| 00:03 | <TabAtkins> | Ah, but bless you, details. Trivial to hack support for, and just the right element for my needs. |
| 00:12 | <Philip`> | TabAtkins: Aha, it works if I do a transparent-PNG-fix style thing |
| 00:12 | <Philip`> | without needing an actual PNG |
| 00:12 | <Philip`> | just using filter:progid:DXImageTransform.Microsoft.Alpha(opacity=80); |
| 00:12 | <Philip`> | (at least in IE8, I assume it'll work in IE6/7 too) |
| 00:19 | <Hixie> | hm |
| 00:19 | <Hixie> | i don't say how to handle <link type=""> if the type can't be parsed |
| 00:19 | <Hixie> | wonder what the spec should say for that case |
| 00:23 | <Philip`> | Fatal error, YSOD |
| 00:25 | <Hixie> | mmhm |
| 00:26 | <Hixie> | webkit seems to ignore type="" altogether |
| 00:54 | <TabAtkins> | Philip`: Oh, duh. I knew MS had a plain opacity filter. Silly me. |
| 01:23 | <TabAtkins> | roc: Specced repeating gradients. Let me know if something doesn't match your implementation. |
| 01:23 | <roc> | ta |
| 01:38 | <roc> | TabAtkins: looks great |
| 01:41 | <Hixie> | watcha guys working on that has all these gradient thingies? |
| 01:42 | <roc> | CSS gradients spec |
| 01:42 | <roc> | sorry, off topic |
| 01:42 | <Hixie> | oh, cool |
| 01:42 | <Hixie> | not at all off topic, don't worry |
| 01:42 | <Hixie> | this channel has no topic |
| 01:42 | <Hixie> | the only thing that's off-topic is logic |
| 01:50 | <Hixie> | anyone remember what IE's equivalent of event.target is? |
| 01:53 | <miketaylr> | event.srcElement? |
| 01:54 | <Hixie> | thanks |
| 01:54 | <miketaylr> | np |
| 02:52 | <TabAtkins> | roc: Oh man, gradient support looks beautiful. Thanks so much. |
| 02:52 | <roc> | pleasure |
| 02:52 | <roc> | thanks for writing a decent spec |
| 02:53 | <roc> | a good spec, atually |
| 02:54 | <TabAtkins> | Woo! |
| 02:54 | <TabAtkins> | /spec high-five |
| 04:53 | <othermaciej> | we should update WebKit's gradient implementation |
| 05:00 | <Hixie> | what does an HTTP client do when an HTTP _server_ (not proxy) replies with 407? |
| 05:01 | <weinig> | othermaciej: how different is the new spec? |
| 05:02 | <othermaciej> | weinig: different (arguably better) syntax from what we had I think |
| 05:02 | weinig | should take a look |
| 05:02 | <othermaciej> | Hixie: I would assume either "get really confused" or "report an error" |
| 05:03 | <othermaciej> | oddly, it's not something we've tested |
| 05:04 | <othermaciej> | weinig: if I'm well enough to come in tomorrow I may borrow you part of the day to work on some standards stuff - we need to give feedback on File API and maybe you can help me, and maybe I can take a whack at helping you update Web IDL to ES5 |
| 05:04 | <weinig> | othermaciej: I would very much like that |
| 05:04 | <othermaciej> | that being said, my throat still feels like it's full of razor blades duct taped to angry badgers, so I have my doubts about being in tomorrow |
| 05:05 | <weinig> | othermaciej: :( |
| 05:06 | <othermaciej> | weinig: it sound like a lot of our team is out of action with something possibly similar, so I'm hesitant to go into the office til I'm confident that I am not contagious |
| 05:08 | <weinig> | othermaciej: yeah, it does seem like there is a hit out on the safari team, flu wise |
| 05:16 | <Hixie> | 0 e-mails |
| 05:16 | <Hixie> | 1 XXX marker (which I am not fixing this year, it's "define WebSQL") |
| 05:16 | <Hixie> | 22 bugs! |
| 05:16 | <Hixie> | where did they come from! |
| 05:16 | <Hixie> | we were at zero! |
| 05:16 | <Hixie> | man, this never ends |
| 05:51 | <JoePeck> | Where would I go to find out about all the html node list / collection types? NodeList, HTMLCollection, HTMLOptionsCollection. Are there others? |
| 05:55 | <MikeSmith> | fyi: error messages for obsolete messages now patched in v.nu source to include suggestions about what to use instead |
| 05:55 | <MikeSmith> | e.g., http://qa-dev.w3.org:8888/?doc=http://dev.w3.org/html5/tests/validation/full/invalid/obsolete/acronym.html |
| 05:55 | <MikeSmith> | "The acronym element is obsolete. Use the abbr element instead." |
| 05:57 | <MikeSmith> | still need to add reporting for obsolete attributes |
| 05:59 | <Hixie> | JoePeck: HTML5 defines some of them, the rest are probably in DOM Core |
| 05:59 | <JoePeck> | Hixie: thanks, bdash pointed me to http://www.whatwg.org/specs/web-apps/current-work/#collections-0 |
| 06:09 | Hixie | tries to remember why we need a secondary insertion mode at all |
| 06:12 | <Hixie> | oh it's so that <table><tr><td><svg></table> closes the table |
| 08:36 | <Hixie> | well |
| 08:36 | <Hixie> | technically i'm at 0,1,0 |
| 08:36 | <Hixie> | i guess it's time to find out what i'm supposed to do to go to last call with all these specs |
| 08:38 | <Philip`> | Quick, somebody file more bugs! |
| 08:39 | <Philip`> | "If you are satisfied with this response, please change the state of this bug to CLOSED." - does that need to be done by the reporter even for ones reported anonymously through the spec comment box? |
| 08:40 | <Hixie> | othermaciej said no, i think |
| 08:41 | <Hixie> | not sure though |
| 08:41 | <othermaciej> | no one *has* to do it |
| 08:41 | <othermaciej> | but if it's not done, then eventually the bug times out into NoReply |
| 08:41 | <Philip`> | Is it something that preferably should be done? |
| 08:42 | <othermaciej> | it would be nice to have a record of agreement for some comments, at least once HTML WG Last Call begins |
| 08:42 | <othermaciej> | and also, I like it when the bug system accurately reflects reality |
| 08:42 | <othermaciej> | so I personally prefer it |
| 08:42 | <othermaciej> | but no one will yell at you if you don't do it |
| 08:44 | <Hixie> | Philip`: if you agree with all of them, just search for your IP and mass-close them :-) |
| 08:44 | <Philip`> | Reflecting reality sounds like a worthwhile goal |
| 08:44 | <Hixie> | oh, hey, it turns out the whatwg charter doesn't have a "last call" stage |
| 08:44 | <Hixie> | we just keep doing call for comments until we're done |
| 08:44 | <Hixie> | that's pretty sensible |
| 08:44 | <Philip`> | Hixie: Mass-closing doesn't help much when I have to individually read every one to work out if I agree or not :-) |
| 08:45 | <Hixie> | Philip`: well you can read them in your e-mail client |
| 08:45 | <Hixie> | Philip`: but if you go to the bug, then sure |
| 08:45 | <Philip`> | Oh |
| 08:45 | <Philip`> | Problem: I can't close the bugs |
| 08:45 | <Philip`> | since I'm not the owner or anything |
| 08:47 | <Hixie> | zcorpan has rights to give you rights to give yourself the rights to close the bug, iirc |
| 08:47 | <Hixie> | (i do not, again iirc) |
| 08:52 | <Philip`> | Seems like a flaw in the LC process if reporters can't close their bugs without asking zcorpan |
| 09:02 | <Philip`> | MikeSmith: Can somebody like you give me bug-editing privileges? |
| 09:06 | <othermaciej> | Philip`: MikeSmith almost certainly can give you the privileges and probably will be willing to if you are a heavy bug filer |
| 09:09 | <MikeSmith> | Philip`: I will give you editbugs now |
| 09:09 | <Philip`> | I have something like a dozen that I mostly want to close now, and will probably have more in the future since I only read up to about p30 out of 654 :-) |
| 09:11 | <Philip`> | MikeSmith: Thanks! |
| 09:11 | <Philip`> | MikeSmith: (Do you need to know my email address or anything?) |
| 09:12 | <MikeSmith> | Philip`: excors one, right? |
| 09:12 | <Philip`> | MikeSmith: Yep |
| 09:14 | <Philip`> | MikeSmith: Seems to work fine |
| 09:15 | <Philip`> | Don't know if I really need the "member of the XSL or XQuery WG" permission bit but I won't complain |
| 09:16 | <MikeSmith> | Philip`: yeah, it does that for everybody .. I have no idea why |
| 09:16 | <Philip`> | Sounds like they're even more open than the HTML WG |
| 09:16 | <Philip`> | since anybody can become a member, whether they want to be or not |
| 09:16 | gsnedders | stretches and goes back to html5lib work |
| 09:36 | <Philip`> | "omitting tags doesn't lead to people dying" - but what if you're using HTML to control a nuclear power station? |
| 09:36 | <Hixie> | aw man, are you really gonna make me rewrite that algorithm using position/input |
| 09:37 | <MikeSmith> | or to control an electric chair |
| 09:38 | <MikeSmith> | (another use case to consider) |
| 09:39 | <Philip`> | Hixie: Feel free to rewrite it however you want, as long as it's not more ambiguous than the rest of the spec :-) |
| 09:39 | <Hixie> | i was at zero bugs for 30 minutes |
| 09:40 | <Hixie> | 30 minutes! |
| 09:40 | <Philip`> | MikeSmith: Then errors in the markup might accidentally lead to not killing somebody, which seems like exactly the opposite problem |
| 09:40 | <Hixie> | admittedly i was rather lost for those 30 minutes, not sure what to do with myself |
| 09:40 | <Hixie> | i mean, i've had things on this pile for 6 years now |
| 09:40 | <Hixie> | and suddenly there was no pile |
| 09:40 | <Hixie> | so it's reassuring to know the pile grows back so fast |
| 09:40 | <Hixie> | but still |
| 10:11 | <gsnedders> | <!DOCTYPE html PUBLICa> throws two parse errors in the DOCTYPE |
| 10:11 | <gsnedders> | Both at the "a" |
| 10:11 | <gsnedders> | Firstly, in "anything else" in After DOCTYPE public keyord state |
| 10:11 | <gsnedders> | Then a second parse error in the Before DOCTYPE public identifer state when it is reconsumed there |
| 10:12 | <Philip`> | Does that matter? |
| 10:12 | <gsnedders> | Normally we've tried to avoid a single error giving multiple parse errors, no? |
| 10:12 | <Hixie> | file a bug |
| 10:13 | <Hixie> | telling me how to fix it |
| 10:13 | <gsnedders> | Hixie: You never used to require that! |
| 10:13 | <Hixie> | i'm lazier now :-P |
| 10:13 | gsnedders | is too |
| 10:13 | Hixie | just changed the status of all the sections that were in pre-LC to LC |
| 10:14 | <gsnedders> | So WHATWG is in LC now? |
| 10:14 | <Hixie> | not yet |
| 10:15 | <Hixie> | actually technically as i discovered earlier, whatwg doesn't have an lc state :-) |
| 10:15 | <gsnedders> | s/an/a/ |
| 10:15 | <Hixie> | "an ell sea" |
| 10:15 | <Philip`> | "a lick" |
| 10:16 | <gsnedders> | "a last call" |
| 10:16 | <Hixie> | should i add section markers to all <h3>s and <h4>s that don't have one? |
| 10:16 | <Hixie> | i mean, status markers |
| 10:16 | gsnedders | grumbles at being unable to change the email address in the W3C bug tracker |
| 10:24 | <gsnedders> | Hixie: The space characters cases within the before DOCTYPE public/system identifier states can never be reached. To fix that, remove those cases. |
| 10:25 | <Hixie> | type that into the box and hit "send review comments" :-) |
| 10:25 | <gsnedders> | 8018 |
| 10:25 | <Hixie> | ta |
| 10:29 | <erlehmann> | Hixie, what is the next state called then ? waiting for implementations ? |
| 10:30 | <Hixie> | yeah |
| 10:31 | gsnedders | is trying to get html5lib shipped in around a week |
| 10:31 | <Hixie> | cool |
| 10:32 | <gsnedders> | (I got to the point where the impl passes all TCs yesterday) |
| 10:32 | <gsnedders> | It's just neither the TCs nor the impl are entirely up to date :) |
| 10:33 | <Hixie> | i know _that_ feeling |
| 10:34 | <Philip`> | gsnedders: At least the spec is up to date, so the situation is not all bad |
| 10:34 | <Philip`> | Well, up to date except for the bug you just filed |
| 10:34 | <Hixie> | he filed 2! :-) |
| 10:35 | <gsnedders> | Well, it seemed easier to fix all the failing tests first |
| 10:38 | <gsnedders> | I also got stuff properly tested in both namespaced and non-namespaced mode |
| 11:46 | <Philip`> | I'm glad we have a relatively regular calendar system now, and don't have to make do with datetime strings like "nine months after the death of Tiberius, on the eighteenth day before the Kalends of January, just as the sun rose" |
| 12:15 | <Dashiva> | Philip`: You still have to deal with it if you're a classical scholar |
| 12:15 | <Philip`> | Dashiva: Fortunately I'm not |
| 12:15 | <Dashiva> | Although wouldn't 18 days before kalends be referenced relative to ides instead? |
| 12:16 | <Philip`> | http://www.medievalgenealogy.org.uk/cal/jansas.htm says no |
| 12:16 | <Dashiva> | Or nones, for that matter. Just not kalends. |
| 12:17 | <Dashiva> | Oh, right, december has ides on the 13th |
| 12:17 | <Philip`> | I don't quite understand why the 3rd day before Kalends of January is the day before the day before Kalends of January |
| 12:19 | <Dashiva> | It's a translation issue |
| 12:19 | <Dashiva> | There was a special word for the day before |
| 12:19 | <Philip`> | Anyway, you probably shouldn't disagree with C. Suetonius Tranquillus about Roman calendars :-p |
| 12:20 | <gsnedders> | Hixie: Does http://www.w3.org/Bugs/Public/show_bug.cgi?id=8019 look all right? |
| 12:20 | <Hixie> | yeah |
| 12:20 | <Dashiva> | "Pridie" |
| 12:20 | <Philip`> | Dashiva: Even with a special word, why is it "3rd"? |
| 12:20 | <Dashiva> | I've forgotten so much |
| 12:21 | Philip` | didn't even know so much |
| 12:21 | <Dashiva> | Philip`: They count inclusively |
| 12:23 | <Dashiva> | You see the same with easter |
| 12:23 | <Dashiva> | Jesus is supposed to be dead three days, but he cheats and gets up on the second day because he's counting inclusively |
| 12:25 | <Philip`> | kalends, pridie kalendas, tertio kalendas (going backwards) it seems |
| 12:25 | <Philip`> | I suppose it makes sense if you don't translate as "day before" |
| 12:28 | <gsnedders> | Philip`: You have any intention to update your generated tests soon? |
| 12:28 | <Philip`> | gsnedders: The tokeniser ones? |
| 12:29 | <gsnedders> | Philip`: Yeah |
| 12:30 | <gsnedders> | Philip`: With my proposed edit to the spec in 8019 404 tests fail, almost all because of DOCTYPE changes at end of Sept |
| 12:30 | <Philip`> | gsnedders: No, unless I run out of other things to work on first, or someone gives me lots of money to do it :-p |
| 12:31 | <Philip`> | (I think it'd involve lots of rewriting the OCaml tokeniser to match years of spec changes) |
| 12:31 | <Philip`> | (which is a bit tedious) |
| 12:37 | <bogphanny> | Can someone please give me an example of how the <keygen> tag would typically be used in a real life example? |
| 12:55 | <gsnedders> | Philip`: OK, I've manually updated them now |
| 13:00 | <gsnedders> | Hixie: you still there? |
| 13:00 | <Hixie> | yup |
| 13:01 | <gsnedders> | RCDATA state when it comes to & goes to character reference in data state, and then after parsing entity it goes to data state. bug, right? |
| 13:01 | <Hixie> | yeah |
| 13:02 | gsnedders | files yet another bug |
| 13:03 | <Hixie> | good lord |
| 13:04 | <murr4y> | you miss those 30 minutes now, eh? :D |
| 13:05 | gsnedders | has only filed three bugs today :P |
| 13:06 | <Hixie> | http://www.whatwg.org/issues/data.html?period=0 is depressing |
| 13:06 | <Hixie> | (well ok not really) |
| 13:06 | <Hixie> | (it's pretty awesome really) |
| 13:06 | <Hixie> | but still! |
| 13:07 | gsnedders | just sees yellow block in all browsers |
| 13:07 | murr4y | too :p |
| 13:07 | <murr4y> | (all 2, ff3/opera10) |
| 13:09 | <Hixie> | y'all need better browsers |
| 13:09 | <gsnedders> | Estimated date for last e-mail based on the data above: 2009-10-23 |
| 13:12 | <gsnedders> | Hixie: If you can fix 8029 soon… |
| 13:12 | <Hixie> | i expect to bring the bug list to 0 tomorrow again |
| 13:13 | <gsnedders> | yay! |
| 13:13 | gsnedders | generally just wants to get html5lib out ASAP |
| 13:13 | <gsnedders> | (Last release was in Jun 2008) |
| 13:14 | <Hixie> | i'm preparing the blog post for when we go to last call |
| 13:14 | <Hixie> | anything i should mention? |
| 13:15 | <gsnedders> | "We need a test suite, kthxbai" |
| 13:16 | <Hixie> | i don't plan to start looking at the test suite til 2011, according to the time table |
| 13:16 | <Philip`> | Got to leave plenty of time for untested buggy implementations to establish legacy for future generations |
| 13:17 | <Philip`> | Then we can write test suites and say "ha ha you got it all wrong" and it'll be too late to change anything |
| 13:17 | <Hixie> | well there are plenty of bits of testsuite all over the place |
| 13:17 | <Philip`> | Sounds like a good plan to me |
| 13:18 | <Hixie> | what we really need is someone who can do this as a fulltime job |
| 13:20 | <annevk42> | chaals talks about standards: "WHAT-WG: Hacking with friends" |
| 13:20 | <Philip`> | Hixie: What we really need is employers who let people do it as a fulltime job |
| 13:21 | <Hixie> | ok i think i'm ready to go to cfc/lc as soon as the bug list is back to zero again |
| 13:22 | <Hixie> | i just need to change the web workers, vocab, html5, and complete.html drafts to class=cfc instead of class=draft and post the blog post |
| 13:24 | <annevk42> | a stable copy would be nice |
| 13:24 | <annevk42> | like we did with Web Forms back in the days when it was in Last Call |
| 13:24 | <Hixie> | yeah |
| 13:24 | <Hixie> | the scripts won't work in that though |
| 13:24 | <Hixie> | at least the status and bug reporting ones won't |
| 13:24 | <annevk42> | makes sense |
| 13:24 | Philip` | will have to keep filing new bugs |
| 13:25 | <Philip`> | Someone needs to set up an IRC bot that alerts everyone when there's only one bug left |
| 13:25 | <Hixie> | Philip`: your bugs are starting to get so editorial that they're almost trivial to deal with, you need to find more serious problems if you want to slow me down. :-) |
| 13:25 | <Philip`> | so we can all rush and add more |
| 13:25 | gsnedders | has found three technical issues today, hence he > Philip` |
| 13:26 | <Hixie> | oh gsnedders did the change to the 'reset' algorithm fix the foreign content stuff? |
| 13:26 | <gsnedders> | Hixie: What where? |
| 13:27 | <Hixie> | the "reset the insertion mode" algorithm changed earlier today |
| 13:27 | <gsnedders> | Oh, I've not changed html5lib for that yet |
| 13:27 | Philip` | probably shouldn't have started reading the spec from the beginning, if he wanted to find serious issues |
| 13:28 | <gsnedders> | I just got the tokenizer up to the end of September (excluding things that have changed this month too) |
| 13:28 | <Hixie> | ah ok |
| 13:28 | <annevk42> | so Philip` finally got bored with finding bugs in XML tools and is now taking on the full HTML5 spec? :) |
| 13:28 | <Hixie> | i'd love to know if the reset change i did earlier today actually fixes the problem |
| 13:28 | <Hixie> | i wasn't 100% convinced it would |
| 13:29 | <Hixie> | btw to fix the entity stuff i'm just gonna have to create a new state for entities in rcdata, right? |
| 13:29 | <gsnedders> | Hixie: right |
| 13:29 | <Hixie> | k that'll be easy then |
| 13:30 | <gsnedders> | I didn't say I was finding hard technical issues :) |
| 13:32 | <Philip`> | annevk42: I wouldn't say "full", because I'm sure I'll stop before I get to the end of it :-p |
| 13:35 | <Hixie> | right. I'm gonna go to bed so that the 13 bugs can become 50 bugs and give me some stuff to do tomorrow |
| 16:07 | <gsnedders> | "script data escaped dash dash state"… oh dear… |
| 19:46 | <TabAtkins> | Quick question: In C, do variable names and function names live in separate namespaces? (Unrelated to current namespace-related discussion. ^_^) |
| 19:47 | <boogyman> | hola |
| 19:47 | <Philip`> | TabAtkins: No |
| 19:48 | <Philip`> | TabAtkins: or yes |
| 19:48 | <TabAtkins> | Hrm. So I can't have a function named "add" and a variable named "add"? |
| 19:48 | <Philip`> | You can have a global function called "add" and a local variable called "add" (which will shadow the function) |
| 19:48 | <TabAtkins> | Damns, I thought they were separate. Kk. |
| 19:49 | <TabAtkins> | How does that not cause problems? ::boggles:: |
| 19:49 | <Philip`> | The syntax lets you write stuff like &add which will apply to both functions and variables |
| 19:49 | <boogyman> | can someone please explain the functionality http://dev.w3.org/html5/spec/Overview.html#ping |
| 19:50 | <Philip`> | TabAtkins: Why would it cause problems? |
| 19:50 | <TabAtkins> | boogyman: If a user clicks on a link with @ping, the UA will do a normal navigation to whatever the @href is, but it will *also* fire off GET requests to all the urls in @ping. |
| 19:50 | <Philip`> | If you have a function and say "int add = 2;" then 'add' (in that scope) will refer to that variable, regardless of what's defined in global scope, which seems quite sensible |
| 19:51 | <TabAtkins> | Philip`: Well, you basically have a big set of global variables you have to manage, and be careful not to shadow. |
| 19:51 | <TabAtkins> | Because then you can't go "add(add,3)" |
| 19:51 | <Philip`> | POST, not GET |
| 19:51 | <Philip`> | (for ping) |
| 19:51 | <Philip`> | TabAtkins: You can just rename the local variable in that case |
| 19:52 | <Philip`> | (C++ lets you write ::add to refer to the global scope, but that's ugly) |
| 19:53 | <TabAtkins> | Philip`: Maybe the C-likes don't have the same problems with names that Lisps do. We name a lot of functions with words that also are useful as argument names. |
| 19:53 | <TabAtkins> | Like the word "list", which shadows the "list" function in Scheme. (In CL it doens't shadow, because they live in different namespaces.) |
| 19:53 | <boogyman> | TabAtkins: so @ping provides a method to multi-submit the corresponding node? |
| 19:54 | <TabAtkins> | boogyman: If "node" means what I assume it does, then yes? |
| 19:54 | <TabAtkins> | It's meant to allow you to, frex, track how often a particular link is clicked. |
| 19:54 | <Philip`> | TabAtkins: C-like languages don't have the problem of dynamic scoping, which makes it more easily manageable |
| 19:55 | <TabAtkins> | Lexical, you mean. C-likes *do* use dynamic scoping. ^_^ |
| 19:55 | <Philip`> | TabAtkins: I think you've got it precisely the wrong way round |
| 19:55 | <boogyman> | so not necessarily multi-submit the node for processing, but submit the node to @href for processing and @ping for "tracking" purposes? |
| 19:56 | <TabAtkins> | boogyman: Yes. |
| 19:56 | <boogyman> | ok, thx |
| 19:56 | <Philip`> | There's nothing dynamic about scoping in C, it's just based on the nesting in the syntax, hence being lexical scoping |
| 19:57 | <TabAtkins> | Philip`: No, you've got it wrong way round. Lexical scoping is what allows closures - you resolve free variables based on what's present at the time the function is defined. Dynamic means you resolve based on what's present at call-time. |
| 19:58 | <TabAtkins> | http://en.wikipedia.org/wiki/Scope_(programming) |
| 19:58 | <TabAtkins> | Rather, not at the time the function is defined, but simply at what's present in the surrounding text where the free variable occurs. |
| 19:59 | <Philip`> | TabAtkins: No, *you've* got it the wrong way round :-p |
| 19:59 | <Philip`> | http://foldoc.org/lexical+scope vs http://foldoc.org/dynamic+scope etc - C does lexical (static) scoping |
| 20:00 | <TabAtkins> | In C: |
| 20:00 | <TabAtkins> | int x = 0; |
| 20:00 | <TabAtkins> | int f() { return x; } |
| 20:00 | <TabAtkins> | int g() { int x = 1; return f(); } |
| 20:00 | <TabAtkins> | What does g() return? |
| 20:00 | <Philip`> | |
| 20:00 | <Philip`> | because it's not dynamic scoping |
| 20:01 | <Philip`> | and the references to x are resolved statically at compile-time |
| 20:01 | <Philip`> | based on lexical information, not dynamic run-time information |
| 20:01 | <TabAtkins> | I could have *sworn* C-likes used dynamic scoping. Guess I'm misremembering my high-school days. |
| 20:02 | <TabAtkins> | In any case, CL and Scheme are both lexical-scope languages, except for their special variables. |
| 20:02 | <Philip`> | It would be really weird if C did dictionary lookups on every variable reference |
| 20:03 | <TabAtkins> | It doesn't need dictionary lookups to implement dynamic scope. You just keep a stack around for each binding, pushing and popping as you move between scopes. |
| 20:06 | <TabAtkins> | Man, what is *up* with that foldoc article. I *assure* you that "(setf x 0) (defun f () x) (defun g () (let ((x 1)) (f))) (g)" returns 0. Emacs-lisp is a dynamic-scope language, and a number of early lisps were as well, but *all* modern lisps are lexical. |
| 20:07 | <Philip`> | What's "modern"? |
| 20:07 | <Philip`> | That page was last updated 13 years ago |
| 20:08 | <TabAtkins> | Common Lisp and anything created after it. |
| 20:08 | <TabAtkins> | Wow, no wonder. CL was a new guy at that point. |
| 20:08 | <TabAtkins> | Jeezus. |
| 20:09 | <TabAtkins> | So in conclusion: I had it the right way round, I was just misremembering my C. |
| 20:13 | <TabAtkins> | Unrelated: Just got my first "Good Answer" badge on SO with a question about <video>. ^_^ |
| 21:20 | and` | finds it ironic that his last e-mail on character encoding problems is affected by that very issue. |
| 21:20 | <Dashiva> | That's not ironic |
| 21:21 | <Dashiva> | Ironic would be if your email was about having fixed the character encoding problems :) |
| 21:23 | <and`> | Right. I actually tried to fix a problem with missing curly quotes by adding ASCII quotes and ended up with both. |
| 22:06 | <Hixie> | Philip`: the "sets of..." are all strings |
| 22:31 | <Philip`> | Hixie: Indeed, but they're called sets and it says they're sets |
| 22:32 | <Philip`> | when they're actually just strings that represent sets |
| 22:51 | <inimino> | Philip`: as opposed to being just bytes in some datatype that represent a set? ;-) |
| 22:54 | <Philip`> | inimino: As opposed to being a mathematical construct that is a set, and that can be manipulated in a normal set-based way and can be parsed from strings and serialised to strings |
| 22:59 | Philip` | happily manages to get his C++ code call JS code that calls C++ code, but unhappily realises he implemented infinite recursion |
| 23:05 | <Hixie> | where does it say they are sets? |
| 23:05 | <Hixie> | oh, i see |
| 23:07 | <Philip`> | I don't think it's worth changing the spec, since it's easy enough to work out they're strings |
| 23:07 | <Hixie> | i changed the "is a set" bit |
| 23:07 | <Hixie> | i didn't change the names of the things |
| 23:07 | <Hixie> | the names are opaque! |
| 23:07 | <Hixie> | or so i shall claim |
| 23:08 | <Hixie> | ok, e-mail is back to zero |
| 23:08 | <Hixie> | 17 bugs! |
| 23:08 | <Hixie> | man |
| 23:09 | Philip` | quite likes RDF's clear separation of lexical spaces and value spaces |
| 23:09 | <Philip`> | Hixie: It's not the 50 that you asked for :-( |
| 23:09 | <Hixie> | heh |
| 23:10 | <Hixie> | (agreed about lexical vs value being a useful seggregation -- i'll know better for next time, i guess) |
| 23:11 | <Hixie> | ok shower, meeting, bbl. |
| 23:17 | <othermaciej> | I have some bugs I need to file |
| 23:17 | <othermaciej> | maybe now is a good time |