| 00:40 | <MikeSmith> | Hixie: requests to wiki.whatwg.org hanging for a long time waiting for response |
| 00:40 | <Hixie> | yeah i noticed that myself a few minutes ago |
| 00:41 | <MikeSmith> | where long is like 10s or more |
| 00:41 | <Hixie> | load average is really high on the machine |
| 00:41 | <Hixie> | no idea why |
| 00:41 | <MikeSmith> | maybe getting hit by a crawler |
| 00:42 | <Hixie> | some weird process running as root |
| 00:43 | <Hixie> | i'm either being attacked or dreamhost is checking for attacks, maybe |
| 00:47 | <MikeSmith> | oh wow |
| 00:47 | <MikeSmith> | well I hope it's dreamhost |
| 00:47 | <Hixie> | yeah, i think it is |
| 00:49 | <Hixie> | looks like a huge grep command looking for infected php files |
| 00:54 | <MikeSmith> | ah |
| 00:56 | <Hixie> | here's a sample command line: |
| 00:56 | <Hixie> | grep -Zl coder by Orlenok.do.am\|ru-otveti1.cc\|mail.ru-0tveti.com\|mail.ru-otveti2.cc\|ru-0tveti1.com\|com-app-view.cc\|homeincometds.ru\|workathome55.ru\|workathomejournal.. |
| 01:01 | <Hixie> | yeah it has to be dreamhost. the process is running as a subprocess of ssh, so if it was an attacker, they wouldn't need to grep for anything, they could just run their malware. |
| 01:07 | <zewt> | who needs the internet to dos a server, when the provider is doing it for them |
| 01:36 | <MikeSmith> | hahah |
| 01:37 | <MikeSmith> | indeed |
| 01:44 | <MikeSmith> | I guess some would argue you're dos'ing yourself by running php-based services at all |
| 01:45 | <MikeSmith> | could be even worse I guess, could be running wordpress too |
| 01:47 | <MikeSmith> | tantek: Is there a way I can have the microformats.org e-mail me a notification whenever a page I'm watching is changed? |
| 01:47 | <MikeSmith> | * microformats.org wiki |
| 01:48 | <MikeSmith> | or a feed I can subscribe to? |
| 01:48 | <tantek> | MikeSmith - not that I know of - but there is a feed of recent changes, that you could filter through for changes to the page |
| 01:48 | <MikeSmith> | naomi ok |
| 01:48 | <MikeSmith> | oops |
| 01:48 | <MikeSmith> | tantek: ok I'll try the recent-changes feed |
| 02:11 | <Hixie> | MikeSmith: we run both wordpress and mediawiki :-| |
| 02:21 | <MikeSmith> | Hixie: way to live dangerously man |
| 02:36 | <tantek> | MikeSmith - it's a common pattern :) we setup the same thing on microformats.org |
| 02:37 | <tantek> | though with indiewebcamp we ditched wordpress |
| 02:37 | <tantek> | because of course, everyone should just blog on their own domain ;) |
| 02:37 | <MikeSmith> | yeah I'm just trolling |
| 02:38 | <MikeSmith> | but I am looking forward to the day when better Web software stacks are more common and easier to deploy and maintain |
| 02:38 | <tantek> | MikeSmith - you're welcome to lurk in #indiewebcamp and watch us develop multiple such stacks ;) |
| 02:39 | <MikeSmith> | tantek: I will drop by |
| 02:42 | <Hixie> | MikeSmith: did you hear about facebook's "Hack"? |
| 02:42 | <MikeSmith> | Hixie: no |
| 02:42 | <MikeSmith> | programming language? |
| 02:42 | <Hixie> | http://code.facebook.com/posts/264544830379293/hack-a-new-programming-language-for-hhvm/ |
| 02:42 | <Hixie> | it's... php. |
| 02:43 | <tantek> | hey - I hear that's Tim Bray's favorite ;) |
| 02:43 | <MikeSmith> | wow it really is still just PHP basically |
| 02:44 | <Hixie> | yeah. |
| 02:44 | <MikeSmith> | https://fbcdn-dragon-a.akamaihd.net/hphotos-ak-ash3/t39.2365-6/851566_537609726356988_1715700065_n.jpg |
| 02:45 | <MikeSmith> | way to break new ground! |
| 02:45 | <Hixie> | also what's with people being unable to come up with good programming language names |
| 02:45 | <Hixie> | "go", "hack" |
| 02:45 | <Hixie> | have you any idea how unsearchable those terms are |
| 02:46 | <tantek> | Hixie: cassis.js :) |
| 02:46 | <MikeSmith> | yeah really dumb name choices |
| 02:46 | <Hixie> | [programming language hack] is not going to help anyone |
| 02:46 | <MikeSmith> | yeah what's another word for "sunk cost"? |
| 02:46 | <MikeSmith> | or "keep throwing good money after bad" |
| 02:47 | <Hixie> | i actually can't decide if "go" is a worse name or "hack". go is basically a stop word. "hack" is a word used for every programming language exploit. |
| 02:47 | <MikeSmith> | seriously there is some term for this, where you're deeply invested in some bad software choice that you can't replace it |
| 02:49 | <MikeSmith> | Erlang is the best language name |
| 02:49 | <MikeSmith> | as far as languages that are actually intended to be used |
| 02:50 | <MikeSmith> | hey somebody should name a language "Kerrang!" |
| 02:50 | <zewt> | fca87fee-6c22-4533-8001-8ab6d8630f13++ |
| 02:52 | <MikeSmith> | INTERCAL's a pretty good name too, because it sounds like other older lanaguge names, and sounds like it's legitimately intended to be useful |
| 02:52 | <zewt> | or the nonsensically named git product named "stash" |
| 02:52 | <zewt> | why not call it "commit" while you're at it |
| 02:52 | <MikeSmith> | heh |
| 03:10 | <Hixie> | MikeSmith: "ada", "pascal", "fortran", "cobol" are all pretty googleable names |
| 03:10 | <Hixie> | "perl" too |
| 03:10 | <Hixie> | and "python", in context |
| 03:15 | <MikeSmith> | Hixie: yeah but all those names lack expressiveness |
| 03:15 | <Hixie> | heh |
| 03:15 | <MikeSmith> | a highly expressive language should have a highly expressive name |
| 03:15 | <MikeSmith> | like "Khaaaaaaaan!" |
| 03:40 | <MikeSmith> | Hixie: in other news I'm trying to figure out how to implement the spec requirements for "shorcut icon" without resorting to a regexp |
| 03:41 | <MikeSmith> | ーIf the "shortcut" keyword is present, it must be come immediately before the icon keyword and the two keywords must be separated by only a single U+0020 SPACE character. |
| 03:41 | <MikeSmith> | I mean that part |
| 03:42 | <MikeSmith> | hmm |
| 03:44 | <MikeSmith> | I guess it's easy enough if I just save the value of the previous token for each time I check the next one |
| 03:52 | <MikeSmith> | it hurts when the solution is "create yet another string" |
| 03:53 | <Hixie> | MikeSmith: do you know the offset of the start of the keyword? |
| 03:53 | <Hixie> | MikeSmith: if so, you can just do a substring search and a previous-keyword search |
| 03:54 | <MikeSmith> | Hixie: yeah I can determine the offset |
| 03:54 | <MikeSmith> | maybe that way would be better |
| 03:54 | <Hixie> | (but that's a weird requirement. maybe it should just be a keyword immediately before) |
| 03:56 | <MikeSmith> | Hixie: ah yeah |
| 03:57 | <MikeSmith> | that would definitely be better |
| 03:57 | <Hixie> | (file a bug if you want that, i'm kinda afk this week) |
| 03:57 | <MikeSmith> | I realize now that's actually what had been preventing me from implementing a check for this |
| 03:57 | <MikeSmith> | k |
| 03:57 | <Hixie> | i'd want to check why i had written it like that before |
| 03:57 | <MikeSmith> | yeah sure |
| 03:57 | <Hixie> | seems like a really weird requirement, not the kind of thing i'm likely to put in on a whim |
| 03:58 | <MikeSmith> | hmm yeah maybe IE doesn't actually recognize it if there are multiple spaces or a newline or tab there |
| 03:58 | <Hixie> | maybe, i dunno |
| 03:58 | <Hixie> | though maybe we no longer care if it's old IE? |
| 03:58 | <Hixie> | who knows |
| 03:59 | <MikeSmith> | yup anyway I'll get the bug raised so we don't forget |
| 05:02 | <SamB> | Hixie: are you not allowed to explain stuff like that somewhere in the spec itself for some reason? |
| 05:03 | <Hixie> | i try not to include too much rationale in the spec because otherwise it gets out of control and makes the spec unreadable |
| 05:04 | <Hixie> | i sometimes leave it in comments |
| 09:45 | <zcorpan> | Hixie: MikeSmith: "shortcut icon" is for ie compat |
| 09:46 | <MikeSmith> | yeah I know |
| 09:46 | <zcorpan> | http://web.archiveorange.com/archive/v/IJ6c3Cpqf8c6cE1xi7N3 |
| 09:46 | <MikeSmith> | question Hixie and I have is, does IE really require that there be only a single space there |
| 09:47 | <MikeSmith> | will it work if it's multipel spaces or tabs or newline |
| 09:47 | <MikeSmith> | other whitespace |
| 09:47 | <zcorpan> | see above |
| 09:47 | <MikeSmith> | if so then it's a lot easier for me to implement a check for it in the validator |
| 09:47 | <MikeSmith> | ah OK |
| 09:47 | MikeSmith | reads |
| 09:49 | <MikeSmith> | zcorpan: doesn't seem to be a definitive answer there about whether IE still works as expected with e.g. rel="shortcut icon" or whatever |
| 09:50 | <zcorpan> | MikeSmith: i guess newer IEs haven't been tested |
| 09:50 | <MikeSmith> | hsivonen demos at http://hsivonen.com/test/moz/shortcut-icon/ don't seem to be checking for the multiple/other whitespace case either |
| 09:51 | <MikeSmith> | zcorpan: blog post from JonathanNeal seems to indicate that at least IE10 still requires "shortcut icon" as opposed to just "icon" |
| 09:51 | <zcorpan> | rel="shortcut icon" with multiple spaces between "shortcut" and "icon" is not IE-compatible.) said hsivonen |
| 09:51 | <MikeSmith> | http://www.jonathantneal.com/blog/understand-the-favicon/ |
| 09:51 | <MikeSmith> | ah OK |
| 09:51 | <MikeSmith> | well shit |
| 09:52 | <MikeSmith> | I guess I'll just give up an use a negative forward lookup regexp then |
| 09:52 | <zcorpan> | i think it makes most sense to check for "shortcut icon" exactly (but ascii-case insensitive), but i guess that doesn't match the current spec |
| 09:53 | <zcorpan> | but we could change the spec to say that :-) |
| 09:53 | <MikeSmith> | yeah but for the validator case that wouldn't win us much |
| 09:54 | <MikeSmith> | we're just tokeninizing the whole value before doing anything else with it |
| 09:54 | <zcorpan> | if (keyword.toAsciiLower() == 'shortcut') assert(originalValue.toAsciiLower() == 'shortcut icon') |
| 09:55 | <MikeSmith> | oh |
| 09:55 | <MikeSmith> | yeah |
| 09:55 | <MikeSmith> | that'd be better |
| 09:55 | <MikeSmith> | sorry I didn't know what you meant |
| 09:55 | <MikeSmith> | yeah really that's what the requirement should be |
| 09:56 | <MikeSmith> | though apparently it doesn't need to be lowercase |
| 09:56 | <zcorpan> | that's why the code is lowercasing it before comparison :-) |
| 09:56 | <MikeSmith> | rel="SHORTCUT ICON" is what the IE does recommend |
| 09:56 | <MikeSmith> | ah OK |
| 09:56 | <MikeSmith> | yeah |
| 09:56 | <MikeSmith> | ok I'll file another bug |
| 09:56 | <MikeSmith> | spec bug |
| 09:57 | <MikeSmith> | actually I'll just change the one I already filed |
| 10:03 | <MikeSmith> | zcorpan: https://www.w3.org/Bugs/Public/show_bug.cgi?id=25151 |
| 10:19 | <MikeSmith> | damn is there really no simple way to find any associated bugs in a chrome code review |
| 10:22 | zcorpan | wonders what's up with r8556 |
| 10:23 | <MikeSmith> | http://html5.org/tools/web-apps-tracker doesn't show an r8556.. |
| 10:24 | <MikeSmith> | oh |
| 10:24 | <MikeSmith> | there was some discussion about that here a couple days ago I htink |
| 10:27 | <MikeSmith> | ーRe-investigate performance regression of moving DOM attributes to prototype chains and try to re-re-re-re-re-reland the CL. |
| 10:28 | <MikeSmith> | groundhog day |
| 10:51 | <MikeSmith> | hsivonen: so bugzilla.validator.nu is still setting a cookie sometimes with ".bugzilla.validator.nu |
| 10:51 | <MikeSmith> | ... as the domain |
| 10:51 | <MikeSmith> | with the leading dot |
| 10:51 | <MikeSmith> | even after I deleted all my bugzilla.validator.nu cookies |
| 10:53 | <MikeSmith> | it's not a big deal -- I just re-delete them again after I notice it starts causing me login problems again |
| 10:53 | <MikeSmith> | just wanted to let you know |
| 10:55 | <MikeSmith> | the problem is that it sets duplicate cookies, one with the domain with that leading dot and one without it |
| 10:56 | <MikeSmith> | if it were setting it once just with the leading dot I guess it would work fine |
| 11:15 | <annevk> | MikeSmith: web-apps-tracker only shows commits to source fwiw |
| 11:32 | <MikeSmith> | ok |
| 11:47 | <nickstenn> | Hi all. I'm doing a bit of research about user(-agent) control over the DOM, and am looking for interesting things to read. |
| 11:47 | <nickstenn> | Specifically, if rollout of CSP is going to kill bookmarklets which load remote resources, what options are available to authors other than vendor-specific browser extensions? |
| 11:51 | <nickstenn> | I guess, to rephrase, the question is as follows -- we are, for good reason, trying to make sure that page authors have more control about what code runs on their pages. At the moment, the only mechanism that users have to run their code on a page (that I know of) is vendor-specific, in the form of a browser extension. |
| 11:52 | <nickstenn> | Does anyone know of proposals to allow user-trusted code to run in the DOM in a standard way? |
| 11:53 | <annevk> | nickstenn: nope |
| 11:53 | <nickstenn> | annevk: yeah, that's kind of the conclusion I'd drawn |
| 13:41 | <ondras> | so, resetStyleInheritance vs. applyAuthorStyles |
| 13:42 | <ondras> | are both still in the spec? which spec is relevant? I am apparently unable to locate the correct one |
| 13:47 | annevk | has never seen those APIs |
| 13:49 | <ondras> | wow. |
| 13:49 | <ondras> | annevk: they are supposed to forbid page CSS to interfere with shadow dom's styling |
| 14:18 | <annevk> | I'm surprised TextDecoder() always eats the BOM |
| 14:19 | <annevk> | That seems completely inappropriate for a number of cases |
| 14:19 | <annevk> | We should add a flag for that |
| 14:20 | <ondras> | annevk: are you aware of any other API to solve the host-css-in-shadow-dom use case? |
| 14:20 | <annevk> | Not really |
| 15:45 | <annevk> | Does IDL have hooks yet for ArrayBuffer handling and such? |
| 15:58 | <TabAtkins> | ondras: Both of those are gone. |
| 15:59 | <TabAtkins> | ondras: Outside CSS just never affects shadow DOM, unless you explicitly reach into the shadow with the ::shadow pseudo-element. |
| 16:00 | <TabAtkins> | Inheritance *does* still work though, unless you explicitly reset it yourself (such as by using "all:initial;" or something). |
| 16:04 | <dglazkov> | good morning, Whatwg! |
| 17:15 | <Domenic_> | TabAtkins: what does "inheritance" mean? |
| 17:28 | <dglazkov> | Domenic_: I think he meant CSS inheritance |
| 17:29 | <Domenic_> | dglazkov: what is CSS inheritance though, and how is it different from outside CSS affecting shadow DOM... |
| 17:30 | <dglazkov> | Domenic_: http://www.w3.org/TR/css3-cascade/#inheriting |
| 17:30 | <Domenic_> | ah ok, so e.g. fonts will affect the shadow DOM |
| 17:30 | <Domenic_> | makes sense |
| 17:59 | <annevk> | Except <input> and such override it :-( |
| 18:47 | <annevk> | People, I rewrote most of Encoding to make a little bit more sense |
| 18:48 | <annevk> | So if you find strange bugs, it's highly likely I just introduced them |
| 18:49 | <annevk> | And if so, please tell me so I can get it back to its former rather polished state |
| 18:53 | <ondras> | TabAtkins: nice, thanks! This kinda makes all available ShadowDOM articles/tutorials obsolete I guess :) |
| 19:23 | SamB | goes in search of the current "shadow DOM"-related spec for those not writing Mozilla Chrome |
| 19:23 | <SamB> | er. |
| 19:23 | <SamB> | S/Chrome/chrome/ |
| 19:24 | <SamB> | or specs ... |
| 19:26 | <SamB> | TabAtkins: is it just me or shouldn't there be a link to the Shadow DOM spec from http://dev.w3.org/csswg/css-scoping/ ? |
| 19:55 | <TabAtkins> | Domenic_: Values will trickle through the tree via inheritance, but selectors can't reach in unless they do so on purpose. |
| 19:56 | <TabAtkins> | SamB: Yeah, there should be, but I didn't want to hunt down a link when I was writing it. |
| 19:57 | <TabAtkins> | ondras: Yeah, that's the peril of writing tutorials for something which is still super-experimental. |
| 19:58 | <SamB> | TabAtkins: maybe http://w3c.github.io/webcomponents/spec/shadow/ or http://www.w3.org/TR/shadow-dom/ would be handy then? |
| 19:58 | <TabAtkins> | dglazkov: Which link should I use? |
| 19:59 | <jsbell> | annevk: I haven't spotted any other glitches in Encoding. Nice refactor. |
| 20:02 | <Domenic_> | No TRs :P |
| 20:04 | <jsbell> | Did Firefox eliminate new-less constructors recently, or was I dreaming? |
| 20:05 | <SamB> | jsbell: in their javascript code, you mean? |
| 20:07 | <jsbell> | SamB: Yeah. Support for calling DOM constructors like functions instead of w/ new operator. Last I checked FF/IE supported that, Safari/Chrome did not. |
| 20:08 | <jsbell> | I thought I saw something fly by on twitter about it, but literally may have dreamt it. |
| 20:08 | <SamB> | jsbell: oh, I thought maybe you meant internal coding style |
| 20:09 | <jsbell> | Ah, found it... https://bugzilla.mozilla.org/show_bug.cgi?id=980945 |
| 20:15 | <SamB> | jsbell: only for typed arrays, then? |
| 20:16 | <jsbell> | SamB: https://twitter.com/annevk/statuses/445864332187168768 and replies made it sound more broad |
| 20:16 | <ondras> | so, the image onload event. is there some way to make sure that actually all the image file data has been transfered? |
| 20:17 | <ondras> | I mean, the browser can fire onload once the image is retrieved and can be displayed |
| 20:17 | <ondras> | but that does not 100% imply that all data has been transfered, right? |
| 20:17 | <jsbell> | SamB: but I see no change in aurora yet, so waiting to see |
| 20:17 | <SamB> | ondras: are you worried it might have loaded only, say, half of the image and then gotten interrupted |
| 20:18 | <ondras> | SamB: no, I am worried that some exif metadata might be located at the end of the file |
| 20:18 | <ondras> | SamB: and I plan on measuring bandwidth by downloading images |
| 20:18 | <ondras> | SamB: so I want to make sure that an X kilobyte file was completely transferred |
| 20:19 | <SamB> | jsbell: I don't see a patch on that bug, anyway, only to fix the in-tree JS |
| 20:24 | <SamB> | ondras: well, the step that sets the state to "completely available" is only supposed to run once the resource has been fetched |
| 20:24 | <SamB> | which is also what sends the load and loadend events |
| 20:25 | <SamB> | well, I mean, same sequence of steps, all after the resource has been fetched |
| 20:26 | <SamB> | but it seems like this WOULD happen for an interrupted transfer too |
| 20:28 | SamB | wonders if it he has what it takes to make the full cross-referencing functionality work in multipage ... he'd presumably need to learn [more?] JavaScript ... |
| 20:56 | <ondras> | SamB: well the interrupted transfer is not really my case. I am troubled by an jpg image that has (a lot of) exif metadata located after image data. |
| 21:25 | <SamB> | ondras: maybe you should use other test images? |
| 21:26 | <SamB> | ondras: but really I don't think the wording in the standard allows the user-agent to take the shortcut you're worried about |
| 21:27 | <SamB> | (also I kind of doubt they're doing much to optimize specifically for huge EXIF) |
| 23:12 | SamB | wonders what rel="bookmark" means |
| 23:15 | <SamB> | (see <http://gpl-violations.org>;) |
| 23:46 | <Gege> | hi |
