| 13:22 | <GPHemsley> | TabAtkins, et al.: What's the story with pseudo-elements like ::after that require there to be content in the primary element? (E.g. why doesn't img::after work?) |
| 13:28 | <Ms2ger> | GPHemsley, img::after is a child of img, and replaced elements don't render their children |
| 13:29 | <GPHemsley> | Ms2ger: Why is ::after a child? |
| 13:29 | <Ms2ger> | Anyone have thoughts on the serialization of doctypes whose publicId/systemId contain quotes? |
| 13:29 | <Ms2ger> | GPHemsley, that's just how they work :) |
| 13:30 | <GPHemsley> | Ms2ger: Any chance that could change to something more intuitive? |
| 13:30 | <Ms2ger> | Not really |
| 13:31 | <Ms2ger> | Say you have "h1:before {content: counter(chapno, upper-roman) ". "}" |
| 13:32 | <Ms2ger> | Would you want that number to be on a separate line? |
| 13:46 | <MikeSmith> | Ms2ger: escape the quotes somehow? |
| 13:53 | <Ms2ger> | MikeSmith, how? :) |
| 13:56 | <MikeSmith> | Ms2ger: "somehow" in the annevk "User agents are encouraged to expose parse errors somehow." spirit |
| 13:56 | Ms2ger | makes is "Escape the quotes with extreme prejudice" |
| 14:00 | <Ms2ger> | MikeSmith, seriously though, afaict there's no way |
| 14:01 | <MikeSmith> | Ms2ger: I guess not if you want them to be round-trippable at least |
| 14:02 | <Ms2ger> | That's kind of the idea :) |
| 14:03 | <MikeSmith> | but if a valid doctype isn't allowed to have quotes in it, then I'd think nobody should expect a doctype with quotes in it will necessarily surive as-is after being (re)serialized |
| 14:04 | <Ms2ger> | They'd better nor :) |
| 14:04 | <Ms2ger> | not, even |
| 14:05 | <MikeSmith> | so you could have the serialization algorithm just, say, just skip them them |
| 14:05 | <MikeSmith> | (instead of attempting to escape them or do something else with them) |
| 14:05 | <MikeSmith> | but yeah otherwise I guess there's no way |
| 14:06 | <MikeSmith> | and by "skip them" I mean never append them to the output stream |
| 14:09 | <Ms2ger> | Yeah |
| 14:11 | Ms2ger | needs to get a recent IE somehow |
| 14:18 | <odinho> | Ms2ger: Not that hard? Just d/l an virtualbox image or something? |
| 14:18 | <odinho> | MS gives that out for free. At least did last time I needed it. |
| 14:18 | <Ms2ger> | Do you know where those are? |
| 14:18 | <Ms2ger> | I got a IE10-prerelease version that way |
| 14:18 | <odinho> | http://www.modern.ie/en-us/virtualization-tools < guess this one |
| 14:19 | <odinho> | Yeah, virtualbox for linux, and you get lots of ie's on lots of os'es. |
| 14:19 | <Ms2ger> | Oh, that's actually a MS site? |
| 14:19 | <odinho> | Mm |
| 14:19 | <Ms2ger> | Whois says yes |
| 14:24 | <Ms2ger> | Ta |
| 19:40 | <gsnedders> | hsivonen: Do you have any data for how often the moz char detector gets the encoding? |
| 19:42 | <malinus> | Hello. How exactly am I supposed to use the "find" method? I want to search for specific tags, and have my parsed html document. |
| 20:10 | <malinus> | When I'm trying to follow the example and set the encoding. I get this nice little exception thrown at me: "TypeError: Cannot explicitly set an encoding with a unicode string" from: html5lib.parse(source.text, treebuilder="lxml", encoding=source.encoding) |
| 21:16 | <SimonSapin> | Hi malinus. What is 'source' in your code? |
| 21:17 | <malinus> | nvm. I've figured it out |
| 23:06 | <SimonSapin> | Do I need to do something like create an account to follow what’s going on in Critic? https://critic.hoppipolla.co.uk/r/445 |