| 00:05 | <Hixie> | k |
| 00:06 | <Hixie> | i just needed somewhere to offload HTML WG bugs when they were really CSS bugs |
| 00:06 | <Hixie> | so the misc component is enough for my purposes |
| 00:07 | <Hixie> | it's fascinating to me how browser vendors are always complaining about how it should be easier to follow the spec changing, but of the dozens of people who have subscribed to the spec's notification mechanism, only a handful are browser vendors |
| 00:08 | <Hixie> | i mean, who are all these people who are following the spec so closely? |
| 00:09 | <Philip`> | People who like to subscribe to everything possible and then filter it out instead of actually reading it? |
| 00:11 | dglazkov | realizes how demeaning the term "browser vendor" sounds. Heeeeere's some browsers! Get a browser! Fresh, hot browsers! |
| 00:12 | <dglazkov> | it totally misrepresents the slow and painful slog that is writing browser software. |
| 00:12 | <dglazkov> | can we rename us to browser sisyphuses? |
| 00:13 | <dglazkov> | or possibly something without sissy and fusses in them? |
| 00:14 | <dglazkov> | ah. I forgot a :) |
| 00:14 | <dglazkov> | just in case I offended Hixie |
| 00:14 | <dglazkov> | :) |
| 00:15 | <Hixie> | :-P |
| 00:15 | <Hixie> | you'll have to work harder than that to offend me :-P |
| 00:15 | <dglazkov> | I dare not. The bar is too high. |
| 00:16 | <divya> | treebeard browsers? |
| 00:16 | <Hixie> | Philip`: actually because of the info in the subscription mails, i think it's better now to just subscribe to all the topics than it is to use the commit list |
| 00:16 | <Hixie> | i should see how many people subscribed to that though |
| 00:17 | <Hixie> | wow, 249 people |
| 00:18 | <Hixie> | wow, including quite a few browser vendors |
| 00:18 | <Hixie> | philip's theory is proved right, i guess |
| 01:53 | <MikeSmith> | dglazkov: the Japanese transliteration of "vendor" is the same as the transliteration of "bender" |
| 01:54 | <MikeSmith> | so let's start saying "browser benders" instead |
| 01:55 | <MikeSmith> | btw, speaking of benders: "Electricity attracts devils and demons. Other instruments attract other spirits. An acoustic guitar attracts Casper. A mandolin attracts Wendy. But an electric guitar attracts Beelzebub." http://blog.wfmu.org/freeform/2009/03/captain-beefhearts-10-commandments-of-guitar-playing.html |
| 02:24 | <MikeSmith> | web-apps-tracker busted? |
| 02:24 | <MikeSmith> | annevk: ↑ |
| 02:24 | <MikeSmith> | http://html5.org/tools/web-apps-tracker |
| 02:25 | <MikeSmith> | wow |
| 02:25 | <MikeSmith> | load average: 3.88, 4.97, 6.40 |
| 02:31 | <Hixie> | well that means it's getting better |
| 02:33 | <MikeSmith> | yeah |
| 02:33 | <MikeSmith> | um, I think it's my fault |
| 02:33 | <MikeSmith> | because I tweeted about the e-mail regexp |
| 02:34 | <MikeSmith> | geeks love that crap |
| 02:34 | <MikeSmith> | fish food |
| 02:35 | MikeSmith | apologizes to html5.org server for harshing his mellow |
| 02:35 | <Hixie> | that explains whatwg.org being down too then |
| 02:35 | <Hixie> | who knew that would be that popular |
| 02:37 | <MikeSmith> | code snippet of any kind is like catnip for the geek |
| 02:39 | <Hixie> | i guess |
| 02:39 | <Hixie> | but this is silly |
| 02:39 | <Hixie> | i can't even get into hixie.ch to see what's actually happening |
| 02:39 | <Hixie> | wtf |
| 02:40 | <jamesr_> | hixie.ch down? |
| 02:40 | <jamesr_> | is it all the same dreamhost? |
| 02:40 | <Hixie> | same server as whatwg.org |
| 02:40 | jamesr_ | was trying to make a testcase :( |
| 02:40 | <Hixie> | i could increase the virtual host allocation briefly |
| 02:40 | <Hixie> | but that sounds like work |
| 02:42 | <MikeSmith> | not sure it was actually my bad |
| 02:42 | <MikeSmith> | but the timing seems more than coincidental |
| 02:43 | <MikeSmith> | please call me El Destructo |
| 02:43 | <Hixie> | i'll check the logs if i remember later when it comes back up |
| 02:43 | <MikeSmith> | k |
| 02:43 | <MikeSmith> | twitter regexp lovebomb |
| 02:44 | <MikeSmith> | jwz would be proud |
| 02:44 | <Hixie> | clearly everyone is just like "holy crap, hixie started editing again". :-P |
| 02:44 | <Hixie> | bbiab |
| 02:47 | <MikeSmith> | "Hixie: The Resurrection" |
| 02:48 | <MikeSmith> | now with more regexp |
| 02:59 | <MikeSmith> | load average: 10.79, 8.84, 8.36 |
| 03:32 | <Hixie> | well looks like the bulk of the load i'm getting is from 2607:f298:1:105::72a:42f3 |
| 03:34 | <Hixie> | no idea who that is but i'm guessing html5.org maybe? |
| 03:34 | <MikeSmith> | eh? |
| 03:34 | <MikeSmith> | oh |
| 03:34 | <MikeSmith> | ipv6 |
| 03:36 | <MikeSmith> | Hixie: nope |
| 03:36 | <MikeSmith> | I think |
| 03:37 | <MikeSmith> | html5.org is 2607:f298:1:103::d75:8e6d |
| 03:37 | <MikeSmith> | or claims to be at least |
| 03:37 | <Hixie> | same subnet then |
| 03:37 | <MikeSmith> | ah, OK |
| 03:37 | <Hixie> | whatever host it is is hammering me doing OPTIONS * on hixie.ch |
| 03:37 | <MikeSmith> | I dunno jack about reading ipv6 addresses |
| 03:38 | <Hixie> | which is what i see when someone does a lot of svn stuff |
| 03:38 | <MikeSmith> | hmm |
| 03:38 | <MikeSmith> | of course web-apps-tracker calls svn |
| 03:38 | <MikeSmith> | on whatever other host |
| 03:40 | <Hixie> | oh wait |
| 03:40 | <Hixie> | i'm an idiot |
| 03:40 | <Hixie> | 2607:f298:1:105::72a:42f3 is hixie.ch, and an OPTIONS * request from that IP on this dashboard means nothing's happening -_- |
| 03:40 | <Hixie> | i'm gonna go now before i say anything else dumb. |
| 03:43 | <MikeSmith> | heh |
| 03:43 | <MikeSmith> | seems to be OK now at least |
| 03:44 | <MikeSmith> | nerd interest now fully turned back to rage comic subreddit |
| 03:45 | <MikeSmith> | Hixie: btw, go see Paul if you haven't already |
| 03:45 | <MikeSmith> | my daughter talked me into going to see it with her |
| 03:45 | <MikeSmith> | a |
| 03:45 | <MikeSmith> | and we laughed all the way through |
| 03:46 | <MikeSmith> | her because she's a 13 year old and me because I'm mentally still pretty much still 13 |
| 03:48 | <MikeSmith> | Hixie: and in other news, when you have time please glance through http://dvcs.w3.org/hg/url/raw-file/f1e294d18a68/Overview.html and let me know if it looks like it's headed in the right direction |
| 03:52 | <Hixie> | MikeSmith: did you end up editor? |
| 03:53 | <MikeSmith> | well, shadow editor for now |
| 03:53 | <Hixie> | poor sucker |
| 03:53 | <MikeSmith> | heh |
| 03:53 | <Hixie> | from a quick glance it looks good |
| 03:53 | <MikeSmith> | so far I just attempted to melt together what Adam had done in the IETF draft and what you had in the spec and what Adam had for the URL API |
| 03:54 | <MikeSmith> | the parsing algorithms are from Adam's IETF draft rather than what was in the spec |
| 03:54 | <MikeSmith> | because I assume Adam has good reasons for doing those differently |
| 03:54 | <Hixie> | two quick comments if you're editing tha spec: |
| 03:54 | MikeSmith | nods |
| 03:54 | <Hixie> | 1. there's a bunch of bugs in the w3c bugzilla to look at |
| 03:54 | <Hixie> | look for things assigned to adam or with "url" in the whiteboard or summary |
| 03:55 | <Hixie> | 2. one of the core design decisions to make is whether any arbitrary string should always parse or whether it's possible for parsing to fail |
| 03:55 | <Hixie> | different browsers differ on #2 |
| 03:55 | <Hixie> | it would be good to make an executive decision on 2 early on since it would inform feedback on the algorithms |
| 03:55 | <Hixie> | and will affect other specs the most |
| 03:56 | <MikeSmith> | yeah, knew that fail part from talking with Adam and perusing the Webkit test cases |
| 03:56 | <MikeSmith> | OK |
| 03:56 | <MikeSmith> | (about the executive decision) |
| 03:56 | <Hixie> | also if you can sever any ties to the URL RFCs i think everyone's life in #whatwg will be easier on the long run :-) |
| 03:56 | <MikeSmith> | oh |
| 03:57 | <MikeSmith> | OK |
| 03:57 | <MikeSmith> | will try to do that then |
| 03:57 | <Hixie> | well i expect that's politically too difficult to do |
| 03:57 | <MikeSmith> | well, we can take that as a challenge, then, I guess |
| 03:57 | <Hixie> | hah |
| 03:57 | <Hixie> | it would mean adding a whole bit about authoring conformance criteria and semantics |
| 03:57 | <Hixie> | so it's also quite a bunch of work |
| 03:57 | <MikeSmith> | oh |
| 03:58 | <MikeSmith> | that doesn't sound like fun |
| 03:58 | <Hixie> | it would have the advantage of finishing the dichotomy of IRI vs URI |
| 03:58 | <MikeSmith> | yeah |
| 03:58 | <Hixie> | yeah it wouldn't be fun |
| 03:58 | <Hixie> | why do you think i deferred to the rfcs :_P |
| 03:58 | <MikeSmith> | heh |
| 03:59 | <MikeSmith> | plan is for me and Anne to work on this when he gets back next month |
| 03:59 | <MikeSmith> | hopefully he might find it slightly more fun than working on the Encodings spec |
| 04:00 | <Hixie> | hah |
| 04:00 | <Hixie> | it's like a hot potato |
| 04:01 | <MikeSmith> | heh |
| 04:01 | <MikeSmith> | yeah, hoping to just end it |
| 04:02 | <MikeSmith> | it's been fun but 3 years of hot potato is probably enough .. the novelty of it wears off a bit |
| 04:02 | <Hixie> | 3? |
| 04:02 | <Hixie> | we first started speccing url parsing more than half a decade ago |
| 04:02 | <Hixie> | no? |
| 04:02 | <Hixie> | maybe it just feels that long |
| 04:05 | <MikeSmith> | I mean the more recent related drama with the Bearded Ones about it |
| 04:05 | <MikeSmith> | metaphorically (neck)bearded |
| 04:05 | <MikeSmith> | which I metaphorically count myself to be a junior member of |
| 04:06 | <MikeSmith> | I've just not yet achieved anywhere near that level of ability for how-many-angels-on-the-head-of-a-pin discussion |
| 04:07 | <MikeSmith> | anyway |
| 04:07 | <MikeSmith> | we will make it right eventually |
| 04:07 | <MikeSmith> | along with other needed stuff |
| 04:09 | <MikeSmith> | though hoping that Anne's current enthusiasm doesn't wane |
| 04:09 | <MikeSmith> | and he doesn't get distracted by taking up model rocketry or something |
| 04:11 | <MikeSmith> | though I have to say Anne has a way of doing spec writing that takes it down to the bone |
| 04:11 | <MikeSmith> | Hixie: your specs have some personality |
| 04:11 | <MikeSmith> | one can feel the love |
| 04:30 | <MikeSmith> | Hixie: ain't finding any open bugs assigned to Adam except https://www.w3.org/Bugs/Public/show_bug.cgi?id=10213 |
| 04:31 | MikeSmith | wonders if Adam might be using multiple addresses |
| 04:42 | <jamesr> | hmm. question: i have a page.html and page.css. in page.html within an inline <svg> element i have a set of <pattern>s defined in a <defs> and a few <rect>s that are visible. in page.css, i have fill:url(#patternId) set for a few rect |
| 04:42 | <jamesr> | the patterns show up in webkit but not in gecko or presto |
| 04:43 | <jamesr> | is this user error or expected? |
| 04:43 | <jamesr> | specifying fill:blue or fill:hsl(5, 6, 6) works in all browsers |
| 04:44 | <shepazu> | jamesr: have you tried fill:url(page.html#patternId) ? |
| 04:44 | jamesr | tries |
| 04:45 | <jamesr> | aha! that works |
| 04:45 | <jamesr> | what if i wanted to move these patterns to an external svg to reuse? would that work? |
| 04:45 | <jamesr> | also, where am i supposed to look this sort of thing up? |
| 04:45 | <shepazu> | nasty spec bug in CSS around relative uris in property values… I think roc's proposed a fix for that |
| 04:46 | <roc> | jamesr: that is a webkit bug I guess |
| 04:46 | <jamesr> | yeah looks like WK uses the wrong base |
| 04:46 | <shepazu> | jamesr: it should work, but I think WebKit is still waiting on a patch for that… and I don't know the status of external resources in SVG for Gecko |
| 04:47 | <roc> | external resources work fine in Gecko |
| 04:47 | <shepazu> | great |
| 04:47 | <roc> | in fact, that's what page.html#patternId will do |
| 04:47 | <roc> | I think it will reload page.html as an external resource |
| 04:47 | <shepazu> | ugh |
| 04:47 | <roc> | but I'm not 100% sure about that |
| 04:47 | <roc> | anyway |
| 04:47 | <jamesr> | roc: gecko has an ugly bug with tiling |
| 04:48 | <roc> | jamesr: you can put your resources in a style-resources.html file and reference that |
| 04:48 | <roc> | it would be kinda neat if you could have an XML island in your CSS file and reference resources there |
| 04:48 | <roc> | tiling bug? do tell |
| 04:48 | <jamesr> | ah, presto has the cracks too. i have a repeated diagnoal line |
| 04:49 | <jamesr> | aha! it only looks good in chrome because we don't AA the line at all |
| 04:49 | <jamesr> | in gecko/presto the line is antialiased and the ends of the line segment look different from the middle |
| 04:49 | <roc> | :-) |
| 04:49 | <jamesr> | so it doesn't tile correctly |
| 04:50 | <jamesr> | not sure if the way patterns repeat is supposed to be "raster a square, repeat" or something logical where the line is supposed to be continuous across tiles |
| 04:50 | <roc> | I believe it's the former |
| 04:50 | <jamesr> | do SVG specs punt on antialiasing behavior like CSS? |
| 04:50 | <jamesr> | and HTML? |
| 04:51 | <roc> | yes |
| 04:51 | <roc> | you don't really want rasterization to be locked into a spec, do you? |
| 04:53 | <jamesr> | as an implementor hell no |
| 04:55 | <jamesr> | well hurray! now my page works in gecko/presto but not webkit :) |
| 04:55 | <shepazu> | :( |
| 04:56 | <shepazu> | maybe you can patch webkit... |
| 04:56 | <roc> | if only you knew someone who was an awesome Webkit hacker! |
| 04:59 | <jamesr> | our svg impl is kind of ... unique |
| 04:59 | <jamesr> | pretty sure <text> in an external SVG still only works if the svg is loaded as an <object type="image/svg+xml"> |
| 05:00 | <shepazu> | jamesr: you might look at this bug https://bugs.webkit.org/show_bug.cgi?id=12499 |
| 05:00 | <shepazu> | which someone is patching |
| 05:00 | <jamesr> | that's a scary low bug number |
| 05:01 | <jamesr> | my use case isn't related to <use> is it? |
| 05:02 | <shepazu> | not directly, probably not at all, but the external refs thing is what I was thinking of |
| 05:02 | <jamesr> | activity as of last week though |
| 05:05 | <MikeSmith> | didn't now webkit actually supported font-feature-settings. had thought it was just in gecko (and now trident) |
| 05:06 | <MikeSmith> | though apparently not supported on osx yet |
| 05:06 | <MikeSmith> | https://bugs.webkit.org/show_bug.cgi?id=69826#c0 |
| 05:08 | <jamesr> | hm, but it works elsewhere? |
| 05:11 | <MikeSmith> | jamesr: yeah, apparently |
| 05:11 | <MikeSmith> | didn't know til I saw a G+ posting from IET about it today |
| 05:12 | <MikeSmith> | https://plus.google.com/108784309450471378435/posts/ZywfbW1MmJj |
| 05:13 | <jamesr> | is the syntax from http://blogs.msdn.com/b/ie/archive/2012/01/09/css-corner-using-the-whole-font.aspx for real? |
| 05:13 | <jamesr> | "ss06" on; ? |
| 05:13 | <jamesr> | -ms-font-feature-settings: "c2sc" 1,"smcp" 1; |
| 05:15 | MikeSmith | compares with https://developer.mozilla.org/en/CSS/-moz-font-feature-settings |
| 05:16 | <MikeSmith> | hmm |
| 05:16 | <MikeSmith> | su |
| 05:16 | <jamesr> | so it's just a pass-through to the opentype API |
| 05:16 | <MikeSmith> | pp |
| 05:16 | <MikeSmith> | |
| 05:16 | <MikeSmith> | |
| 05:16 | <MikeSmith> | |
| 05:16 | <MikeSmith> | oops |
| 05:16 | <MikeSmith> | yeah, but spec says the whole list is supposed to be quoted |
| 05:18 | <MikeSmith> | actually, no it doesn't |
| 05:18 | <MikeSmith> | <feature-tag-value> = <string> [ <integer> | on | off ]? |
| 05:18 | <MikeSmith> | and example: font-feature-settings: "smcp", "swsh" 2; |
| 05:19 | <MikeSmith> | at least if that's actually the current spec |
| 05:20 | <MikeSmith> | which maybe it's ain't since it's under TR/ |
| 05:20 | MikeSmith | compares it to http://dev.w3.org/csswg/css3-fonts/ |
| 05:20 | <MikeSmith> | where's nattokirai when you need him? |
| 05:22 | <nattokirai> | hey |
| 05:22 | <MikeSmith> | yeah, editor's draft has it the same |
| 05:22 | <MikeSmith> | http://dev.w3.org/csswg/css3-fonts/#ltfeature-tag-valuegt |
| 05:22 | <MikeSmith> | nattokirai: hey man |
| 05:22 | <nattokirai> | neat, ms has font feature support |
| 05:22 | <MikeSmith> | Y NOT GECKO MATCH TEH SPEC? |
| 05:22 | <nattokirai> | but note that moz uses *old* syntax |
| 05:22 | <MikeSmith> | ah |
| 05:22 | <MikeSmith> | well |
| 05:22 | <nattokirai> | old syntax: font-feature-settings: "liga=1"; |
| 05:23 | <MikeSmith> | yar |
| 05:23 | <nattokirai> | new syntax: font-feature-settings: "liga" 1; |
| 05:24 | <nattokirai> | jamesr: it's a pass-thru to the shaping API |
| 05:24 | <nattokirai> | in the MS case, DirectWrite, in our case, the Harfbuzz shaping library |
| 05:28 | <nattokirai> | jamesr: but keep in mind that's the low-level feature mechanism |
| 05:28 | <nattokirai> | jamesr: there are higher-level properties that will expose access to all the commonly used features |
| 05:29 | <jamesr> | i don't know anything about shaping engines. it does look pretty ugly, though |
| 05:29 | <nattokirai> | yup |
| 05:30 | <nattokirai> | the higher-level properties are more CSS-like |
| 05:30 | <nattokirai> | e.g. http://dev.w3.org/csswg/css3-fonts/#font-variant-ligatures-prop |
| 05:31 | <nattokirai> | jamesr: note that in the spec, these can also be set within @font-face rules |
| 05:32 | <nattokirai> | jamesr: as a form of per-font default |
| 06:11 | <hsivonen> | annevk: do you know the answer to bz's question: https://bugzilla.mozilla.org/show_bug.cgi?id=716579#c8 ? |
| 06:14 | <MikeSmith> | nattokirai: so gecko doesn't rely on Core Text? |
| 06:14 | <jamesr> | what's tantek go by here? |
| 06:15 | <nattokirai> | MikeSmith: currently only for Indic scripts |
| 06:15 | <MikeSmith> | jamesr: usually "ta ntek" |
| 06:15 | <MikeSmith> | nattokirai: OK |
| 06:15 | <jamesr> | hm, guess he's not around |
| 06:15 | <nattokirai> | b/c OSX only has AAT fonts for Indic |
| 06:15 | <MikeSmith> | oh |
| 06:15 | <MikeSmith> | jamesr: yeah, he's not here super often |
| 06:16 | <nattokirai> | where AAT == Apple Advanced Typography |
| 06:16 | <MikeSmith> | or at least has not been so much in the past couple years |
| 06:16 | <MikeSmith> | though been around more recently |
| 06:16 | <nattokirai> | which is in many ways better then OpenType but no longer supported by font vendors |
| 06:16 | <MikeSmith> | I see |
| 06:16 | <MikeSmith> | nattokirai: I asked because the Core Text dependency seems to be what's complicating the WebKit implementation |
| 06:17 | <nattokirai> | yup |
| 06:17 | <nattokirai> | that's the webkit mo, to do text shaping via the os |
| 06:18 | <nattokirai> | and coretext support for opentype is somewhat limited |
| 06:18 | <MikeSmith> | yeah, I saw that in Kenichi's comment as well |
| 06:19 | <nattokirai> | you're looking at a bug somewhere? |
| 06:19 | <MikeSmith> | yeah, https://bugs.webkit.org/show_bug.cgi?id=69826#c0 |
| 06:19 | <MikeSmith> | tradeoffs abound I guess |
| 06:20 | <MikeSmith> | it seems like the general Gecko approach is to minimize platform code as much as possible |
| 06:20 | <MikeSmith> | is that right? |
| 06:23 | <nattokirai> | well, not entirely but in this case it makes everything simpler to do the shaping ourselves |
| 06:24 | <nattokirai> | osx support is limited and android non-existent |
| 06:24 | <MikeSmith> | ah |
| 06:25 | <nattokirai> | in particular, we wanted to avoid uniscribe on windows for security reasons |
| 06:25 | <nattokirai> | related to downloadable fonts |
| 06:25 | <MikeSmith> | oh |
| 06:25 | <jamesr> | like http://www.computerworld.com/s/article/9221498/Duqu_exploits_same_Windows_font_engine_patched_last_month_Microsoft_confirms ? |
| 06:26 | <MikeSmith> | hmm |
| 06:27 | <MikeSmith> | jamesr: did MS actually patch that yet or not? |
| 06:27 | <jamesr> | i dunno, i don't follow that stuff terribly closely |
| 06:27 | <MikeSmith> | OK |
| 06:28 | <jamesr> | some people @google built http://code.google.com/p/ots/ to try to mitigate those sorts of issues on windows |
| 06:28 | MikeSmith | takes a look |
| 06:28 | <MikeSmith> | hey it's tantek |
| 06:28 | <MikeSmith> | maybe |
| 06:29 | <jamesr> | tantek: yo! |
| 06:29 | <tantek> | hey MikeSmith you called? |
| 06:29 | <MikeSmith> | jamesr was looking for you |
| 06:29 | <jamesr> | just fyi and you may already know this, but google-chrome-browser.com is nothing to do with us |
| 06:30 | <tantek> | yeah it looked a little spammy but shall we say it's been hard to search for official chrome docs recently (perhaps due to the demotion of chrome sites due to the paid blogging fiasco?) |
| 06:30 | <jamesr> | i think only google.com/chrome was demoted, our developer documentation shouldn't have been hit |
| 06:31 | <tantek> | or perhaps there is no chrome documentation of beforeload |
| 06:31 | <jamesr> | most likely |
| 06:32 | <MikeSmith> | so that site is from whoever https://twitter.com/#!/ChromeBrowser is |
| 06:33 | <MikeSmith> | which twitter account I seem to remember doing something obnoxious last year |
| 06:33 | <MikeSmith> | like, anti-Firefox and/or anti-IE trolling |
| 06:34 | <jamesr> | people spreading information is great but i don't want people getting confused, especially since it looks like deep links don't have the disclaimer that google-chrome-browser.com does |
| 06:39 | <MikeSmith> | cool to see mention there for Boris Zbarsky's $1000 security-bug bounty |
| 06:39 | <MikeSmith> | seems like it's not the first time he found one |
| 06:41 | <MikeSmith> | jamesr: so can OpenType sanitization be performed fast enough that it could be done in the browser? |
| 06:42 | <annevk> | hsivonen: replied |
| 06:44 | <jamesr> | MikeSmith: i'm not intimately familiar with it, but i don't think that this would be a super hot path |
| 06:44 | <MikeSmith> | I see |
| 06:46 | <hsivonen> | annevk: thanks |
| 06:52 | hsivonen | wonders how different UI Spanish is between Argentina, Chile and Mexico or how different UI Bengali is between India and Bangladesh |
| 06:52 | <hsivonen> | or how different UI English is between the UK and South Africa |
| 07:05 | <annevk> | someone should HTTP behavior for data URLs |
| 07:05 | <annevk> | define /\ |
| 07:13 | <annevk> | http://blog.fonts.com/2012/01/09/monotype-imaging-and-google-collaborate-to-make-web-fonts-better/ |
| 07:14 | <annevk> | http://www.w3.org/Submission/MTX/ RF |
| 07:15 | <annevk> | lot of C code there |
| 07:26 | <hsivonen> | I'm disappointed that http://en.wikipedia.org/wiki/Rule_110 doesn't link to a List of Numbered Rules |
| 07:28 | <hsivonen> | annevk: so we had to get everyone to deploy WOFF first |
| 07:30 | <annevk> | they seemed to have some kind of alliance with Microsoft before, kind of surprised this is announced jointly with Google |
| 07:30 | <annevk> | prolly some money involved |
| 07:48 | <MikeSmith> | that looks like a genuine RF-for-any-possible-purpose license |
| 07:48 | <tantek> | hsivonen - here's a list of numbered rules: http://en.memory-alpha.org/wiki/Rules_of_Acquisition |
| 07:52 | <annevk> | MikeSmith: yeah, looks great, although is 15% savings worth the added complexity? especially with a format documented as a bunch of C code... |
| 07:52 | <annevk> | MikeSmith: reminds me of webm, wonder if that has improved yet... |
| 07:53 | <tantek> | MikeSmith, would be nice if folks would simply pick a standard license like CC0/OWFa rather than something "that looks like a genuine RF-for-any-possible-purpose license" |
| 07:53 | <MikeSmith> | yeah, there's no real spec for webm yet afaik |
| 07:53 | <MikeSmith> | tantek: yeah, but large companies don't seem to do that |
| 07:54 | <annevk> | heh, twitter brought html5.org down? |
| 07:54 | <MikeSmith> | I guess the the existing webm implementation can count as the spec |
| 07:54 | <annevk> | ooh, it brought svn.whatwg.org down |
| 07:54 | <MikeSmith> | and for MTX too |
| 07:54 | <MikeSmith> | annevk: yeah, not sure but seems like |
| 07:54 | <MikeSmith> | due to the timing |
| 07:54 | <Hixie> | i didn't investigate closely what happened on my end |
| 07:55 | <Hixie> | so not sure |
| 07:55 | <annevk> | hmm, I wonder if caching is still turned on |
| 07:56 | hsivonen | had forgotten about those numbered rules |
| 08:03 | <annevk> | hsivonen: fwiw; we've a patch for BOM overriding HTTP already |
| 08:03 | <annevk> | hsivonen: we = Opera |
| 08:04 | <hsivonen> | annevk: ok. I guess I should get around to writing one, too |
| 08:04 | <hsivonen> | annevk: for all file formats or just HTML? |
| 08:05 | <annevk> | hsivonen: all |
| 08:06 | <annevk> | hsivonen: though unlike WebKit we'd still allow user overrides |
| 08:08 | <MikeSmith> | man, twitter says 100+ people retweeted that e-mail regexp tweet |
| 08:08 | <MikeSmith> | in the wrong hands twitter is a dangerous weapon |
| 08:08 | <divya> | it is dangerous in all cases |
| 08:09 | <MikeSmith> | divya: twitter luvs you |
| 08:09 | <MikeSmith> | you should love it right back |
| 08:09 | <divya> | hahahah |
| 08:10 | <divya> | its so trivial to RT or say anything in 140 chars :( |
| 08:10 | <MikeSmith> | that's the fun of it! |
| 08:11 | <divya> | :) |
| 08:11 | Philip` | looked at MTX a while ago, and decided that he hated it, since reverse-engineering incomplete fragments of untestable pseudo-C code doesn't sound like fun |
| 08:11 | <MikeSmith> | there you go, more fun |
| 08:12 | <MikeSmith> | in other news, there's now an actual Browser Automation API spec |
| 08:12 | <MikeSmith> | aka WebDriver |
| 08:12 | <MikeSmith> | (which name probably really needs to actually change to Browser Automation API) |
| 08:12 | <MikeSmith> | http://dvcs.w3.org/hg/webdriver/raw-file/tip/webdriver-spec.html |
| 08:14 | <Ms2ger> | Bah, respec |
| 08:14 | <MikeSmith> | respec needs love too |
| 08:14 | <Ms2ger> | the Web DOM Core specification ([DOM-LEVEL-3-CORE]). |
| 08:14 | <MikeSmith> | it needs our help to make itself better |
| 08:14 | <Ms2ger> | What the hell |
| 08:15 | <MikeSmith> | yeah, that needs to be fixed |
| 08:15 | <MikeSmith> | I'm pretty sure the default respec biblio thing is generated from a file at W3C that only lists TR docs |
| 08:15 | <wilhelm> | It's a very, very early draft. We have a F2F tomorrow where we'll do some real work on it. (c: |
| 08:16 | <MikeSmith> | hey it's wilhelm |
| 08:16 | <wilhelm> | Any and all feedback is very welcome. |
| 08:16 | <MikeSmith> | hope somebody and skype me into the meeting |
| 08:17 | wilhelm | installs Skype. |
| 08:19 | <wilhelm> | Why do I need a native application for this? Hasn't someone made a web app yet? :P |
| 08:19 | <wilhelm> | Wait, perhaps Google+ works. |
| 08:20 | <izhak> | Guys, sorry for very simple stupid question, but is HTML5 is valid XML? |
| 08:20 | <Ms2ger> | No |
| 08:20 | <tantek> | depends on your HTML5 |
| 08:20 | <Hixie> | depends what you mean by "HTML5" |
| 08:20 | <tantek> | :D |
| 08:21 | <tantek> | izhak, I've got some HTML5 that's also valid XML on my site. but it takes some amount of work to pull that off. |
| 08:21 | <izhak> | and what do I mean by HTML5 ? |
| 08:21 | <izhak> | :D |
| 08:21 | <jamesr> | there are documents in the intersection, i believe |
| 08:21 | <Hixie> | generally speaking unless you have a pretty important reason for it you shouldn't really worry about XML when you're doing HTML |
| 08:22 | <izhak> | Can HTML5 be determined by a DTD? |
| 08:22 | <Hixie> | what do you mean by "determined"? |
| 08:22 | <izhak> | Can we write such DTD which would be able to validate any HTML5 document? |
| 08:23 | <izhak> | Well, OK I think I understand now, I mean syntax.. HTML5 syntax |
| 08:24 | <jamesr> | no |
| 08:25 | <izhak> | Is there any simple answer to the question, why not to use XHTML which is perfectly deterministic? |
| 08:26 | <wilhelm> | HTML is also perfectly deterministic. |
| 08:26 | <wilhelm> | And doesn't come with the disadvantages of XML. |
| 08:26 | <izhak> | wilhelm: it's not, I'm a complete newbie but it's not exactly |
| 08:26 | <jamesr> | what do you mean? |
| 08:27 | <tantek> | izhak, XHTML does not confer any advantages over XML-well-formed HTML5. |
| 08:27 | <jamesr> | there's a deterministic algorithm to decide if a given string of bytes is a valid HTML document, if that's what you mean |
| 08:27 | <tantek> | izhak - longer explanation of that here: http://tantek.com/2010/302/b1/xhtml-dead-long-live-xml-valid-html5 |
| 08:27 | <izhak> | I mean it's very simple to find out relative layout of elements in XML, than in HTML |
| 08:28 | <tantek> | izhak, XML does not determine layout - CSS does that. |
| 08:29 | <izhak> | tantek: I mean nesting, closing etc |
| 08:29 | <tantek> | izhak - you can do the same with HTML5 - see my blog post. |
| 08:30 | <izhak> | Yeah, thanks, that was the text i was searching |
| 08:30 | <wilhelm> | MikeSmith: I think I got Google+ Hangouts to work. I'll invite you tomorrow morning. I'll be at the Google office 15-30 minutes before the meeting starts. |
| 08:30 | <MikeSmith> | wilhelm: OK |
| 08:30 | <izhak> | HTML 4 parser is a mess, while XML parser is pretty simple, tell me does HTML5 allow that mess also? |
| 08:31 | <Philip`> | Writing an HTML5 parser is more work than writing an XML parser (since HTML5 has more parsing rules), but very few people need to write parsers (and everyone else can reuse one written by someone else), whereas many people need to output HTML code, and guaranteeing the output is well-formed XML is extremely hard, so it doesn't make sense to choose the approach that simplifies parsers |
| 08:31 | <izhak> | Or it's as regorous as XML is |
| 08:32 | <Hixie> | Philip`: unless you're implementing the script stuff, i think it might actually be less work to do an HTML parser since with XML you don't have a spec to follow but instead have to determine it from the validity description |
| 08:32 | <Philip`> | izhak: HTML4 parsing is undefined; HTML5 parsing is specified precisely in complete detail |
| 08:32 | <jamesr> | writing an XML parser isn't exactly trivial |
| 08:33 | <izhak> | jamesr: relatively to html4's it's just a toytaks |
| 08:33 | <izhak> | *toytask |
| 08:33 | <jamesr> | given that html4 parsing is undefined, it's not really a valid comparison |
| 08:34 | <izhak> | jamesr: actually it is cause we actually have html4 parsers |
| 08:34 | <jamesr> | not interoperable ones |
| 08:34 | <Hixie> | technically HTML4 parsing is defined, in SGML. not that such a parser would actually work on the web ;-) |
| 08:34 | <Philip`> | (HTML5's specified parsing algorithm is a crazy mess, but that doesn't matter since you can just implement what the spec says and not worry about it) |
| 08:36 | <izhak> | Unfortunately we've got lot's of html4 pages all over the web and they seem not to be going to leave in a nearest future as in farest |
| 08:37 | <Philip`> | Fortunately HTML5 parsers can parse all those HTML4 pages |
| 08:37 | <Hixie> | the HTML standard today handles all previous versions of HTML (or more precisely, all existing HTML content) |
| 08:37 | <izhak> | So it's not rigorous and allow writing html4 stuff |
| 08:38 | <izhak> | And if it allows people will use that |
| 08:38 | <Philip`> | If by "not rigorous" you mean "not fragile", then yes :-) |
| 08:38 | <Hixie> | what do you mean by "rigorous"? |
| 08:38 | <izhak> | vicious circle |
| 08:38 | <Hixie> | and what do you mean by "allow"? |
| 08:39 | <izhak> | I mean determinism |
| 08:39 | <izhak> | Oh I see, my English is very bad |
| 08:40 | <izhak> | rigorous = strict, well defined, deterministic |
| 08:42 | <izhak> | Anyways, it seems that I understand now, w3c is going by a strict, ideal but unachievable way to switch the world to strict xml, and you guys going by more realistic way |
| 08:42 | <Philip`> | The HTML5 parsing algorithm isn't strict (it allows any random stream of bytes as input and won't produce an error and abort) but is well defined and deterministic |
| 08:43 | <Philip`> | (There are multiple implementations that produce precisely the same output for any random stream of bytes) |
| 08:43 | <roc> | modulo bugs |
| 08:44 | <izhak> | Philip`: hmm, great |
| 08:44 | Ms2ger | wonders why Brad Kemper used GB2312 |
| 08:45 | <izhak> | Philip`: does html 5 parser do charset normalization? |
| 08:47 | <izhak> | "The HTML5 parsing algorithm isn't strict (it allows any random stream of bytes as input and won't produce an error and abort) but is well defined and deterministic" - cool! This sentence makes me feel good:D |
| 08:48 | <jamesr> | roc: in gecko when script force a style resolution+layout, do y'all ever resolve styles for only a subset of the document or do you always resolve the whole thing? |
| 08:49 | <Ms2ger> | And <di>, woot |
| 08:49 | <roc> | the whole thing |
| 08:50 | <roc> | I can think of lots of optimizations I'd rather try than changing that :-) |
| 08:50 | <Philip`> | izhak: What do you mean by charset normalization? |
| 08:51 | <jamesr> | seems scary (and normally not worth it) |
| 08:51 | <Philip`> | izhak: (The parser basically just figures out the character encoding somehow and decodes the input, and then the rest of the algorithm deals purely with Unicode strings) |
| 08:51 | <roc> | I'm sure you could create a benchmark where it would be totally worth it |
| 08:51 | <roc> | please don't! |
| 08:51 | <izhak> | Philip`: yes, that what I meant:) |
| 08:52 | <izhak> | Philip`: is it defined in what particular encoding? |
| 08:52 | <izhak> | I mean utf-8 or utf-16 |
| 08:52 | <wilhelm> | Ms2ger: Re: the DOM3 reference, please prod the editors and/or public-test-infra⊙wo (c: |
| 08:53 | <Ms2ger> | Oh, and "Web DOM Core" is two Microsoft interventions out of date |
| 08:55 | <Philip`> | izhak: http://whatwg.org/html#determining-the-character-encoding says how to determine what encoding to use when decoding |
| 08:55 | <izhak> | Philip`: thanks |
| 09:02 | <jgraham> | wilhelm: Isn't it public-browser-tools-testing that one should prod? |
| 09:03 | <jgraham> | (gotta love the short and easy to remember email addresses) |
| 09:04 | hsivonen | notes that a conforming XML parser has to support all the internal subset goo |
| 09:05 | <wilhelm> | jgraham: Eh, probably. |
| 09:05 | wilhelm | subscribes. :P |
| 09:06 | <hsivonen> | izhak: for both HTML and XML, character encodings are a loophole that you can drive a truck through |
| 09:07 | <hsivonen> | izhak: in the sense that neither HTML nor XML parsing is well-defined when an ill-defined character encoding is used |
| 09:07 | <hsivonen> | and the set of available encodings isn't locked down by either spec |
| 09:08 | <hsivonen> | see http://www.tbray.org/ongoing/When/200x/2003/10/17/UTF8-plus for an example of one of the editors of XML trying to exploit the definitional loophole left in XML |
| 09:08 | <izhak> | hsivonen: I do believe that what you say is far from real world nowadays |
| 09:09 | <hsivonen> | izhak: see the log of this channel for annevk's findings of differences in how different browser implement various character encodings |
| 09:11 | <hsivonen> | izhak: once the character encoding has been dealt with, HTML parsing is actually more deterministic than XML parsing, because XML parsing still involves an optional feature that HTML doesn't have: whether or not external entities are processed |
| 09:13 | <hsivonen> | since both HTML and XML share the character encoding hole but HTML doesn't have the optional feature that external entities are for XML, it's a myth these days that XML is somehow more deterministic |
| 09:15 | <izhak> | hsivonen: tha fact that html5 parses both html4 mess and xml still being deterministic is enough for me to love it:), concerning what you say, I think charset problems are not really a problems today, all stuff is quite worked over, waddya think |
| 09:16 | <hsivonen> | izhak: unfortunately, for implementations, charset problems are still problems today. for authors, they aren't when authors stick to UTF-8 |
| 09:16 | <hsivonen> | and declare that they are sticking to UTF-8 |
| 09:18 | <izhak> | well, that's what I didn't think about, thanks, I'll learn |
| 09:39 | <izhak> | Hah, eventually it became clear to me.. That's just a question of "Break or not to break?" (on error). Such a thin border, and yes what to do if not to break:). |
| 09:39 | <izhak> | tantek, cool article thanks |
| 13:02 | <MikeSmith> | matjas: I'm wondering what mail clients actually support IDN mail addresses yet |
| 13:02 | <MikeSmith> | re: https://www.w3.org/Bugs/Public/show_bug.cgi?id=15489 |
| 13:03 | <matjas> | what do you mean by “support”? displaying the email address’s Unicode form as opposed to its Punycoded ASCII representation? |
| 13:03 | <matjas> | or, like, “work” |
| 13:03 | <MikeSmith> | yeah, "work" -- as in, allowing you to actually send an e-mail message to such an address |
| 13:04 | <MikeSmith> | allowing to put the e-mail address into a To field in the UI for the mail client |
| 13:53 | <annevk> | replacing a SSD apparently takes a while... |
| 13:53 | <annevk> | meanwhile using my old T60 with Ubuntu |
| 13:53 | <annevk> | good times |
| 13:53 | <annevk> | except for the keyboard, that sucks |
| 14:18 | <MikeSmith> | matjas: every mail client I've tried so far barfs on IDN email addresses when I put them in a To field |
| 14:18 | <MikeSmith> | annevk: T60 is IBM laptop? |
| 14:19 | <annevk> | effectively, it's from lenovo |
| 14:19 | <MikeSmith> | ah yeah |
| 14:19 | <MikeSmith> | you like the keyboard? |
| 14:19 | <annevk> | but carries the old IBM ThinkPad logo |
| 14:19 | <annevk> | it's kind of broken, that's the problem :) |
| 14:20 | <annevk> | Enter no longer gives tactile feedback for instance |
| 14:22 | <MikeSmith> | oh |
| 14:22 | <MikeSmith> | just pretend it's a soft keyboard, and then it's fine |
| 14:22 | <matjas> | MikeSmith: I don’t really see why that matters though. Why wait until email clients fix their stuff to fix the HTML spec? |
| 14:22 | <annevk> | I guess, it's no big deal |
| 14:22 | <matjas> | MikeSmith: are you thinking of <form action=mailto:>? |
| 14:23 | <MikeSmith> | matjas: no, I'm thinking input@type=email |
| 14:25 | <MikeSmith> | matjas: and maybe I'm misunderstanding your point but I'm not sure there's anything that needs fixing in the spec because the current restrictions are by design |
| 14:25 | <MikeSmith> | that is, not an oversight |
| 14:26 | <matjas> | MikeSmith: What’s the reason foo@mañana.com is invalid while foo⊙xc isn’t? Any pointers? |
| 14:26 | <MikeSmith> | no pointers |
| 14:26 | <MikeSmith> | maybe others on the channel know |
| 14:27 | <MikeSmith> | annevk maybe |
| 14:27 | <MikeSmith> | as far as the rationale |
| 14:29 | <annevk> | not sure, I always found it slightly confusing and I think there might still be issues with both url and email |
| 14:34 | <MikeSmith> | I assume the rationale is that it doesn't make much sense for the spec and browsers to recognize IDN e-mail addresses as valid for input@type=email if existing MUAs don't recognize them in their corresponding input mechanisms |
| 14:35 | <MikeSmith> | and I don't know why existing mail clients don't yet recognize IDN e-mail in their input fields, I don't know |
| 14:36 | <MikeSmith> | but I'd assume it's not due just to collective laziness |
| 15:07 | <Ms2ger> | MikeSmith, the idea for input type=email is that the actual value is the punycode version, and the UA shows the Unicode version in its UI |
| 15:07 | <zewt> | i'd think it'd make more sense to have the unicode version in mailto:, and have the MUA encode it when it receives it |
| 15:08 | <zewt> | eg. aim for users never having to type or see the encoded version |
| 15:09 | <MikeSmith> | Ms2ger: really? |
| 15:09 | <Ms2ger> | See the spec, there's a note |
| 15:09 | MikeSmith | looks |
| 15:10 | <annevk> | Ms2ger: is it the same for URLs? |
| 15:10 | <zewt> | (just like you should be able to use unicode in URLs without encoding it first--no idea if you can, not familiar enough with that yet) |
| 15:10 | <Ms2ger> | Yeah |
| 15:10 | <zewt> | (but if you can't, I'd consider it a dead-end broken i18n effort...) |
| 15:11 | <annevk> | I don't think you can trust the MUA to handle non-ASCII |
| 15:11 | <annevk> | but the MUA can translate it back in its own UI |
| 15:11 | <MikeSmith> | so I find "User agents may transform the value for display and editing; in particular, user agents should convert punycode in the value to IDN in the display and vice versa." |
| 15:11 | <zewt> | annevk: the browser can always encode it before handing it off to the MUA |
| 15:11 | <zewt> | at navigation time, i mean |
| 15:11 | <zewt> | i guess copy-to-clipboard is less clear |
| 15:12 | <zewt> | (but also less important) |
| 15:16 | <zewt> | (the one or two times I've seen people using IDN addresses in email it made me think "that guy's using a feature long before it's ready", since it just showed up encoded in gmail, in quote headers, etc) |
| 15:18 | <zewt> | by the way, doesn't Android have a special keyboard layout for email addresses? (eg. @ available without shifting) |
| 15:19 | <zewt> | that seems like a more day-to-day useful use case for type=email |
| 15:19 | <MikeSmith> | Gmail doesn't actually let you enter IDN e-mail addresses into its UI |
| 15:19 | <annevk> | that's more of an inputmode feature tied to type=email |
| 15:20 | <annevk> | we don't want to introduce types for all the inputmode features we might have at some point |
| 15:23 | <zewt> | i'd expect it to depend on whether it's actually orthogonal to input mode or not |
| 15:24 | <zewt> | some input modes might make sense with textarea, I suppose |
| 15:52 | <bga_> | http://psicomatico.net/wp-content/uploads/2011/02/jqueryninja.jpg |
| 15:53 | <ksweeney> | haha |
| 15:53 | <ksweeney> | shouldn't the jquery ninja be chaining everything in there? |
| 15:54 | <bga_> | only who has black belt :) |
| 15:55 | smaug____ | so much prefers Javascript Ninja |
| 15:58 | <AryehGregor> | smaug____, you and I don't have to write complicated webpages that work in a large range of browsers on a regular basis. |
| 15:59 | <hsivonen> | Hixie: is it intentional that document.close() doesn't return early if document.close() has been called? |
| 15:59 | <hsivonen> | Hixie: that is, should each document.close() call attempt to tokenize? |
| 16:00 | <smaug____> | AryehGregor: yeah. Fortunately I don't need to use buggy and memory hungry js libraries to support multiple browsers |
| 16:00 | <AryehGregor> | Yep, me either. |
| 16:00 | <AryehGregor> | Lucky for us. |
| 16:01 | <jgraham> | AryehGregor: You have to write evil webpages that break a large range of bgrowsers though. Surely that must count for something? |
| 16:02 | <AryehGregor> | jgraham, yes. It would be much harder if I had to use jQuery. |
| 16:02 | <jgraham> | If you start using jQuery for testcases I might have to arrange to have you destroyed |
| 16:03 | <Ms2ger> | I'm going to submit some Mozilla tests that use $() just to get your blood pressure up :) |
| 16:03 | hsivonen | notes that Mozilla uses mochikit for tests and mochikit browser-sniffs |
| 16:03 | jgraham | knows :( |
| 16:04 | <Ms2ger> | jgraham, did that make it hard to steal our tests? :) |
| 16:18 | AryehGregor | adds tests for case-sensitivity of all the keywords used in CSS 2D Transforms |
| 16:22 | <AryehGregor> | It turns out having been a math major is really useful for anything involving 2D graphics. |
| 16:22 | <AryehGregor> | Like, knowing what linear transforms are and how they relate to matrices. |
| 16:23 | <AryehGregor> | Although I have yet to find a use for, say, Jordan normal forms or Perron's theorem. |
| 16:23 | <AryehGregor> | (Jordan normal forms are useless in practice because nondiagonalizable matrices have codimension zero in the space of matrices, I guess) |
| 16:26 | <jgraham> | Ms2ger: I don't know that we made a serious attempt to import lots of Mochitests |
| 16:30 | <gsnedders> | The attempt I made wasn't very serious, at least. |
| 16:56 | <dglazkov> | good morning, Whatwg! |
| 19:19 | <annevk> | XML Prague wants Docbook |
| 19:19 | <annevk> | euh |
| 19:19 | <annevk> | DocBook |
| 19:34 | <AryehGregor> | Suppose I have an inner div nested inside the outer div, and the inner div has a nonzero left/right margin. Is there any way to get their border boxes to coincide? |
| 19:34 | <AryehGregor> | (yes, I know this is a weird question) |
| 19:35 | <Ms2ger> | Negative.. Padding, no, that wouldn't work |
| 19:35 | AryehGregor | has developed doubts regarding the sanity of what he's currently trying to do |
| 19:35 | <AryehGregor> | Yeah, you can't have negative padding or border, so . . . |
| 19:35 | <AryehGregor> | Well, I could use relative positioning or something, of course. I mean in normal flow. |
| 19:35 | <Ms2ger> | Can you add a third div? :) |
| 19:35 | <AryehGregor> | Possibly. How would that help? |
| 19:36 | <Ms2ger> | Put it in between and give it negative margins |
| 19:36 | <AryehGregor> | Sneaky. |
| 19:36 | <Ms2ger> | The universal solution to CSS problems-- add another div |
| 19:37 | <AryehGregor> | It just occurred to me, though, that it's enough that the border boxes' *centers* coincide. I just want the transform-origin to be the same. |
| 19:37 | AryehGregor | tries that |
| 19:37 | <Ms2ger> | The centers shouldn't be hard |
| 19:38 | <Ms2ger> | Match up padding-left on the outer with margin-right on the inner, no? |
| 19:38 | <AryehGregor> | I'll just put a few pixels of padding on the outer divs. |
| 19:38 | <AryehGregor> | That should do it. |
| 19:38 | <AryehGregor> | (I actually have three nested divs) |
| 19:45 | <AryehGregor> | Oh, no, wait. I actually do need the border boxes the same actual size if I want percentages to work right. |
| 19:45 | <AryehGregor> | Or I can adjust those by hand. |
| 19:45 | <AryehGregor> | Feh. |
| 19:45 | <AryehGregor> | I guess that's what I'll do. |
| 19:46 | <AryehGregor> | Bah. Complicated. |
| 19:46 | AryehGregor | is getting a headache |
| 19:55 | <AryehGregor> | Okay, I finally think I have it. Sheesh. |
| 20:02 | <annevk> | hsivonen: with file upload and a custom scheme the file upload field jumps down visually in Opera when you hit validate |
| 20:02 | <annevk> | hsivonen: on validator.nu |
| 21:07 | <vishal3212> | hello |
| 21:35 | <jgraham> | Random observation of the day: half the web seems to be articles abou hiring developers, but I have never seen a single article abot hiring testers |
| 21:35 | <jgraham> | *about |
| 21:36 | <jgraham> | That isn't apropos anything in particular, it's just a thing I noticed |
| 23:07 | <TabAtkins> | In the XML serialization, are all the entities still defined and available? |
| 23:07 | <TabAtkins> | Or just the XML entities? |
| 23:36 | <Hixie> | TabAtkins: what other kinds of entities would you mean? |
| 23:36 | <Hixie> | TabAtkins: oh you mean like the mathml ones? |
| 23:37 | <Hixie> | TabAtkins: there's some DOCTYPEs you can specify that magically include the others |
