| 00:00 | gsnedders | wonders how to start parsing the parser spec… |
| 00:04 | <othermaciej> | roc: that is really weird |
| 00:05 | <othermaciej> | roc: file a bug / report by email / whatever? it seems like a wonky API design and it seems unlikely it would be needed for compat |
| 00:08 | <zewt> | it's useful, so you can say foo.toggle("x", bool) instead of if(bool) foo.add("x"); else foo.remove("x"); |
| 00:08 | <Hixie> | should have been foo.set(x, bool) |
| 00:08 | <zewt> | yep |
| 00:09 | <zewt> | iirc i've added DOMTokenList.prototype.set that does exactly that in the past |
| 00:11 | <zewt> | can we have a national death to yoda conditionals day or something |
| 00:12 | <othermaciej> | zewt: I can see how it’s useful, but yeah, it should be a separate function instead of overloading toggle() |
| 00:13 | <zewt> | othermaciej: if it's possible to change it then i'd say yeah, but it's already out there |
| 00:13 | <zewt> | though i'd probably be inclined to still say DOMElementList.prototype.set = DOMElementList.prototype.toggle in my code, to make it clearer |
| 00:17 | <othermaciej> | looks like we have had it a while |
| 00:17 | <othermaciej> | still very silly for toggle() to have a not-toggle mode |
| 00:18 | <zewt> | we all agree on that :) |
| 00:21 | <roc> | I just don't know where to file the bug |
| 00:22 | <zewt> | i assume dom |
| 00:22 | <zewt> | (anne's dom) |
| 00:22 | <zewt> | but is there really any chance of being able to change this? (and is it worth the effort it would take to find out, and change something that is probably implemented everywhere?) |
| 00:27 | <roc> | https://www.w3.org/Bugs/Public/show_bug.cgi?id=25869 |
| 01:24 | <gsnedders> | "The algorithm for the outline also associates each node in the DOM tree with a particular section and potentially a heading." |
| 01:24 | <gsnedders> | Liar! |
| 01:25 | <gsnedders> | It only associates element nodes! |
| 02:06 | <MikeSmith> | gsnedders: if you're reading the outline algorithm right now, I wonder if you might check and see what outline the following should produce: |
| 02:06 | <MikeSmith> | <!DOCTYPE html><section><h1>section</h1></section><h1>top-level</h1> |
| 02:07 | <MikeSmith> | gsnedders: because your outliner currently gives: |
| 02:07 | <MikeSmith> | 1. top-level |
| 02:07 | <MikeSmith> | 1. section |
| 02:08 | <MikeSmith> | gsnedders: but the outliner in the validator gives: |
| 02:09 | <MikeSmith> | 1. [implied heading] |
| 02:09 | <MikeSmith> | 1. section |
| 02:09 | <gsnedders> | My outliner is massively out of date wrt the spec |
| 02:09 | <MikeSmith> | 1. top-level |
| 02:10 | <MikeSmith> | gsnedders: well on that one point I think the spec may have always had it that way and your outliner implemented it wrong |
| 02:10 | <MikeSmith> | but not sure because h5o has it wrong also I think |
| 02:10 | <gsnedders> | h5o? |
| 02:11 | <MikeSmith> | gsnedders: https://code.google.com/p/h5o/ |
| 02:12 | <MikeSmith> | gsnedders: I'm not positive but I recall vaguely recall that when I checked the spec before, I found that my outliner code conforms to the spec, and also recal that when I asked Hixie he said yeah I had it right |
| 02:25 | <MikeSmith> | annevk: about html5-tracker, I wonder if you'd have less of a hard time if you hooked it into the git html-mirror repo that foolip set up, rather than having it keep using svn |
| 02:25 | <MikeSmith> | annevk: and no I'm not volunteering |
| 03:03 | <gsnedders> | Hixie: from the POV of generating code from spec text, it'd be easier if the bogus comment state and the CDATA section state followed the model of everything else instead of magically consuming large chunks… |
| 03:37 | <Hixie> | gsnedders: yeah, i noticed a few things like that myself |
| 03:37 | <Hixie> | gsnedders: my tokeniser outputs entirely different tokens when it comes to text nodes than the spec claims that it should, too |
| 04:09 | <MikeSmith> | gsnedders: btw about the thing outline-algorithm thing I mentioned earlier, https://www.w3.org/Bugs/Public/show_bug.cgi?id=20068 is where I discussed it with Hixie last year and I see from that I ended up changing my outliner code to conform to the spec |
| 06:29 | <annevk> | TabAtkins: oh, I thought you might have some input as to what the platform needs from ES |
| 06:35 | <TabAtkins> | annevk: Ah, Kk. Didn't recognize it as a request for feedback. |
| 06:38 | <annevk> | TabAtkins: I'm trying to capture some needs from platform and then I hope to find someone to present them to TC39 |
| 06:45 | <TabAtkins> | Yeah, cool. Need to add "some way to legitimately subclass Maps and Sets". |
| 06:53 | <annevk> | TabAtkins: can you dig up a pointer for additional rationale on that? might not entirely be self-evident |
| 06:57 | <TabAtkins> | Yeah, can do. |
| 07:05 | <annevk> | MikeSmith: yeah, foolip suggested the same and is not volunteering either |
| 07:15 | <annevk> | jorendorff: I don't recall I was tasked with talking to Hixie, I'm not even sure I know enough about modules to do so :-) |
| 07:15 | <annevk> | jorendorff: I've been encouraging dherman to file a bug or write an email, not sure why he hasn't done it |
| 07:16 | <annevk> | jorendorff: seems a bit weird to talk about <module> at conferences without it having ever been given as feedback on the spec |
| 07:30 | <annevk> | foolip: https://bugzilla.mozilla.org/show_bug.cgi?id=664544 |
| 07:35 | <foolip> | annevk: thanks |
| 07:36 | <foolip> | annevk: can you post that blink-dev if you think it'd help? |
| 07:36 | <annevk> | foolip: sure, pretty sure I pushed for removal at Opera around that time |
| 07:36 | <foolip> | annevk: btw, you reported the bug that got it removed in core :) |
| 07:36 | <annevk> | heh |
| 09:46 | <jgraham> | annevk: So are you going to review https://critic.hoppipolla.co.uk/r/1619 ? |
| 09:50 | <annevk> | jgraham: looks good, do you want to reference http://url.spec.whatwg.org/#concept-url-parser and http://xhr.spec.whatwg.org/#the-open%28%29-method maybe? |
| 09:52 | <annevk> | jgraham: done |
| 09:52 | <jgraham> | annevk: Thanks! |
| 09:53 | <annevk> | jgraham: basically asking me on IRC is the best way to get me to do one of these (provided they're not too big) |
| 09:53 | <jgraham> | annevk: Understood |
| 09:58 | <Ms2ger> | Aha |
| 09:58 | Ms2ger | pulls up the list |
| 09:59 | <jgraham> | annevk: Also, schoolboy error saying that with Ms2ger in the channel ;) |
| 10:00 | <annevk> | MikeSmith: foolip: your "use git" idea is interesting, but how does that work with the existing SVN IDs that we use? |
| 10:00 | <annevk> | MikeSmith: foolip: they are in the commit messages, but I don't think there's a way I can invoke git diff and make that work, is there? |
| 10:01 | <jgraham> | annevk: Context? |
| 10:02 | <annevk> | jgraham: I'm trying to figure out if I create a git clone https://github.com/whatwg/html-mirror/ if I can build web-apps-tracker on top |
| 10:02 | <MikeSmith> | annevk: no way that I know of, but I'm not a git expert |
| 10:02 | MikeSmith | looks through git-svn docs |
| 10:04 | <jgraham> | I'm pretty sure git-svn must maintain a mapping, but if that's using the github stuff I don't know if it's exposed |
| 10:05 | <MikeSmith> | right yeah |
| 10:05 | <jgraham> | (that is, if that mirror is based on github running git-svn rather than you running it) |
| 10:05 | <annevk> | I don't have git-svn on DreamHost |
| 10:14 | <MikeSmith> | annevk: well if you did have git-svn apparently you could get the svn revision number by giving the git has to "git svn find-rev" |
| 10:14 | <annevk> | MikeSmith: how about the other way around? |
| 10:15 | <annevk> | MikeSmith: what I want is, given two svn revisions, get a diff |
| 10:15 | <MikeSmith> | same thing but preface the arg with "r" I think |
| 10:16 | <annevk> | that's hanging the server :-) |
| 10:16 | <annevk> | this does not seem like a cheap command |
| 10:16 | <MikeSmith> | :( |
| 10:16 | <annevk> | I could cache the return values I suppose, but r8000 is still running... |
| 10:16 | annevk | fetches more tea |
| 10:17 | <annevk> | "Unable to determine upstream SVN information from HEAD history" |
| 10:17 | <annevk> | well that's great |
| 10:19 | <jgraham> | Uh, this must be doing a linear search of the commits or something |
| 10:19 | <jgraham> | Yeah, I get the same result |
| 10:19 | <darobin> | it's possibly faster to get the information from searching the git log... |
| 10:19 | <annevk> | even for a fairly recent commit it fails hard |
| 10:20 | <jgraham> | git log --grep="git-svn-id: http://svn.whatwg.org/webapps@8000"; |
| 10:20 | <darobin> | annevk: git log --grep='http://svn.whatwg.org/webapps@8000'; |
| 10:20 | <darobin> | duh |
| 10:21 | <darobin> | it's pretty instant |
| 10:21 | <darobin> | except it'll fail for numbers < 1000 |
| 10:21 | <jgraham> | Add --format=format:%H |
| 10:22 | <darobin> | you have to append a space |
| 10:22 | <annevk> | yes |
| 10:22 | <annevk> | then parse the return value |
| 10:24 | <jgraham> | No, just add the --format I suggested |
| 10:25 | <annevk> | sorry jgraham, missed your suggestion |
| 10:26 | <annevk> | git log --grep="git-svn-id: http://svn.whatwg.org/webapps@{X} " --format=format:%H |
| 10:29 | <jgraham> | Yes |
| 10:29 | <annevk> | Anyone know enough Bash so we can turn this into a gitsvndiff script? |
| 10:30 | <annevk> | And then I can invoke that from Python and woohoo, tower of hacks |
| 10:31 | <jgraham> | Why not just write it in python? |
| 10:33 | <annevk> | jgraham: then I'd have to install GitPython or some such; I rather just have something that works with the default DreamHost setup |
| 10:38 | <MikeSmith> | I'm sure there are people around here who are better at shell scripting than me |
| 10:38 | <MikeSmith> | hint hint |
| 10:38 | <jgraham> | annevk: git_sha1 = subprocess.check_call(["git", "log", "--grep='git-svn-id: http://svn.whatwg.org/webapps%d '" % (svn_id,), "--format=format:%H"]) |
| 10:38 | <MikeSmith> | if not I can write a shell script for it |
| 10:40 | <jgraham> | s/check_call/check_output/ |
| 10:40 | <annevk> | jgraham: is it better to do that than have all the subprocesses in a single script? |
| 10:41 | <jgraham> | Oh and I missed an @ |
| 10:42 | <jgraham> | annevk: the overhead of starting a subprocess isn't going to kill you here |
| 10:42 | <jgraham> | I think doing the simplest thing is best, and I'm sure the simplest thing doesn't involve bash |
| 10:43 | <MikeSmith> | yeah I agree strongly |
| 10:44 | <jgraham> | (oh and no need for the quotes in the grep argument ofc) |
| 10:44 | <jgraham> | subprocess.check_output(["git", "log", "--grep=git-svn-id: http://svn.whatwg.org/webapps@%s " % (8000,), "--format=format:%H"]) |
| 10:44 | <jgraham> | actually runs |
| 10:57 | <annevk> | jgraham: had to use os.popen() as that command wasn't supported |
| 10:58 | <jgraham> | annevk: What version of python |
| 10:58 | <jgraham> | ? |
| 10:58 | <jgraham> | Also, I assume you imported subprocess? |
| 10:59 | <annevk> | 2.6.6 |
| 10:59 | <annevk> | did |
| 10:59 | <jgraham> | (popen is the wrong choice pretty much always; if you don't have check_process, you should probably use .call()) |
| 10:59 | <annevk> | well I guess I'm first going to make this work and then accept pull requests |
| 11:01 | <annevk> | already 1PM :( |
| 11:01 | <jgraham> | Python 2.6 is about 4 years out of date and 2.6.6 appears to not even be the latest 2.6 release :( |
| 11:06 | <annevk> | It's clearly a sad day |
| 11:09 | <odinho> | omfg, python 2.6. blast from the past. |
| 11:11 | <jgraham> | yeah it's so old annevk is actually programming with a beard and flares |
| 11:12 | <odinho> | Sounds like the annevk I know. |
| 11:33 | <foolip> | annevk: the ids are in the commit message, but it could be worth creating tags to each commit (or something like that) to make it quicker to find a commit by SVN number |
| 11:33 | <foolip> | annevk: is the current web-apps-tracker experiencing problems, other than the ever-growing diff cache? |
| 11:33 | <annevk> | foolip: it's slow and it's a perf problem for Hixie |
| 11:34 | <annevk> | If my git repo is in ./X/ |
| 11:34 | <annevk> | How do I run git commands in ./? |
| 11:34 | <annevk> | --exec-path and such don't seem to work as I expect |
| 11:34 | <foolip> | annevk: GIT_DIR probably? |
| 11:34 | foolip | experiments |
| 11:34 | <jgraham> | annevk: If you are running from a script, you can just set cwd |
| 11:35 | <jgraham> | .call(["git", "whatever"], cwd="./") or something |
| 11:35 | <jgraham> | Er |
| 11:35 | <jgraham> | .call(["git", "whatever"], cwd="./X") or something |
| 11:35 | <annevk> | jgraham: if I want to use .call(), how do I get the output? |
| 11:37 | <foolip> | annevk: git --git-dir=./opus/.git --work-tree=./opus log also worked for me (opus is your X) |
| 11:38 | <jgraham> | Hmm, so the documentation recommends stdout_data, stderr_data = subprocess.Popen(["git", "foo"], cwd="./X", stdout=subprocess.PIPE).communicate() |
| 11:40 | <annevk> | foolip: cool |
| 12:00 | <foolip> | annevk: if you intent do try this out and don't want to generate the refs for svn revision numbers on your end, I could put them in the html-mirror repo itself |
| 12:00 | <foolip> | just let me know |
| 12:00 | <annevk> | foolip: I'll let this channel know whenever I get stuck |
| 12:01 | <annevk> | mostly just trying to get the tool to do the same thing at the moment |
| 12:28 | <jorendorff> | annevk: oh, ok, sorry for the mistake (re modules) |
| 12:31 | <annevk> | jorendorff: no worries, happy to help out if that's needed |
| 13:15 | <annevk> | foolip: I think I got something working that can replace the current setup, the only thing that's needed is some kind of hook that updates all the git stuff once the spec is updated |
| 13:33 | <foolip> | annevk: I have a thing that updates the git mirror automatically (Hixie added a ping URL to his publish script), would it suffice to have a web hook on the GitHub html-mirror repo? |
| 13:34 | <foolip> | I'm guessing you already have a hook that Hixie pings, but which be racy with my update script |
| 13:34 | <annevk> | foolip: yeah that might work |
| 13:34 | <annevk> | foolip: need to sort a few things still |
| 13:42 | <foolip> | annevk: you probably have access to https://github.com/whatwg/html-mirror/settings/hooks/new when you need it |
| 13:42 | <annevk> | foolip: good point |
| 13:46 | <jgraham> | foolip: Ever tried to take all the commits touching a set of directories /A /B /C in one git repo and add them to a second repo at a different path like /prefix/A' /prefix/B' /prefix/C'? I sort of have the feeling it is possible, perhaps using filter-branch, but I haven't actually tried it |
| 13:46 | <foolip> | jgraham: yes, I've done that |
| 13:46 | <foolip> | step 1 is to use git filter-branch --subdirectory-filter A/ |
| 13:48 | <annevk> | foolip: so basically all I need is a script that does git --git-dir=html-mirror/.git pull |
| 13:48 | <foolip> | jgraham: step 2 depends on what kind of history you want in the final repo, but just fetching the filtered branch from the target repo and merging it with subtree something something might work |
| 13:49 | <foolip> | with extra complexity if you need to keep it in sync and not just do it once |
| 13:49 | <foolip> | annevk: what language would the hook be implemented in? |
| 13:51 | <foolip> | in any event, what I've done is to have a script which creates a batch job to do the real work, as opposed to doing it in web-facing script directly |
| 13:51 | <jgraham> | foolip: Yeah, I think the problem is that I don't want to do it just once, so I have to be certain that the commit ids don't change |
| 13:52 | <jgraham> | and, preferably, avoid doing the work N times. |
| 13:52 | <jgraham> | Well N**2 times |
| 13:52 | <foolip> | jgraham: and it's important to remove everything except the filtered directory, to save space? |
| 13:52 | <foolip> | otherwise you could just merge the original repository and remove the bits you don't want to see |
| 13:52 | <annevk> | foolip: shell |
| 13:53 | <annevk> | foolip: I just want to sync |
| 13:53 | <jgraham> | foolip: Interesting idea |
| 13:53 | <jgraham> | pull in the whole other repo and then add a commit to delete the unwanted parts? |
| 13:54 | <foolip> | jgraham: yeah, or do the whole move+delete dance as part of the merge commit, whichever gives you the prettiest history in your mind |
| 13:55 | <foolip> | annevk: well, maybe try the simple thing first and improve if needed :) |
| 13:56 | <annevk> | Man, my beta version is already so much faster than the current web-apps-tracker setup |
| 13:56 | <foolip> | that's not surprising :) |
| 13:59 | <foolip> | annevk: if I get hit by a bus and html-mirror stops updating, https://github.com/foolip/html-webhook is the thing that keeps it in sync |
| 14:01 | <annevk> | http://html5.org/tools/web-apps-tracker now runs on the new script, no auto-update yet |
| 14:01 | <annevk> | if you view a diff you can use &context to get more context |
| 14:10 | <MikeSmith> | nice |
| 14:11 | MikeSmith | tries ?from=1&to=8648 |
| 14:27 | <annevk> | MikeSmith: killed that |
| 14:28 | <darobin> | heh |
| 14:31 | <MikeSmith> | annevk: smart man |
| 14:40 | <jgraham> | Is that different to just git diff [first sha1].. ? |
| 14:59 | <annevk> | jgraham: ooh, can I kill the second lookup? I didn't even try! |
| 15:05 | <jgraham> | annevk: Well in the case it is HEAD, yes |
| 15:05 | <annevk> | jgraham: it almost never is |
| 15:06 | <jgraham> | So I don't understand what's slow with MikeSmith's query. Is it formatting the diff? |
| 15:06 | <jgraham> | Because getting the diff from git seems to be more or less instant |
| 15:08 | <annevk> | jgraham: oh I killed the feature |
| 15:09 | <annevk> | jgraham: because it was rarely used |
| 15:09 | <zewt> | man, the spec is sort of unusable for me right now, freezes in chrome constantly |
| 15:09 | <jgraham> | annevk: Oh. I guess if I used this at all I would use that |
| 15:12 | <annevk> | Oh. |
| 15:19 | <foolip> | annevk: cool, it feels pretty fast now |
| 15:19 | <foolip> | too bad I can no longer use it to check if html-mirror is in sync with svn :) |
| 15:19 | <foolip> | gotta go, have a nice weekend! |
| 15:21 | <dglazkov> | good morning, Whatwg! |
| 15:22 | <annevk> | Hixie: no longer hammering your SVN setup |
| 15:28 | <annevk> | If people want to change things around, now you can https://github.com/whatwg/web-apps-tracker |
| 16:11 | <annevk> | Hixie: as for talking about cleanup, prolly next week somewhere, your Wednesday morning is usually convenient |
| 16:36 | <Ms2ger> | wat: http://lists.w3.org/Archives/Public/www-style/2014May/0303.html |
| 17:47 | <Hixie> | i don't understand why chrome is taking so long to apply the :hover effects on the spec now |
| 17:47 | <Hixie> | it didn't used to |
| 17:47 | <Hixie> | there's no transitions |
| 17:48 | <Hixie> | i mean i'm talking like 4 seconds |
| 17:48 | <dglazkov> | esprehn: ^^^ |
| 17:50 | <SamB> | Hixie: downgrade! |
| 17:50 | <dglazkov> | Hixie: can you file a bug? |
| 17:53 | <Hixie> | sure |
| 17:55 | <dglazkov> | yay |
| 17:55 | <Hixie> | https://code.google.com/p/chromium/issues/detail?id=376847 |
| 17:56 | <Hixie> | man that's really annoying |
| 17:56 | <Hixie> | wtf |
| 17:58 | <TabAtkins> | Ms2ger: WAT |
| 18:38 | <mathiasbynens> | TypeError: 'null' is not an object (evaluating 'configUI.appendChild') |
| 18:38 | <mathiasbynens> | http://www.whatwg.org/specs/web-apps/current-work/updater.js:71 in initUpdater |
| 18:38 | <mathiasbynens> | http://www.whatwg.org/specs/web-apps/current-work/updater.js:84 |
| 18:38 | <mathiasbynens> | heads up Hixie ^ |
| 18:53 | <mathiasbynens> | Hixie: http://whatwg.org/html/tokenization.html#table-charref-overrides includes “0x000D to 0x001F” but http://whatwg.org/html/parsing.html#preprocessing-the-input-stream only has “U+000E to U+001F”. Is U+000D missing from the latter by accident, or is this intentional? |
| 18:54 | <mathiasbynens> | this is the only difference between those two lists of invalid code points |
| 19:07 | <jsbell> | mathiasbynens: There's a whole paragraph on U+000D in parsing - that appears to answer the question ("there are never any CR characters in the input to the tokenization stage") |
| 19:07 | <mathiasbynens> | jsbell: ah, gotcha. thanks! |
| 19:30 | <Hixie> | mathiasbynens: yeah, known issue (the updater.js thing) |
| 19:30 | <Hixie> | mathiasbynens: it's intentional, you shouldn't ever got to an U+000D at that point in the second list |
| 19:45 | <IZh> | Hi. If you play several tracks with different length simultaneously, will it lasts until the end of longest track? |
| 19:45 | <SamB> | IZh: what do you mean? |
| 19:46 | <IZh> | I mean long video and short audio. |
| 19:46 | <SamB> | are they being played in a coordinated fashion, or independantly? |
| 19:46 | <IZh> | On the same media controller. |
| 19:47 | <SamB> | okay, so you've reason to expect there to be interaction ;-) |
| 19:49 | <IZh> | SamB: You are right. This will be better than playing silent video by default. :-) |
| 19:52 | <SamB> | IZh: http://www.whatwg.org/specs/web-apps/current-work/multipage/the-video-element.html#media-controller-duration would seem to indicate that it takes the union of the timelines of the individual resources |
| 19:53 | <SamB> | Hixie: hmm, I kind of miss the glow effect on the target element |
| 19:53 | <IZh> | SamB: The union? Not the maximum? Hmm... Is it possible to make a playlist with sequential playing? |
| 19:54 | <Hixie> | SamB: people wanted no yellow, so... |
| 19:54 | <SamB> | aww |
| 19:55 | <TabAtkins> | I found the yellow glow pretty ugly. The arrow in the margin is much better. ^_^ |
| 19:56 | <SamB> | I don't really care about the color, I guess I just liked the part where it highlighted the phrase in question |
| 19:57 | <SamB> | also the arrow doesn't look too accurate? |
| 19:57 | <Ms2ger> | Make it papayawhip |
| 19:58 | <SamB> | http://www.whatwg.org/specs/web-apps/current-work/multipage/the-video-element.html#media-controller-duration has the arrow pointing a line below where I'd expect |
| 19:58 | <SamB> | at least, that's how it looks to me |
| 19:59 | <Hixie> | i can move the arrow up a bit, hold on |
| 19:59 | <SamB> | (firefox 24 ESR, looks like no blocked script) |
| 20:00 | <TabAtkins> | SamB: It's just a bit of CSS, nothing to do with script. |
| 20:00 | <SamB> | ah, that's more like I'd expect |
| 20:00 | <Hixie> | fixed |
| 20:01 | <SamB> | of course it looks a bit clipped now since you can't tell the browser "only please scroll up to avoid clipping the arrow" |
| 20:01 | <SamB> | but that's life for you |
| 20:03 | <Hixie> | yeah not much i can do about that |
| 20:12 | <Hixie> | Domenic: (just reading some of the bugs liek https://www.w3.org/Bugs/Public/show_bug.cgi?id=20925 ...) oh man, please don't change webidl again |
| 20:12 | <Hixie> | Domenic: every time webidl changes (as opposed to fixing bugs or adding features) i end up with massive amounts of make-work that do nothing to improve the platform |
| 20:12 | <SamB> | Hixie: you could *make* a way ;-P |
| 20:13 | <Hixie> | SamB: that's more a CSS thing, bug zcorpan or TabAtkins :-P |
| 20:14 | <SamB> | hmm, isn't there a nasty thing that happens with pseudo-framed pages and anchors, too? |
| 20:14 | <TabAtkins> | What do you mean? |
| 20:14 | <SamB> | where there's a fixed box that covers the link target? |
| 20:18 | <TabAtkins> | SamB: Still not understanding. |
| 20:18 | <SamB> | hmm |
| 20:21 | <SamB> | TabAtkins: say http://web.archive.org/web/20140506093351/http://www.whatwg.org/specs/web-apps/current-work/multipage/the-video-element.html#user-interface |
| 20:24 | <TabAtkins> | SamB: So what's the problem? |
| 20:24 | <SamB> | TabAtkins: the target is under the thing at the top there |
| 20:24 | <TabAtkins> | Oh, yes, indeed. |
| 20:25 | <SamB> | I think I've seen such on sites that actually have such headers of their own, but couldn't remember any off the top of my head |
| 20:30 | <TabAtkins> | Yeah, I've seen that before. It's often due to people wanting to do a variable-height header with scrollable thing underneath, which wasn't really possible before flexbox. |
| 20:38 | <Hixie> | is es-discuss⊙mo the way to contact the people behind the js module stuff? |
| 20:39 | <Ms2ger> | I believe so |
| 20:40 | <TabAtkins> | Yup. |
| 20:56 | <Ms2ger> | "If a browsing context P has a Document D with an element E that nests another browsing context C inside it, then C is said to be nested through D, and E is said to be the browsing context container of C." |
| 21:36 | <smaug____> | why it is so difficult to understand that sync XHR sucks really really badly |
| 21:39 | <TabAtkins> | smaug____: Because it's easier to use if you don't care about responsiveness at all. |
| 21:43 | gsnedders | wonders if we can use some NLP toolkit to find sentences in the spec with ambiguous parse-trees. |
| 21:46 | <SamB> | gsnedders: how would you know which ones were ambiguous in a problematic fashion? |
| 21:46 | <gsnedders> | SamB: Many NLP toolkits give you all possible parse-trees, so if a sentence permits > 1 parse-tree, it's ambiguous |
| 21:49 | <gsnedders> | But ambiguity isn't per-se a problem, because often the grammatically ambiguous cases are semantic nonsense. |
| 21:51 | <gsnedders> | like the typical example of "I shot an elephant wearing my pyjamas", which can be parsed such that the subject is wearing their pyjamas, or the elephant is wearing their pyjamas. Obviously the second is nonsense. |
| 21:54 | <Hixie> | the _second_ is the nonsense one?! |
| 21:54 | <gsnedders> | Hixie: when did you last see an elephant wearing pyjamas? |
| 21:55 | <Hixie> | i think it's a hell of a lot more likely that someone would shoot an elephant that was wearing their pyjamas, than it is that an elephant would surprise someone in pyjamas and that that person would just shoot them |
| 21:55 | <Hixie> | when did you last see someone in pyjamas and an elephant at the same time? |
| 21:55 | <gsnedders> | In the film that that quote comes from? |
| 21:56 | <Hixie> | oh well if we're including fiction, then i've almost certainly seen elephants wearing pyjamas |
| 21:56 | <TabAtkins> | How would the elephant even get my pajamas on? |
| 21:56 | <Hixie> | kid's books, cartoons... |
| 21:56 | <Hixie> | kids' books, even |
| 21:57 | <gsnedders> | TabAtkins: I think that's rather the point :P |
| 21:57 | <TabAtkins> | I know, I was directing the question at Hixie. |
| 21:58 | <gsnedders> | I need to be trying to do fewer things at once. This is impractical. |
| 21:58 | <gsnedders> | I should just leave my spec -> code converter till later. |
| 22:13 | <zewt> | a spec -> code converter is more commonly referred to as "a programmer" |
| 22:19 | <SamB> | zewt: he wants the spec to BE a program, clearly |
| 22:20 | <jgraham> | gsnedders is tyring to replace himself with a short shell script |
| 22:21 | <zewt> | surely perl |
| 22:37 | <jgraham> | zewt: He's just a masochist, he's not mad |