| 00:00 | <Hixie> | ok try grabbing it now |
| 00:00 | <Hixie> | http://www.whatwg.org/specs/web-apps/current-work/dev-index |
| 00:01 | <benschwarz> | Hixie: how brittle |
| 00:01 | <benschwarz> | and hard to read :) |
| 00:02 | <benschwarz> | still breaks… |
| 00:03 | <Hixie> | ah, it wants two elements in the head, because the "real" spec has a paragraph with the logo before the <header> :-) |
| 00:03 | <Hixie> | hold on |
| 00:05 | <Hixie> | regenning |
| 00:05 | <Hixie> | ready |
| 00:05 | <Hixie> | http://www.whatwg.org/specs/web-apps/current-work/dev-index |
| 00:10 | <benschwarz> | trying again |
| 00:13 | <benschwarz> | success |
| 00:13 | <benschwarz> | Hixie: http://developers.whatwg.org/ |
| 00:13 | <Hixie> | note: i just updated some of the mailman settings for the WHATWG list. Please let me know ASAP if you notice anything broken. Thanks! |
| 00:14 | <jamesr_> | benschwarz: shiny |
| 00:14 | <benschwarz> | jamesr_: a bit to go yet, but we're getting there |
| 00:14 | <Hixie> | I guess I should mark the "This section is non-normative" bits as class=impl |
| 00:16 | <Hixie> | well, maybe not |
| 00:17 | <Hixie> | the spec is still a spec, even though it's missing the UA bits |
| 00:18 | <Hixie> | benschwarz: http://developers.whatwg.org/introduction.html#typographic-conventions -- should I hide some of those from the dev version, or will they at some point get special styles again? |
| 00:18 | <benschwarz> | Hixie "This section is non-normative" doesn't mean anything to anyone… who isn't apart of the whole specification creation process |
| 00:19 | <benschwarz> | Hixie: I'll sort those, thanks for pointing them out |
| 00:20 | <Hixie> | while you're doing those, dropping all the styles means you've dropped the styles for DOM trees, e.g. search for "DOCTYPE: html" on that page and compare to the HTML spec |
| 00:20 | <benschwarz> | Hixie: the opera boys will be happy anyway, they can see their toc |
| 00:21 | <benschwarz> | Hixie: https://github.com/benschwarz/developers.whatwg.org/issues/issue/2 |
| 00:21 | <Hixie> | there's also some specific styles in view-source:http://www.whatwg.org/specs/web-apps/current-work/middle-override-styles you might want to grab |
| 00:21 | <benschwarz> | and yeah, I'll look at those trees, those were always styled badly in my version |
| 00:21 | <Hixie> | #table-example-1 for instance, which tries to make that example look more like the book it was taken from |
| 00:22 | <benschwarz> | Hixie: I'll go through the entire spec and try to make it look right |
| 00:22 | <benschwarz> | all I've really done is slash and burn, then bring over my styles |
| 00:22 | <Hixie> | cool |
| 00:23 | <benschwarz> | which had crazy overrides for the w3c and whatwg stylesheets |
| 00:23 | <Hixie> | should i drop the Conformance requirements section too? |
| 00:23 | <benschwarz> | so it was all a bit much |
| 00:23 | <Hixie> | and maybe some of the other stuff in that section |
| 00:23 | <Hixie> | woah, the UTF-8 section isn't properly marked |
| 00:24 | <benschwarz> | not sure, anything past the introduction is good content |
| 00:25 | <benschwarz> | where is that Hixie ? |
| 00:25 | <Hixie> | 2.2 and 2.4 respectively |
| 00:25 | <Hixie> | 2.4 will be gone as soon as i regen the spec |
| 00:25 | <Hixie> | i'm working on 2.2 |
| 00:25 | <Hixie> | got to make 2.2.1 turn into an h3 |
| 00:25 | <benschwarz> | okay. |
| 00:25 | <Hixie> | i have some special magic for that |
| 00:25 | <benschwarz> | I need to roll now… |
| 00:26 | <Hixie> | later |
| 00:26 | <Hixie> | and thanks! |
| 00:26 | <Hixie> | this stuff is awesome |
| 00:26 | <benschwarz> | Hixie: make it known around this channel that issues are useful! |
| 00:26 | <Hixie> | people are gonna love it |
| 00:26 | <Hixie> | will do |
| 00:26 | <benschwarz> | https://github.com/benschwarz/developers.whatwg.org/issues/ |
| 00:26 | <benschwarz> | I haven't been through it at all, but it'll be good to know what people find straight away |
| 00:27 | <benschwarz> | okay. good work! |
| 00:27 | <benschwarz> | talk laters |
| 00:34 | <benschwarz> | ps Hixie: I'll need a way to determine that the date and the secondary header are different :) |
| 00:34 | <benschwarz> | maybe a <time> tag, hey? |
| 00:34 | <paul_irish> | ;) |
| 00:34 | <Hixie> | how do you mean? |
| 00:35 | <benschwarz> | check the live copy… |
| 00:36 | <benschwarz> | I have to use "header h1 + h2" to target that first h2 |
| 00:36 | <benschwarz> | or the first after the h1, that is |
| 00:37 | <benschwarz> | so if <h2 class="no-num no-toc">7 January 2011</h2> could be a <time> or something? |
| 00:37 | <benschwarz> | its not really a header, in my mind |
| 00:37 | <Hixie> | oh, i see |
| 00:37 | <Hixie> | i can put it in a <p> after the <hgroup> if you like |
| 00:38 | <benschwarz> | sure… that will work |
| 00:38 | <benschwarz> | Are you not wanting to use <time> because of the other specs that have to be generated? |
| 00:39 | <Hixie> | i'm mostly reluctant to use <time> because i don't want to try to figure out how to get anolis to output an ISO8601 date :-) |
| 00:39 | <benschwarz> | Hixie: I guessed that, now you know the plight of other web developers :) |
| 00:40 | <Hixie> | i'm not really a fan of <time> in general |
| 00:40 | <Hixie> | but that's another story |
| 00:40 | <Hixie> | do you want the date on every page, or just the front page? |
| 00:40 | <benschwarz> | just the front |
| 00:41 | <Hixie> | regenning |
| 00:53 | <Hixie> | benschwarz: if you're still around, the file is ready |
| 00:54 | <benschwarz> | Hixie: ha! "make clean && make && ./deploy.sh" is my new friend |
| 00:56 | <benschwarz> | deployed… |
| 00:56 | <benschwarz> | it needs more love… I'll get to it |
| 00:56 | <benschwarz> | okay. I really have to go this time. |
| 00:57 | <Hixie> | later! |
| 01:19 | <cying> | is it okay to set arbitrary attributes on DOM node elements? |
| 01:19 | <cying> | (in Javascript) |
| 01:22 | <benschwarz> | Hixie: can you make the header read "Edition for Web Developers" |
| 01:53 | <Hixie> | benschwarz: done, regenning. should be up in a few minutes. |
| 03:11 | <Hixie> | Web DOM Core says INVALID_MODIFICATION_ERR is "historical", does that mean I shouldn't use it? |
| 03:12 | <Hixie> | what exception should I use for when you call event.dataTransfer.items.add('foo', 'bar') when there is already a 'foo' in the list? |
| 03:13 | <Hixie> | er, already a 'bar' in the list |
| 03:15 | <Hixie> | man, the dnd section is so much better now than it used to be |
| 03:28 | <benschwarz> | Hixie: see the toc? |
| 08:11 | jgraham | wonders if the author view stylesheet license is compatible with the spec license |
| 08:17 | <paul_irish> | jgraham: if it isnt already public domain, ben wouldnt have a problem PD'ing it |
| 08:27 | <Hixie> | benschwarz: toc? |
| 08:27 | <Hixie> | jgraham: why would they need to be compatible? |
| 08:36 | <benschwarz> | Hixie: check the live copy… |
| 08:44 | <benschwarz> | Hixie: I was hoping to get a password-less login happening on whatwg |
| 08:45 | <benschwarz> | I uploaded my keys, but not working– Anything in the config I should know about? |
| 08:45 | <zcorpan> | Hixie: so is web database "Dead Standard"? :) |
| 08:48 | <MikeSmith> | benschwarz: you ssh directory is world readable |
| 08:52 | <MikeSmith> | hmm |
| 08:53 | <MikeSmith> | and there's no key in your .ssh dir |
| 08:57 | <jgraham> | Hixie: I imagine someone might assume that the right to fork the spec also covers the stylesheet |
| 09:17 | <MikeSmith> | david_carlisle: regarding your message a while back about href, I agree |
| 09:18 | <MikeSmith> | I think for that validator we should allow it |
| 09:19 | <MikeSmith> | I have a validator patch ready for switching to your mml2-subset-of-mml2 schema, but I need for Henri to review and OK it before I commit it |
| 09:23 | <david_carlisle> | MikeSmith: OK, thanks, then I guess I should do same in the copy I stuck on the w3c server |
| 09:24 | <MikeSmith> | yeah |
| 09:24 | <MikeSmith> | that'd actually be better |
| 09:24 | <MikeSmith> | because as part of the patch, I wrote a makefile that just grabs your source |
| 09:25 | <MikeSmith> | so if/when you make changes upstream, I can just run that to build the version for the validator |
| 09:25 | <david_carlisle> | ha ha so I now have ultimate power and if I add content mathml |
| 09:25 | <david_carlisle> | and docbook |
| 09:25 | <david_carlisle> | and xforms |
| 09:25 | <MikeSmith> | heh |
| 09:25 | <david_carlisle> | and rdf/xml |
| 09:25 | <david_carlisle> | it'll all make it's way on to the web? |
| 09:26 | <david_carlisle> | sorry, just got in. need coffee |
| 09:26 | MikeSmith | needs some green tea |
| 09:28 | <MikeSmith> | david_carlisle: the validator version necessarily is a patched version of your source |
| 09:28 | <MikeSmith> | because of some cases of conflicts in relaxng patterns names used |
| 09:29 | <david_carlisle> | I could fix that, where are the patches? |
| 09:29 | <MikeSmith> | https://bitbucket.org/validator/validator-patches/qseries?apply=t&qs_apply=mathml2-update |
| 09:29 | <david_carlisle> | probably will look over weekend |
| 09:30 | <MikeSmith> | the raw patch is here: https://bitbucket.org/validator/validator-patches/raw/ba32f45e8515/mathml2-update |
| 09:30 | <MikeSmith> | there are some changes other than just the name changes |
| 09:30 | <MikeSmith> | stuff that's necessarily for making it integrate-able into the HTML5 schema |
| 09:31 | <MikeSmith> | and due to the way it handles annotation-xml |
| 09:31 | <david_carlisle> | don't get me started on the way html5 handles annotation-xml _;) |
| 09:35 | <MikeSmith> | heh |
| 09:35 | <MikeSmith> | david_carlisle: that patch is for several files but I think the main one of interest to you is just the part that adds the "schema/mml2/patch-vnu" file |
| 09:36 | <MikeSmith> | that includes comments for all the changes |
| 09:37 | <MikeSmith> | e.g., "++# changed datatype of id attr from ID to w:xml-name -- mike" |
| 09:37 | <MikeSmith> | so actually it would be great if you could review those at some point and let me know if you are OK with them |
| 09:37 | <david_carlisle> | yes thanks had a quick look through the raw patch data, but probably can't actually do anything with it till later |
| 09:39 | <MikeSmith> | ok |
| 10:01 | <hsivonen> | MikeSmith: does your patch remove the old MathML2 schema files? I see you only removing them from entity-map.txt. |
| 10:01 | <MikeSmith> | hsivonen: it does not remove them but I suppose it should |
| 10:02 | <hsivonen> | MikeSmith: do I read correctly that you implemented the same annotation-xml content model that I had implemented on top of the old MathML 2 schema? |
| 10:02 | <MikeSmith> | yes |
| 10:02 | <hsivonen> | MikeSmith: ok. r=hsivonen for landing it. |
| 10:02 | <MikeSmith> | cheers |
| 10:03 | <MikeSmith> | got another minor htmlparser refinement for the open-elements reporting |
| 10:05 | <MikeSmith> | https://bitbucket.org/validator/htmlparser-patches/qseries?apply=t&qs_apply=report-unclosed-for-eof |
| 10:05 | <MikeSmith> | and raw patch at https://bitbucket.org/validator/htmlparser-patches/raw/e5f6f9c4a410/report-unclosed-for-eof |
| 10:06 | <MikeSmith> | hsivonen: for fixing the case of open title, script, style elements |
| 10:06 | <MikeSmith> | which zcorpan pointed out |
| 10:06 | <hsivonen> | MikeSmith: r=hsivonen, although I'd prefer commit messages to start with a capital letter and end with a period to conform with hg.mozilla.org style. |
| 10:07 | <MikeSmith> | ok |
| 10:07 | <hsivonen> | MikeSmith: would you like me to push directly to hg.mozilla.org, or do you prefer to land to bitbucket first? |
| 10:08 | <MikeSmith> | please push directly to hg.mozilla.org |
| 10:08 | <hsivonen> | ok |
| 10:08 | <MikeSmith> | I don't push to the bitbucket repo for htmlparser except for just that separate patch queue for it |
| 10:09 | <annevk> | heh, someone highlighted my todo list in the logs |
| 10:09 | <MikeSmith> | the repo itself gets mirrored from whatever is at hg.mozilla.org |
| 10:09 | <zcorpan> | MikeSmith: awesome :) |
| 10:10 | <zcorpan> | MikeSmith: i didn't find any further bugs when testing briefly the other day |
| 10:10 | <MikeSmith> | ok |
| 10:10 | <MikeSmith> | but note that Henri made a further refinement yesterday also |
| 10:11 | <MikeSmith> | to prevent duplicate error messages |
| 10:11 | <MikeSmith> | zcorpan: I pushed that and this update to http://www.w3.org/html/check already |
| 10:11 | <hsivonen> | MikeSmith: pushed. http://hg.mozilla.org/projects/htmlparser/rev/0b9e1ee721df |
| 10:11 | <MikeSmith> | hsivonen: thanks |
| 10:12 | <MikeSmith> | I think people are going to really like having this |
| 10:13 | <hsivonen> | yeah, I expect so |
| 10:13 | <MikeSmith> | and maybe having the locators in the stack will prove useful for other stuff |
| 10:13 | <zcorpan> | yeah, this immediately doubled the usability of the validator |
| 10:16 | <MikeSmith_> | zcorpan: hey, I was thinking last night about another open bug that you reported some time ago |
| 10:16 | <MikeSmith_> | the one about all the attribute names being listed for input elements |
| 10:17 | <MikeSmith_> | I think a better way to handle that might be in the javascript for the Web UI |
| 10:18 | <MikeSmith_> | this bug, I mean http://bugzilla.validator.nu/show_bug.cgi?id=339 |
| 10:18 | <MikeSmith_> | we wouldn't need to scrape the spec |
| 10:20 | <MikeSmith> | on the backend |
| 10:21 | <MikeSmith> | would instead need to have the javascript scrape the error message and the remove whichever <li>s need to me removed |
| 10:24 | <MikeSmith> | would not be so pretty but it's going to be kind of ugly regardless of where it's handled (backend/build or the UI) |
| 10:26 | <zcorpan> | MikeSmith: it wouldn't work so well with the group messages feature |
| 10:26 | <MikeSmith> | ah |
| 10:27 | <zcorpan> | MikeSmith: another thing i noticed just now. <input max="1"> has less useful message than <input type="text" max="1"> even though they're semantically equivalent |
| 10:28 | <zcorpan> | MikeSmith: maybe that's not a common problem though |
| 10:29 | <MikeSmith> | what would be better for the <input max="1"> case? |
| 10:31 | <zcorpan> | to say that max is not allowed for the "Text" state |
| 10:31 | <MikeSmith> | ah |
| 10:32 | <MikeSmith> | the thing is, the schema doesn't default it to the text state |
| 10:32 | <MikeSmith> | no attribute defaulting in relaxng |
| 10:34 | <MikeSmith> | so the relaxng backend at least does not know that those two cases are but type="text" |
| 10:34 | <annevk> | schemas really have a lot of limitations |
| 10:50 | <david_carlisle> | annevk: they do, but that one is a limitation of this schema rather than of relaxng, i would guess? |
| 10:51 | <annevk> | dunno |
| 10:51 | <annevk> | it seems that for useful error messages not having schemas would make things easier |
| 10:51 | <hsivonen> | david_carlisle: many of the limitations of "schemas" are limitations of Jing |
| 10:52 | <annevk> | but I'm not sure how much more complex it would be |
| 10:52 | <david_carlisle> | everything is james' fault, i agree |
| 10:52 | <annevk> | well that is, you need custom code anyway, so... |
| 10:53 | <david_carlisle> | annevk: perhaps, but it's harder to drive a context sensitive editor from a load of custom java code |
| 10:54 | <hsivonen> | david_carlisle: the way this has been addressed so far is by punching more and more holes into the RELAX NG schemas (making them less useful for oXygen, nxml-mode, etc.) and filling the holes with custom Java code |
| 10:54 | <david_carlisle> | :-) |
| 10:54 | <MikeSmith> | generic validation tools produce generic error messages |
| 10:55 | <david_carlisle> | but if the schemas are useful for editors, people don't make errors, then the validator never need report anything, or something.... |
| 10:55 | <david_carlisle> | but i guess you will tell me Hixie managed to find an example of an invalid html file in the wild |
| 10:56 | <MikeSmith> | (or rather "general-purpose validation" tools) |
| 10:57 | <david_carlisle> | MikeSmith: I agree really, and think validator.nu or schematron really produce much better error messages for real people |
| 10:58 | <david_carlisle> | schema validation (like dtd) is best viewed as binary pass fail, the error message returned (if any) is often as not misleading as it reports things as seen by the validators internal state rather than what was in the file |
| 10:58 | <MikeSmith> | indeed |
| 10:58 | <david_carlisle> | but I still use dtd every day anyway |
| 10:59 | <MikeSmith> | what for? |
| 10:59 | <david_carlisle> | cause I need food |
| 10:59 | <MikeSmith> | heh |
| 11:00 | <MikeSmith> | "f your desires are not great, a little will seem much to you; for small appetite makes poverty equivalent to wealth." |
| 11:02 | <MikeSmith> | david_carlisle: btw, your existing http://www.w3.org/Math/RelaxNG/mathml3-restricted/ schema seems to already allow href |
| 11:03 | <MikeSmith> | in CommonAtt |
| 11:03 | <annevk> | david_carlisle, people make errors and typically do not use schemas :) |
| 11:04 | <david_carlisle> | annevk: you don't say:-) |
| 11:05 | <david_carlisle> | MikeSmith: does it? oops will check, i was experimenting with or without at one point but I thought I'd checked it in as we'd agreed without href, but will check while I'm looking at pulling in your other diffs |
| 11:06 | <MikeSmith> | ok |
| 11:54 | <annevk> | you can now go back from http://html5.org/tools/web-apps-tracker?from=5743&to=5744 to the overview page by clicking the heading |
| 11:54 | <annevk> | I also removed "Revision" from the title |
| 11:54 | <annevk> | and 410'd web-forms-tracker |
| 11:59 | <MikeSmith> | nice |
| 12:00 | <MikeSmith> | annevk: it would be nice to just be able to have also have a URL with single revision number |
| 12:00 | <MikeSmith> | http://html5.org/tools/web-apps-tracker?r=5744 or whatever |
| 12:01 | <annevk> | I'm not following what Sam is going on about again :/ |
| 12:01 | <MikeSmith> | which would just give the same thing as http://html5.org/tools/web-apps-tracker?from=5743&to=5744 |
| 12:02 | <annevk> | what is the upside? |
| 12:02 | <annevk> | just shorter URLs? |
| 12:04 | <annevk> | hmm, found a mirror: http://kerouanen.appspot.com/html5.org/tools/web-apps-tracker |
| 12:05 | <annevk> | but the mirror is broken |
| 12:15 | <MikeSmith> | annevk: shorter |
| 12:15 | <MikeSmith> | but actually could be even more shorter |
| 12:15 | <MikeSmith> | http://html5.org/r/5744 for example |
| 12:16 | <annevk> | aah, as a URL shortener? |
| 12:16 | <annevk> | where would you use them? |
| 12:16 | <annevk> | I will try to make something like that |
| 12:16 | <MikeSmith> | in tweets or whatever I guess |
| 12:16 | <MikeSmith> | cool |
| 12:17 | <MikeSmith> | david_carlisle: validator.nu repo now has the new MathML schema |
| 12:18 | <MikeSmith> | and I pushed it to http://www.w3.org/html/check for testing |
| 12:18 | <MikeSmith> | so if/when you have time, please try some test cases on it |
| 12:20 | <david_carlisle> | MikeSmith: thanks |
| 12:21 | <MikeSmith> | I have a couple more patches I need for Henri to review, but after I get those reviewed and checked in, hopefully Henri can re-redeploy http://validator.nu and after that I will also update the HTML5 backend on validator.w3.org so that we can have those in sync |
| 12:21 | <MikeSmith> | within the next week or so |
| 12:21 | <hsivonen> | MikeSmith: which other patches? |
| 12:22 | <MikeSmith> | hsivonen: one is https://bitbucket.org/validator/syntax-patches/qseries?apply=t&qs_apply=script-checker |
| 12:23 | <MikeSmith> | for checking text content of <script></script> and <script src=foo></script> |
| 12:25 | <MikeSmith> | but note that it does not implement checking for all the constraints in the ABNF at http://www.whatwg.org/specs/web-apps/current-work/complete/scripting-1.html#restrictions-for-contents-of-script-elements |
| 12:25 | <MikeSmith> | it only intentionally only checks at this point for "<!--" without a matching "-->" |
| 12:26 | <MikeSmith> | and for proper "//" "/* */" comments in the <script src=foo></script> case |
| 12:55 | <annevk> | MikeSmith, it now works |
| 12:55 | <annevk> | MikeSmith, you can also do http://html5.org/r/4000-4010 |
| 12:56 | <webr3> | annevk, that's handy |
| 12:57 | <annevk> | now all that needs changing is that instead of showing "Filter" on the DIFF page it shows a short URL there instead |
| 13:39 | <annevk> | what is it with the twitter tooltips for URLs and adding inappropriate slashes? |
| 13:40 | <zcorpan> | http://whoisryannystrom.com/tests/webcontrol/ is cool |
| 13:40 | <annevk> | btw the pages now link back to the short URL |
| 13:40 | <annevk> | http://html5.org/r/1143 |
| 13:40 | <annevk> | "No websocket or canvas support, SORRY." |
| 13:42 | <matjas> | zomg, the `/` after `r` is redundant!!11eleven |
| 13:43 | <zcorpan> | matjas: the 'r' is redundant too? |
| 13:43 | <matjas> | yeah in a way |
| 13:44 | <matjas> | now if only @whatwg would start tweeting again, those sexy short URLs could be put to use :) |
| 13:44 | <annevk> | too late? |
| 13:44 | <matjas> | it’s never too late for some hot 301 lovin’ |
| 13:45 | <annevk> | mwah, two characters |
| 13:45 | <matjas> | doesn’t matter much anyway |
| 13:45 | <annevk> | someone first get me html.org :) |
| 13:45 | <annevk> | or ht.ml |
| 13:45 | <annevk> | or .html |
| 13:45 | <Lachy> | what ccTLD is .ml? |
| 13:45 | <annevk> | then ill make http://html/1143 work |
| 13:45 | <zcorpan> | annevk: just register the html:// protocol and get it implemented |
| 13:46 | <annevk> | Lachy, Mali |
| 13:46 | <Lachy> | hmm, I wonder what the requirements for registering one are? |
| 13:46 | <annevk> | Lachy, but you cannot get directly under it |
| 13:46 | <zcorpan> | then use html:1143 |
| 13:46 | <Lachy> | oh |
| 13:46 | <annevk> | zcorpan, profit |
| 13:49 | <annevk> | matjas, yeah we should totally fix the twitter feed |
| 13:49 | <zcorpan> | annevk: you could represent the revision number in base 64 to make it shorter |
| 13:49 | <matjas> | and then rot13 it for the lulz |
| 13:49 | <annevk> | no and no |
| 13:50 | <matjas> | annevk: did the python script i found last time not work? |
| 13:50 | <annevk> | I don't think we hooked it up |
| 13:50 | <annevk> | Admittedly I also forgot the password to the twitter account |
| 13:50 | <matjas> | FFFFFFFUUUUUU— |
| 13:51 | <annevk> | Not sure if I have time tonight but maybe Hixie can give me the details and I can set up a script on html5.org that he pings with a message and revision number |
| 13:52 | <annevk> | @WHATWG has a lot of followers but currently it is not very useful |
| 13:55 | <annevk> | matjas, was it http://code.google.com/p/python-twitter/ ? |
| 13:55 | <matjas> | nope, but this looks pretty good |
| 13:57 | <annevk> | not exactly simple |
| 13:57 | <annevk> | http://code.google.com/p/python-twitter/source/browse/examples/tweet.py |
| 13:59 | <annevk> | oh |
| 13:59 | <annevk> | that's just a lot of boilerplate |
| 13:59 | <annevk> | it comes down to |
| 13:59 | <annevk> | status = api.PostUpdate(message) |
| 13:59 | <annevk> | and creating api |
| 14:00 | <annevk> | miketaylr, did you co-create python-twitter? |
| 14:00 | <matjas> | requires authentication though |
| 14:01 | <matjas> | looks like it’s another Mike Taylor, codenamed bear42 |
| 14:03 | <annevk> | don't they all require authentication of some kind? |
| 14:03 | <annevk> | seems quite clear you need to use OAuth |
| 14:04 | <annevk> | somewhat overkill for stuff like this I think, but whatever |
| 14:05 | <miketaylr> | annevk: haha no, it was the smart mike taylor |
| 14:09 | <annevk> | wait what |
| 14:09 | <annevk> | and we hired you? |
| 14:09 | <miketaylr> | shhh |
| 14:09 | <annevk> | ;p |
| 14:09 | <miketaylr> | >_> |
| 14:10 | <matjas> | miketaylr: /nick bear43 |
| 14:11 | <annevk> | I'll email Hixie asking for the details |
| 14:12 | <matjas> | annevk: yay! |
| 14:12 | <matjas> | annevk: looks like there’s a missing </code> after “Short URL:” on pages like http://html5.org/tools/web-apps-tracker?from=5746&to=5747 |
| 14:14 | <matjas> | are you relying on proper parsing/error handling as per HTML5? |
| 14:16 | <annevk> | I used + instead of += |
| 14:16 | <annevk> | fixed now |
| 14:20 | <zcorpan> | annevk: localStorage["tracker-" + ""] should be localStorage["tracker-" + name] ? |
| 14:25 | <annevk> | something like that, yes :/ |
| 14:25 | <annevk> | if it's been broken for so long |
| 14:25 | <annevk> | maybe I should just nuke that? |
| 14:26 | <annevk> | for me it has been broken since revision 5280 |
| 15:00 | <annevk> | Ms2ger, thanks for doing a few things with the encoding data |
| 15:01 | <annevk> | Ms2ger, added a comment to the bug hoping to sway Simon |
| 15:01 | <Ms2ger> | Thanks, I was hoping you would :) |
| 15:05 | zcorpan | switches to developers.whatwg.org for http://simon.html5.org/html5-elements |
| 15:06 | <zcorpan> | benschwarz: if you have a style sheet to make html5-elements match the style of developers.whatwg.org i'd be happy to apply it |
| 15:07 | <zcorpan> | benschwarz: btw the color of the headings is not so great contrast |
| 15:08 | <zcorpan> | or the <dt>s rather |
| 15:13 | <erlehmann> | The WHATWG is committed to the well being of all participants. Cake and grief counseling will be available at the conclusion of the specification. Thank you for helping us help you help us all. |
| 15:16 | <erlehmann> | zcorpan, that is a nice site. but the light blue color really should be changed to enhance the contrast level. |
| 16:49 | <AryehGregor> | Hixie, what are the beforeundo and beforeredo events, exactly? Googling them seems to yield no info. |
| 16:57 | <AryehGregor> | (if anyone else knows, that would be helpful too) |
| 16:58 | <annevk> | maybe you should search for them with the prefix "on" |
| 16:58 | <annevk> | I believe the idea is that using them you can make sure the undo/redo UI is highlighted appropriately, or some such |
| 16:59 | <annevk> | sort of similar to beforecopy et al |
| 16:59 | <AryehGregor> | I did search for them with the prefix "on". |
| 16:59 | <AryehGregor> | I got some Dojo stuff. |
| 16:59 | <Workshiva> | Have you checked MSDN? |
| 17:02 | <AryehGregor> | I don't seem to see anything there. |
| 18:48 | <annevk> | hmm, WHATWG blog is pretty much dead |
| 18:49 | <Ms2ger> | Write something about encodings there :) |
| 18:51 | <annevk> | yeah, could do |
| 18:51 | <annevk> | guess I'll write a post once I fix the twitter account |
| 18:55 | <dglazkov> | Hixie: yt? |
| 18:56 | <dglazkov> | interesting question about querySelector behavior and XBL2 pseudo-element matching |
| 18:57 | <dglazkov> | currently, we're supposed to throw a SYNTAX_ERR if the pseudo-element is unknown. Well, that's no longer reasonable once we introduce the ability to define arbitrary strings as pseudo-elements. |
| 18:57 | <dglazkov> | should we: |
| 18:57 | <dglazkov> | 1) change querySelector behavior to not throw on unknown pseudo-elements |
| 18:58 | <Ms2ger> | Pseudo-elements or pseudo-classes? |
| 18:58 | <dglazkov> | 2) throw SYNTAX_ERR when the pseudo-element in the query was not matched for the given query |
| 18:58 | <dglazkov> | oooh |
| 18:59 | <dglazkov> | 3) since they are in the shadow DOM anyway, we should keep always throwing SYNTAX_ERR when unknown pseudo-element is encountered |
| 19:00 | <dglazkov> | Ms2ger: pseudo-elements: http://dev.w3.org/2006/xbl2/#the-pseudo-attribute |
| 19:00 | <dglazkov> | I think I'll implement #3 for now |
| 19:17 | <annevk> | dglazkov, pseudo-elements are anonymous boxes, I don't think they should get exposed to script |
| 19:17 | <annevk> | (even when they are in fact elements in an XBL binding) |
| 19:28 | <dglazkov> | annevk: so #3 seems like the right choice? |
| 19:28 | <annevk> | yeah |
| 19:28 | <annevk> | I think I had XBL in mind already when designing that part of querySelector |
| 19:33 | <Hixie> | ok here now |
| 19:36 | <annevk> | oh hey |
| 19:36 | <annevk> | see some email about twitter in your inbox |
| 19:37 | <Hixie> | k |
| 19:37 | <Hixie> | got it |
| 19:38 | <cszabo1> | Hixie, did you follow my discussion with Robert O'Callahan about the drawing model on canvas? |
| 19:38 | <Hixie> | where was the discussion? |
| 19:38 | <cszabo1> | on whatwg-mailing list |
| 19:38 | <Hixie> | i guarantee a reply to all spec feedback on the whatwg list |
| 19:39 | <Hixie> | (unless it just repeats something that was said earlier) |
| 19:39 | <Hixie> | so yes, and i'll reply in due course... i can take a look now though if you need a reply sooner |
| 19:40 | <cszabo1> | Hixie, OK, I will just be patient. I was raising among other things, an issue with shadows not being drawn for compositing modes such as copy in the current model. |
| 19:40 | <Hixie> | oh, right, the compositing thread |
| 19:41 | <Hixie> | i'm basically happy to spec whatever is (a) implemented and (b) described to me in terms i can put in the spec |
| 19:41 | <Hixie> | i unfortunately am not qualified to really describe what browsers do from reverse engineering -- my graphics knowledge is lacking |
| 19:41 | <Hixie> | otherwise i'd just reverse engineer the browsers and spec that |
| 19:45 | <Hixie> | annevk: replied. Does anyone actually care though? Given how few complaints we got when it broke (the earliest being months later), it's not clear to me that Twitter messages are helpful here. |
| 19:48 | <Ms2ger> | People might not notice if it's gone, but still appreciate it when they do see it, though |
| 19:49 | <Hixie> | fair enough |
| 20:02 | <cszabo1> | Hixie, do you know whether Microsoft agrees to the current behavior or wants something more along what WebKit does? I heard both versions so far. If Mozilla, Opera and IE behave like the spec says, and they are happy with that, probably WebKit should do the same. |
| 20:06 | <Philip`> | cszabo1: http://lists.w3.org/Archives/Public/public-canvas-api/2010OctDec/0070.html is what they've said |
| 20:11 | <jamesr_> | afaik they haven't actually released an implementation of any compositing modes other than source-over so far |
| 20:12 | <Philip`> | Yeah, their preview releases have skipped it so far |
| 20:13 | <jamesr_> | and fwiw i agree that it would make sense for WebKit to change to the currently spec'd behavior if IE does, even though i think it's kind of dumb |
| 20:13 | <Philip`> | If they implement it to pass the current test cases then they should end up with behaviour like Opera (and like Firefox with fewer bugs, I think) |
| 20:13 | <Philip`> | and they haven't asked for those test cases to be changed at all |
| 20:14 | <cszabo1> | I have just tested the "copy" globalCompositeOperation as it relates to shadows being shown on Firefox and it appears to work not as speced, but according to my proposal, that is: shadows are shown if specified. |
| 20:16 | <Philip`> | Hmm, I'm probably missing a test for that one |
| 20:19 | <jamesr_> | last i checked ffx doesn't perfectly implement the non-default globalCompositeOperations in all corner cases |
| 20:23 | <Ms2ger> | Please file bugs, if Philip` hasn't yet :) |
| 20:24 | Philip` | hasn't bothered filing any bugs any time recently |
| 20:25 | <annevk> | so at one point I installed my own version of Python because Dreamhost was lagging behind |
| 20:25 | <annevk> | is there a way to remove that again? |
| 20:25 | <Philip`> | Just look for failures in http://test.w3.org/html/tests/submission/PhilipTaylor/canvas/index.2d.composite.html if you want :-) |
| 20:26 | <Philip`> | annevk: Delete it from your home directory, if that's where it got installed? |
| 20:26 | <Philip`> | maybe in ~/lib or similar |
| 20:26 | <annevk> | It's in ~/opt/lib/ |
| 20:27 | <annevk> | however I'm afraid to break things |
| 20:27 | <annevk> | e.g. anolis is there too |
| 20:27 | <Philip`> | Ah |
| 20:27 | <AryehGregor> | Some of the spam bugs are just fascinating: http://www.w3.org/Bugs/Public/show_bug.cgi?id=11695 |
| 20:27 | <annevk> | the main problem I'm having right now is that for some reason md5 support is breaking |
| 20:28 | <annevk> | these are the last four lines: |
| 20:28 | <annevk> | md5 = __get_builtin_constructor('md5') |
| 20:28 | <annevk> | File "/home/annevk/opt//lib/python2.5/hashlib.py", line 60, in __get_builtin_constructor |
| 20:28 | <annevk> | import _md5 |
| 20:28 | <annevk> | ImportError: No module named _md5 |
| 20:28 | <annevk> | I think this is the main thing I hate about Python |
| 20:28 | <Philip`> | If you want an easily reversible way, you could remove PATH/PYTHONPATH etc from wherever they're set (maybe ~/.bashrc) so it always runs the default Python instead |
| 20:29 | <annevk> | installing frigging dependencies all the time and nothing ever working straight |
| 20:29 | <AryehGregor> | The mojibake is asking things like "How many spearmen do you have?" and so on, followed by reports "Your soldiers are worth X generals", "Your navy is worth Y generals", "Your military is worth X + Y generals", then "Credit to eliranisrael, author of the code." In Hebrew. |
| 20:30 | <AryehGregor> | a) What is that program supposed to do? b) How did it end up being submitted as spam. |
| 20:30 | <AryehGregor> | ? |
| 20:30 | <annevk> | Philip`, nothing in .bashrc or .bash_profile |
| 20:32 | <Philip`> | AryehGregor: http://board.ikariam.co.il/print.php?threadid=80228&page=1 |
| 20:32 | <Philip`> | Sounds like it's for the game Ikariam |
| 20:32 | <AryehGregor> | Well, Googling is one strategy, yeah. |
| 20:32 | <AryehGregor> | Sometimes I forget that. |
| 20:33 | <cszabo1> | Philip`, Hixie, Here is a simple test proposal: http://waplabdc.nokia-boston.com/browser/users/caszabo/shadow/index.htm and the spec change to go with it is in this thread: http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2011-January/029658.html |
| 20:33 | <Ms2ger> | AryehGregor, you didn't get the memo that Google is not a verb? ;) |
| 20:35 | <AryehGregor> | Ms2ger, I'm not an employee, I'm just providing them with some service as an external vendor. So I don't get their memos. |
| 20:35 | <Philip`> | The bug was posted by someone who uses an ISP that looks Hebrewish |
| 20:35 | AryehGregor | is pretty sure he's observed Google employees using "Google" as a verb, anyway |
| 20:35 | <Philip`> | Maybe it's not spam, maybe they happened to have that program on the clipboard then accidentally ctrl+v+entered it into the spec |
| 20:36 | <AryehGregor> | Well, so then it's accidental spam. |
| 20:36 | AryehGregor | finds it funny to read conversations like this in Hebrew, which he normally only sees in liturgy and rabbinic texts |
| 20:36 | <AryehGregor> | One guy wrote it in Java, another one wants it in C#. Someone else thinks it should go on a webpage, and of course someone has to point out that Java and JS are actually two totally different languages. |
| 20:37 | <AryehGregor> | Oh, it was the guy who wrote it who thought Java was JavaScript, so he wanted it for a webpage. |
| 20:39 | <Ms2ger> | And now he thought we'd explain him how to rewrite it with canvas? |
| 20:39 | <AryehGregor> | Ah, that's probably it. Someone was probably trying to figure out how to do it in HTML and got directed to the HTML5 spec by mistake. |
| 20:40 | <Ms2ger> | Still, INVALID |
| 20:40 | <AryehGregor> | Yeah, I'll back you up on that one. |
| 20:40 | <Philip`> | Surely we should rewrite it into JS for him and then reply on the forum |
| 20:41 | <AryehGregor> | That would be nice. |
| 20:41 | <AryehGregor> | It might make them submit more bugs in that vein, though, which wouldn't be nice to the other HTMLWG members. |
| 20:41 | <Ms2ger> | Then the next one will make his bug into an ISSUE |
| 20:41 | <AryehGregor> | That would be pretty funny, actually. |
| 20:42 | <Ms2ger> | You mean, like, Julian-funny? |
| 20:42 | <AryehGregor> | I imagine the chairs would agree to summarily dismiss such an issue. Which reference is used for ASCII they might feel compelled to treat as a serious issue for the form of things, but a code rewriting request service might be a little much for them. |
| 20:42 | <AryehGregor> | But who knows? |
| 20:43 | <annevk> | Philip`, so it seems I have no need for Anolis |
| 20:43 | <annevk> | Philip`, can I then simply nuke the contents of that folder? |
| 20:43 | <Ms2ger> | You did see his codepoint issue, right, AryehGregor? |
| 20:43 | <Philip`> | annevk: I'd suggest renaming it out of the way, and then making sure nothing catastrophic happens |
| 20:44 | <Philip`> | but otherwise that probably ought to be sufficient |
| 20:44 | <AryehGregor> | Julian's? |
| 20:44 | <AryehGregor> | The one about presentation of the code points? |
| 20:44 | <Philip`> | since bash/Python will end up looking in the default system paths for the files, I think |
| 20:44 | <AryehGregor> | I actually agree with him on that. |
| 20:44 | <AryehGregor> | Although I wouldn't raise it to an issue or file CPs or anything, since it's not worth the energy. |
| 20:46 | <annevk> | Philip`, I renamed them and it seems things still work, except that I cannot use python or python2.5 on the commandline |
| 20:46 | <annevk> | Philip`, it searches for them in the wrong place (i.e. my local dir) |
| 20:46 | <annevk> | and I cannot find where the settings are :/ |
| 20:47 | <Philip`> | Hmm |
| 20:47 | <Philip`> | It should search the whole PATH for them, I thought |
| 20:47 | <annevk> | I get -bash: /home/annevk/opt/bin/python: No such file or directory |
| 20:48 | <annevk> | whereas on the site things still work as they look at /usr/bin/python |
| 20:48 | <Philip`> | What does "echo $PATH" say? |
| 20:49 | <AryehGregor> | annevk, try doing PATH=$PATH |
| 20:49 | <AryehGregor> | It caches the results of $PATH lookups. |
| 20:49 | <AryehGregor> | Or just start a new shell. |
| 20:49 | <AryehGregor> | If you change PATH (even to the same thing), it gets rid of the cache. |
| 20:50 | <annevk> | ah, starting a new shell helps |
| 20:50 | <annevk> | great great |
| 20:52 | <annevk> | should I use http://wiki.dreamhost.com/Python#Virtualenv ? |
| 20:56 | <cszabo1> | Philip`, Did you see my test proposal? |
| 20:56 | <cszabo1> | Does it look ok to you? |
| 21:04 | <Philip`> | cszabo1: Tests should be designed so that if they pass then no red is visible (and preferably if they fail then some red is visible) (and preferably if they pass then they're entirely green) |
| 21:06 | <Philip`> | Maybe that's not incredibly useful when we have JS to verify the correctness of the output, rather than verifying it by eye, but it's the convention so it's probably good to stick with it |
| 21:08 | <Philip`> | Probably we ought to have tests to make sure it does the right thing with semi-transparent globalAlpha, and with other compositing modes |
| 21:08 | <Philip`> | Do you know if Firefox does what you proposed in those cases? |
| 21:10 | <cszabo1> | Philip`: Actually not, I have to fix my test. there should be no green. |
| 21:11 | <cszabo1> | Philip`: I was away from my desk to run the test on Opera (I do not have opera on my computer). Opera sticks to the curent wording of the spec and erases the shadow. |
| 21:16 | <Hixie> | cszabo1: if i understand right, the proposal is to change the shadow steps 3 and 4 with your step 3? the last two steps don't change? |
| 21:17 | <Hixie> | seems reasonable, if so |
| 21:17 | <cszabo1> | Philip`: I fixed my test. Now everybody fails it, because all browsers act different, but I believe that the test now requires a logical output: If I draw a rectangle with a shadow composited with copy, all I am getting is a rectangle with a shadow. Opera eats the shadow (per the spec) Mozilla now leave part of the older image. |
| 21:17 | <Hixie> | assuming Philip` agrees |
| 21:18 | <cszabo1> | Hixie, yes |
| 21:19 | Philip` | hasn't thought enough about the proposal to be sure it doesn't have problems in various cases |
| 21:19 | <cszabo1> | The idea is that if a composite operator is used, that would clear the image where nothing is drawn, the last step clears the shadow of the object, which is not good. |
| 21:20 | <cszabo1> | Also, if a composite operator is used that does not draw where there is prior content (i.e. out and xor), the shadow prevents the object from being drawn. |
| 21:22 | <cszabo1> | Hixie, Philip`, My proposal is to draw the object and it's shadow as one piece such that these efects are aplied between the prior content of the canvas and the new object being drawn. |
| 21:32 | <hsivonen> | Python standard lib relocating md5 is programmer hostile. leads to pointless warnings or effort to update references to a non-deprecated lib |
| 21:38 | <annevk> | http://twitter.com/WHATWG/status/23494775223091201 |
| 21:39 | <annevk> | got the fricking OAuth to work |
| 21:39 | <matjas> | oh god please let it be |
| 21:39 | <matjas> | annevk++ |
| 21:40 | <annevk> | I wonder what happens for messages larger than 140 characters |
| 21:40 | <annevk> | especially when there's a URL in them |
| 21:43 | <annevk> | that simply fails |
| 21:46 | <annevk> | anyone object against this: |
| 21:47 | <Hixie> | annevk: btw, the message part can include newlines |
| 21:47 | <annevk> | str = shorturl + " — " + message |
| 21:47 | <annevk> | str[:140] |
| 21:47 | <Hixie> | also, what i used to do is filter out any commits that has the 'e' flag set |
| 21:47 | <Hixie> | and the bug and flags likely don't need to be in there |
| 21:47 | <Hixie> | that looks good |
| 21:48 | <annevk> | can you filter e on your end? |
| 21:48 | <annevk> | i.e. not ping for e |
| 21:48 | <annevk> | otherwise I can build that in |
| 21:48 | <Hixie> | i can, but it's probably easiest for you to do it on your side, that way you can add features that use that info later if you do anything unrelated to twitter |
| 21:49 | <Hixie> | e.g. you can hook this up to web-apps-tracker to get it to update when svn changes rather that on a cronjob |
| 21:49 | <Hixie> | (just an idea, not saying you should do that) |
| 21:49 | <Hixie> | (i'm a firm believer in not fixing what ain't broken ;-) ) |
| 21:49 | <annevk> | web-apps-tracker is currently live |
| 21:49 | <annevk> | there's no cronjob |
| 21:49 | <Hixie> | ah well there you go |
| 21:50 | <Hixie> | it could be static :-) |
| 21:50 | <Hixie> | if necessary |
| 21:50 | <Hixie> | in the future |
| 21:50 | <annevk> | yeah, might be neat at some point |
| 21:50 | <annevk> | so can I get it like ?rev=3000&message=message&flags=flags ? |
| 21:50 | <Hixie> | sure |
| 21:50 | <Hixie> | and &level=level |
| 21:51 | <Hixie> | and &bug=bug |
| 21:51 | <Hixie> | if you like |
| 21:51 | <Hixie> | that's all the info i have iirc |
| 21:51 | <Hixie> | note that rev is sometimes missing |
| 21:51 | <Hixie> | as is bug |
| 21:51 | <annevk> | why would rev be missing? |
| 21:51 | <Hixie> | sometimes all i do is update the w3c copy |
| 21:51 | <Hixie> | and it doesn't affect the whatwg svn |
| 21:51 | <Hixie> | so there's no svn rev |
| 21:52 | <annevk> | okay, in those cases no tweet will go out |
| 21:52 | <Hixie> | seems reasonable |
| 21:52 | <Hixie> | people can follow mike's twitter for those if they want |
| 21:52 | <annevk> | i'll write some cgi code and then get the URL to you |
| 21:53 | <Hixie> | wtf i just lost my ipv4 address |
| 21:53 | <Hixie> | k |
| 21:53 | <Hixie> | GET or POST? |
| 21:53 | <annevk> | POST would prolly be better |
| 21:53 | <annevk> | not sure it makes a difference to Python |
| 21:55 | <Hixie> | k |
| 22:09 | <annevk> | I tried via GET and that certainly works |
| 22:10 | <annevk> | I also gave you the details |
| 22:12 | <Hixie> | thanks |
| 22:12 | <annevk> | Programming today is screaming at APIs and getting result a few hours later |
| 22:12 | <Hixie> | will add shortly |
| 22:20 | <annevk> | suitable location/bio for @WHATWG? |
| 22:20 | <Hixie> | http://www.whatwg.org/ ? |
| 22:20 | <Hixie> | or the faq |
| 22:20 | <Ms2ger> | Everywhere/There is no cabal |
| 22:20 | <Hixie> | or the blog |
| 22:20 | <Hixie> | oh you mean a string? |
| 22:20 | <annevk> | link is a separate one |
| 22:20 | <Hixie> | "Please leave your sense of logic at the door, thanks!" |
| 22:21 | <annevk> | that and Everywhere! great |
| 22:21 | <Hixie> | or The Web :-) |
| 22:21 | <Hixie> | (as a location) |
| 22:21 | <annevk> | :) |
| 22:22 | <annevk> | you win |
| 22:22 | <Hixie> | :-) |
| 22:25 | <Hixie> | ok here goes, i'm doing my first commit since hooking up the spec to your script |
| 22:26 | <Hixie> | it didn't tweet! |
| 22:26 | <annevk> | I guess you don't have the request? |
| 22:27 | <Hixie> | how do you mean? |
| 22:27 | <Hixie> | i can reconstruct it |
| 22:27 | <Hixie> | is it not in your logs? |
| 22:27 | <matjas> | annevk: you are my hero. |
| 22:28 | <annevk> | it's not working matjas :) |
| 22:28 | <Hixie> | oh i see the problem |
| 22:28 | <Hixie> | i misused the API |
| 22:28 | <matjas> | oh, i saw https://twitter.com/whatwg/status/23502541589319680 and started cheering |
| 22:29 | <Hixie> | hey can i just send a POST instead? |
| 22:29 | <Hixie> | that would make my life slightly easier |
| 22:29 | <annevk> | yes either POST or GET should work |
| 22:29 | <Hixie> | k |
| 22:29 | <annevk> | as long as parameters named rev/message/flags are there |
| 22:29 | <Hixie> | yeah |
| 22:29 | <Hixie> | ok hold on let me find something else to fix |
| 22:29 | <annevk> | if either rev is empty or flags contains an "e" nothing happens |
| 22:30 | <annevk> | and I'm hoping Python and the API do not disagree on how to count to 140 but I suspect there may be problems if lots of non-ASCII characters start getting involved |
| 22:31 | <annevk> | takes a long time for twitter to propagate profile updates to http://twitter.com/WHATWG |
| 22:31 | <annevk> | as in, it has not happened yet |
| 22:38 | <Hixie> | annevk: has for me |
| 22:38 | <annevk> | ah, yeah, good |
| 22:38 | <annevk> | now all we need is updates |
| 22:38 | <Hixie> | working on that |
| 22:38 | <Hixie> | 5 secs |
| 22:39 | <Hixie> | regenning as we speak |
| 22:40 | <Hixie> | it works! |
| 22:40 | matjas | sits back with a big bucket of pop corn |
| 22:40 | <matjas> | EUREKA |
| 22:40 | <annevk> | whoa |
| 22:40 | <matjas> | http://twitter.com/WHATWG/status/23510192872230912 |
| 22:40 | <matjas> | eff yes |
| 22:40 | <annevk> | including the larger than 140 characters |
| 22:40 | <matjas> | via WebAppsTweet :') |
| 22:40 | <annevk> | and my … character i replaced the last one with |
| 22:40 | <Hixie> | well, i guess we have one happy customer :-D |
| 22:41 | matjas | cries tears of joy |
| 22:42 | <Hixie> | one _very_ happy customer |
| 22:42 | <Hixie> | thanks annevk :-) |
| 22:43 | <annevk> | added http://twitter.com/WHATWG/status/23510988938543105 :) |
| 22:43 | <Hixie> | lol you can't thank yourself :-P |
| 22:44 | <annevk> | it's not me, it's the WHATWG |
| 22:44 | <Hixie> | :-P |
| 22:45 | <matjas> | perfect use case for the shiny new short URLs |
| 22:46 | <Hixie> | is there any practical difference between font-size: inherit and font-size: 1em? |
| 22:46 | <Hixie> | it inherits the computed value right? |
| 22:46 | <Hixie> | so it's always a length? |
| 22:46 | <Hixie> | it won't inherit 'smaller' or anything silly like that? |
| 22:47 | <annevk> | right |
| 22:47 | <annevk> | http://www.w3.org/TR/CSS2/fonts.html#font-size-props |
| 22:47 | <Hixie> | i wonder why http://www.w3.org/Bugs/Public/show_bug.cgi?id=11020 then |
| 22:47 | <annevk> | though maybe quirks with tables/form controls? |
| 22:48 | <annevk> | I think browsers do something weird with monospace fonts |
| 22:48 | <annevk> | in Opera they look generally quite big whereas in other browsers they are quite tiny |
| 22:48 | <annevk> | comes up every now and then |
| 22:48 | <Hixie> | yeah i always change the defaults in my browsers to 16px for both |
| 22:48 | <Hixie> | drives me crazy that browsers still ship with the default set to 13px |
| 22:48 | <Hixie> | looks so ugly and wrong |
| 22:49 | <Ms2ger> | The solution for that is font-family: monospace, sans-serif |
| 22:49 | <annevk> | but 1em still means 13px?! |
| 22:50 | <Hixie> | yeah i dunno |
| 22:50 | <annevk> | there is a difference once you set it to 16px |
| 22:50 | <Hixie> | ~/me pimps http://hixie.ch/specs/css/font-size-ui/font-size-ui |
| 22:50 | <Hixie> | er |
| 22:50 | <Hixie> | -~ |
| 22:50 | <annevk> | this is really weird |
| 22:51 | <Ms2ger> | It is |
| 22:51 | <annevk> | there's a difference between * { font-size:1em } and * { font-size:16px } basically, in Gecko |
| 22:52 | <Ms2ger> | http://meyerweb.com/eric/thoughts/2010/02/12/fixed-monospace-sizing/ |
| 22:55 | <annevk> | but he's wrong |
| 22:56 | <annevk> | well, I think he's right in what happens, but there's no justification coming from "medium" |
| 22:57 | <annevk> | because as Hixie just said, the computed value is absolute |
| 22:57 | <Hixie> | 'medium' is absolute |
| 22:57 | <annevk> | oh really? |
| 22:58 | <Hixie> | however, while i could believe that browsers compare the font size to the 'medium' pref size, there's no direct evidence that they do that, only cirucmstancial |
| 22:58 | <Hixie> | i'd be interested in finding out what they actually do |
| 22:58 | <annevk> | CSS 2.1 says "Computed value: absolute length" |
| 22:58 | <annevk> | medium is an absolute-size, but not a length |
| 22:58 | <Hixie> | i'd be even more interested in just having all the browsers default to 16px for both serif and monospace and not let anyone change it (outside of setting a default zoom) |