 < hyatt> are you in strict mode?
03:20 < hyatt> in your html?
03:20 <@bz> yes
03:20 < hyatt> case-sensitivity
03:20 <@bz> er...
03:20 < hyatt> if you're in strict mode
03:20 <@bz> no, I am not
03:20 <@bz> anyway
03:20 < fantasai> case sensitivity?
03:20 <@bz> [border] applies
03:20 < fantasai> what case sensitivity
03:20 <@bz> [border=""] does not
03:21 < fantasai> ?
03:21 <@bz> odd
03:21 < fantasai> bz
03:21 < fantasai> [border]
03:21 < fantasai> no?
03:21 <@bz> sure
03:21 <@bz> I just wonder why the other doesn't apply... ;)
03:21 < fantasai> what's the HTML?
03:21 -!- drepper [drepper@cpe-24-221-190-179.ca.sprintbbd.net] has joined #mozilla
03:22 <@bz> hrmph
03:22 <@bz> I need "myborder"
03:22 <@bz> since "border" triggers our friggin' html.css
03:22 < fantasai> hyatt: aren't attr=(val1|val2|val3) -type attributes case-insensitive?
03:22 < fantasai> (except in XHTML, when they're lower-case)
03:22 < fantasai> bz: it's not in html.css, it seems
03:23 <@bz> border?
03:23 < fantasai> bz: it's in the preshint ss
03:23 <@bz> it may be in attr mapping too
03:23 <@bz> right
03:23 < fantasai> bz: it has to handle border=3
03:23 < fantasai> which html.css can't do
03:24 <@bz> hrmph
03:24 <@bz> so what should <table border rules="groups"> do?
03:24 <@bz> <table border> sets borders on all the frames....
03:24 * bz adds :not([rules])
03:24 < fantasai> bz: groups wins
03:24 < fantasai> because border implies rules=all
03:25 < fantasai> and that's where all those little borders come from
03:25 < fantasai> bz: http://www.w3.org/TR/html4/struct/tables.html#adef-border-TABLE
03:27 < hyatt> fantasai: they should be case-insensitive always in html docs IMO
03:27 < hyatt> fantasai: but i'm not sure moz does it that way
03:27 < hyatt> i think moz may just use quirks/strict as the trigger
03:27 < hyatt> i know moz uses quirks/strict as the trigger for class/id case-sensitivity
03:27 < hyatt> safari uses quirks/stirct for id/class and html/xml doc as the trigger for attr values/names
03:28 < fantasai> yeah, but those are different datatypes
03:28 < fantasai> CDATA and ID, IIRC
03:28 < hyatt> (at least it does internally, 1.2 is just always case-insensitive in html)
03:28 * hyatt stops babbling.
03:28 < fantasai> hehe
03:28 -!- conan [~conan@dhcp-ams-cam-vl10-144-254-193-239.cisco.com] has joined #mozilla
03:28 * fantasai won't hold it against you
03:28 <@dmose> hyatt: rheeeeeeeeeeet!
03:29 < hyatt> i think it's going to be hard to beat the perf of mapped attributes with stylesheet rules
03:29 < hyatt> for this particular case
03:29 < glazou> b������ dmose
03:29 < hyatt> since i think the rules are going to be quite nasty
03:29 <@dmose> hey glazou!
03:29 < hyatt> but hey, i wanted to do it too
03:29 < hyatt> :)
03:29 <@bz> hrmph
03:29 < glazou> hyatt: coming to cannes for ftf ?
03:29 < hyatt> glazou: no, not a chance.
03:29 <@bz> <table border rules=groups> seems unhappy
03:29 < glazou> hyatt: :-(
03:30 < fantasai> bz: lol
03:30 < hyatt> glazou: apple isn't going to foot the bill for globe-trotting
03:30 < hyatt> :)
03:30 < fantasai> bz: if it's unhappy, blame it on the parser
03:30 <@bz> oh, I see why
03:30 < fantasai> bz: because it's supposed to translate that into
03:30 < glazou> hyatt: just move to Paris, and Cannes will be at 1 hour from your work place ;-)
03:30 < fantasai> <table frame=border rules=groups>
03:30 * bz can't spell 'myrules'
03:30 * bz forgot the 'my'
03:30 < fantasai> heh
03:31 -!- Zeron [user@216.199.55.12] has joined #mozilla
03:31 * fantasai wonders if Tantek would strangle her if she showed up in Cannes XP
03:31 < glazou> fantasai: why ???
03:31 < glazou> fantasai: I'd love to see you attend our group meeting as an observer
03:31 < fantasai> I am sure my late CSS2.1 comments are exasperating
03:31 < fantasai> I know they're exasperating Hixie
03:32 < fantasai> glazou: I've actually thought about it a bit
03:32 < glazou> fantasai: don't worry ; the HTML WG has a looooong well-known tradition of sending late comments too :-) and that's worse since it's internal to w3c !
03:32 < glazou> fantasai: the problem is on w3c's side, not yours
03:33 * fantasai of course /wouldn't/ expect to be strangled by anyone at the W3C
03:33 < glazou> you have the right to send comments
03:33 < fantasai> hehe
03:33 < glazou> but W3C does not know how to say "too late for official response", their problem
03:33 < glazou> I am serious
03:33 < glazou> I had the same problem with Selectors comments for CR
03:33 <@bz> so...
03:33 -!- benG5 [benG5@gw.office.mozilla.org] has joined #mozilla
03:33 <@bz> these testcases are treating rules="" as the same thing as "rules not set"
03:33 -!- hyatt is now known as gabrielle
03:33 -!- benG5 is now known as hyatt\
03:34 * bz kicks them
03:34 -!- hyatt\ is now known as hyatt
03:34 * dmose doesn't even want to know
03:34 -!- gabrielle [gabrielle@gw.office.mozilla.org] has quit []
03:34 < fantasai> bz: ...
03:34 < fantasai> bz: what exactly is the problem?
03:35 < hyatt> dmose: rheet
03:35 <@dbaron> hyatt: what's gabrielle's IP?
03:36 <@bz> <table rules="" border="">
03:36 <@bz> what should be the rendering?
03:36 * fantasai tries to work out what exactly glazou meant by "W3C does not know... Selectors comments for CR" and guess that he was referring to late internal official-stamped comments
03:36 < hyatt> dbaron: 125
03:36 <@dbaron> hyatt: thanks
03:36 <@bryner> dbaron: there was no name change on the trunk, the tinderboxes should be ok.
03:36 < fantasai> bz: undefined, I think
03:36 <@bz> heh
03:36 <@dbaron> bryner: I'm changing the name of the tinderbox tree.
03:36 <@bz> oh, wait
03:36 <@bryner> oh, ok
03:36 * bz kicks self
03:36 <@bryner> thanks.
03:37 < fantasai> bz should play soccer
03:37 <@bz> the problem is that in the DOM when people do .rules = ''
03:37 <@bz> that sets rules to ''
03:37 <@bz> instead of unsetting it....
03:37 <@bz> as far as I can tell
03:37 * bz won't worry about that edge case for now
03:39 -!- EliteEagle is now known as Eagle`sl33p
03:41 <@dbaron> hyatt: I'm done
03:41 < hyatt> dbaron: no problem, i'm settled in on the mac now
03:42 -!- bbaetz [~bbaetz@c211-30-0-2.wavrl1.nsw.optusnet.com.au] has joined #mozilla
03:43 -!- mcsmurf_away [~mcsmurf@pD95439EB.dip.t-dialin.net] has quit [Ping timeout: 365 seconds]
03:43 <@bz> bingo!
03:43 <@bz> fantasai: the dynamic testcases work
03:44 < fantasai> bz: cool
03:45 * bz should sleep
03:45 <@bz> css in bug
03:45 <@bz> we should really do border via attr mapping, though
03:45 <@bz> except for its effect on cells
03:45 <@bz> right now that's done via post-reflow callbacks
03:45 <@bz> which is just evil
03:46 < hyatt> why post-reflow?
03:46 <@bz> er, sorry
03:46 <@bz> post-resolve
03:46 < hyatt> right
03:46 <@bz> "evil"
03:46 < hyatt> i was going to say
03:46 < hyatt> safari does it post-resolve
03:46 < hyatt> and then it isn't dynamic
03:46 < hyatt> tee-hee
03:46 <@bz> heh
03:46 < hyatt> it's more like post-frame construction
03:46 -!- PaperTimedOut [Arron@u176147.ap.plala.or.jp] has joined #mozilla
03:46 < hyatt> i am working on fixing that right now in fact.
03:46 < hyatt> how odd and eerie
03:47 <@bz> the problem is writing all these rules the would be slow...
03:47 < hyatt> yeah, the mapped attribute stuff is fast
03:47 < hyatt> and it gets hashed and shared by multiple tables
03:47 <@bz> hyatt: http://bugzilla.mozilla.org/attachment.cgi?id=140929&action=view
03:47 <@bz> hyatt: see whatcha think of the stylesheet there.... feedback welcome on http://bugzilla.mozilla.org/show_bug.cgi?id=43178
03:48 < fantasai> bz: I don't think rules=all should have outer borders
03:48 <@dbaron> we could also implement it in C++ and have one set of rules per table
03:48 <@dbaron> which would be reasonably fast
03:48 <@dbaron> or even one set of rules per table + attribute combination
03:48 <@dbaron> er, attribute combination
03:49 <@dbaron> or I was thinking of scoped stylesheets hooked in by the rules/frame attributes
03:49 <@dbaron> which would require our improving scoped stylesheets a bit
03:51 -!- Ganesh [~Ganesh@203.145.156.130] has joined #mozilla
03:51 <@bz> dbaron: hmm I'm seeing bug 123355 in my debug build
03:51 <@bz> dbaron: pulled a day or two ago
03:51 <@bz> fantasai: comment in bug? Or post updates?
03:51 * dbaron wonders how CSS style rule implements GetStyleSheet given the style sheet inner stuff
03:51 < fantasai> bz: I haven't time to post updates
03:51 -!- PaperSleep [Arron@t172140.ap.plala.or.jp] has quit [Ping timeout: 365 seconds]
03:52 < fantasai> bz: as it is, I don't think I'm going to sleep much, if at all, tonight
03:52 <@bz> fantasai: ok.
03:52 * bz makes the update
03:53 * fantasai has been working on 4510 all weekend, and now has a CSS2.1 email as well as other non-web related things to take care of for tomorrow
03:53 < fantasai> :/
03:53 <@bz> :(
03:53 * bz should sleep too....
03:54 <@bz> dbaron: we clone the sheet when someone tries to access a rule via the dom apis
03:54 <@bz> dbaron: and non-dom people ... well... ;)
03:54 <@bz> dbaron: though we have bugs in the cloning, of course
03:55 <@bz> http://web.mit.edu/bzbarsky/www/testcases/css-weirdness/test.html for example
03:56 * dbaron wonders why nsIStyleRule should have GetStyleSheet is right
03:56 < fantasai> dbaron: was http://lists.w3.org/Archives/Public/www-style/2003Mar/0003.html addressed by the WG or do I have to repost it?
03:56 * fantasai notes that this was posted before the last last call
03:56 -!- stianj [stianj@dodge.jordet.nu] has joined #mozilla
03:57 <@dbaron> fantasai: I think mostly...
03:57 <@dbaron> fantasai: I remember discussing the issue, and the eventual resolution was, well, a bit interesting...
03:57 -!- stianj [stianj@dodge.jordet.nu] has quit [Client Quit]
03:58 < fantasai> dbaron: I'm not seeing anything in the last public draft
03:58 <@dbaron> fantasai: no, since then
03:58 < fantasai> dbaron: unless I'm missing it
03:58 < fantasai> dbaron: what was the resolution?
03:58 < hyatt> i thought my proposal was accepted
03:58 < hyatt> for that
03:58 < hyatt> and i sent out wording to the wg
03:58 < hyatt> for changes to the table module
03:58 <@dbaron> we really shouldn't discuss this in public. :-(
03:59 < hyatt> i'm not sure if my changes were ever accepted into the draft though
03:59 -!- sime [~sime@c211-30-245-2.rivrw5.nsw.optusnet.com.au] has joined #mozilla
03:59 < fantasai> dbaron: ok
03:59 <@dbaron> there definitely are changes in the draft relating to the half-in-the-margin stuff
03:59 * dbaron doesn't know who wrote them
03:59 <@bz> dbaron: probably no good reason....
03:59 < hyatt> fantasai: my proposal was to resolve the top border and bottom border normally and not to go into the margin
03:59 < fantasai> eh?
04:00 <@dbaron> bz: eh?
04:00 < hyatt> fantasai: but for left/right, you computed an initial border by examining the first row of the table
04:00 <@bz> dbaron: nsIStyleRule's GetStyleSheet
04:00 < hyatt> and then spill into the margin if things got bigger in subsequent rows
04:00 * dbaron hopes it's not used for anything important
04:00 * glazou is quite glad he resisted to the *bird requests for Composer
04:00 <@Pavlov> i thought it was cockbird
04:01 < glazou> that's the editor going with pornzilla
04:01 <@bz> firefox?????
04:01 <@Pavlov> lol
04:01 <@bz> dbaron: heh
04:01 -!- stianj [stianj@dodge.jordet.nu] has joined #mozilla
04:01 < glazou> bz: yeah...
04:01 < fantasai> hyatt: what do you mean by "resolve the top border and bottom border normally"?
04:01 <@Pavlov> bz you are so an hour behind
04:01 -!- petern [~chatzilla@segbg.cellnetwork.com] has quit [Quit: ChatZilla 0.9.58p [Mozilla rv:1.5/20031007]]
04:01 -!- Hixie [~ianh@ti200710a080-3830.bb.online.no] has joined #mozilla
04:01 < hyatt> fantasai: you actually look at every cell in the top row to compute the top collapsed border for the table
04:01 < glazou> Pavlov: from a mktg, that's very very very confusing for users
04:02 < glazou> mktg pov
04:02 < hyatt> fantasai: and so you never have margin spill
04:02 <@Pavlov> glazou: cockbird?
04:02 < glazou> Pavlov: naaaah; s/firebird/firefox
04:02 < fantasai> hyatt: right.. but where does the extra space come from?
04:02 <@bz> dbaron: it's not, as far as I can tell... just inspector
04:03 < fantasai> hyatt: i.e. which region of the table box
04:03 < fantasai> hyatt: margin, border, padding?
04:03 < hyatt> fantasai: you can think of the table's border width as being 1/2 of the maximum collapsed border
04:03 < fantasai> ok
04:03 < hyatt> so the table gets a border width to account for the largest collapsed border
04:03 < fantasai> and for the sides, you suggested using a first-row heuristic
04:03 < hyatt> exactly
04:03 <@bz> oh, hmmm
04:03 <@bz> and for inline style
04:03 < hyatt> with excess spilling into the table's margins on the left/right
04:03 <@bz> er, not inlin
04:03 < fantasai> hyatt: you know what the problem is with that?
04:03 < hyatt> if a larger collapsed border is encountered on subsequent rows
04:04 <@bz> it's used to go rule -> document, basically
04:04 <@bz> for CSSOM changes to a rule
04:04 < fantasai> hyatt: you're defining the border box to be bigger than the table
04:04 < fantasai> hyatt: so the background will spill out of it like it does in Quirks in Mozilla
04:04 -!- MadMoose [~chatzilla@0x3ef36cc9.virnxx3.adsl.tele.dk] has joined #mozilla
04:04 <@bz> but at that point the sheet had better have been cloned, so we can just use the DOM apis
04:04 < hyatt> fantasai: correct.
04:04 < fantasai> good morning, Hixie
04:04 < hyatt> fantasai: the background of the table could draw outside the cells
04:04 < fantasai> hyatt: that looks *awful*
04:05 -!- bryner [bryner@gw.office.mozilla.org] has quit [Quit: Leaving]
04:05 < hyatt> fantasai: well, using different border widths is somewhat bizarre anyway
04:05 < fantasai> hyatt: I'd rather have ambiguous wording than that.
04:05 < hyatt> (especially for the outside of the table)
04:05 -!- gabrielle [gabrielle@gw.office.mozilla.org] has joined #mozilla
04:05 < hyatt> it's not like people actually vary border widths along the exterior of the table
04:05 < fantasai> hyatt: oh, really?
04:05 <@bz> hyatt: lol
04:05 -!- Sander [~abuse@203.97.90.154] has quit [Quit: And back he spurred like a madman, shrieking a curse to the sky.]
04:05 < hyatt> so the "collapsei nto the margin" behavior is crazy, since it penalizes the common case
04:06 < fantasai> hyatt: tbody { border: 1px solid } thead {border: 4px solid}
04:06 <@bz> hyatt: you name something the spec allows, and people do it
04:06 < fantasai> hyatt: that makes sense, graphically
04:06 <@bz> hyatt: they do crap the spec does not allow either, but those are easier to wontfix
04:06 < fantasai> hyatt: and it would look messed up if you put a background on it
04:06 < hyatt> anyway, gotta go to dennys
04:06 < hyatt> :)
04:06 -!- hyatt [benG5@gw.office.mozilla.org] has quit []
04:06 -!- kerz [asdf@gw.office.mozilla.org] has quit []
04:07 * fantasai mutters that it doesn't "penalize common sense", what on earth does he mean by it