| 06:08 | <hsivonen> | Oh, maybe "[e]" means editorial, dunno. |
| 06:08 | <hsivonen> | Does anyone happen to remember whether the ambiguous ampersand proposal proceeded in W3C Bugzilla, on public-html or on the whatwg mailing list? |
| 06:28 | <DerekNonGeneric> | Does anyone happen to remember whether the ambiguous ampersand proposal proceeded in W3C Bugzilla, on public-html or on the whatwg mailing list? |
| 06:32 | <DerekNonGeneric> | mailing lists are tough since that was way before my time |
| 06:35 | <DerekNonGeneric> | maybe an email is the best way to make things happen in here, so far no complaints |
| 06:57 | <zcorpan> | hsivonen: I found https://lists.w3.org/Archives/Public/public-whatwg-archive/2008Mar/0023.html |
| 07:02 | <annevk> | hsivonen: I found https://lists.w3.org/Archives/Public/public-whatwg-archive/2007Jun/thread.html#msg217 and https://github.com/whatwg/html/commit/351f1aefeb0082cf88f5c0c94983d10a552b0d35 (also https://github.com/whatwg/html/commit/5cef9e088cc04f24debe53104fd4fb2b7731b4bc) |
| 07:03 | <annevk> | hsivonen: those are the first commits made to the entity part of the parser section since the initial checkin of Web Apps 1.0 (and since they were in June 2007 it was rather easy to find corresponding emails); and just before the introduction of the term of ambiguous ampersands, which seems to be only there for non-normative purposes? |
| 07:04 | <hsivonen> | Did ambiguous ampersand gain normative conformance significance only later? |
| 07:05 | <annevk> | hsivonen: I think the introduction of a specific parser state for it is a much more recent (supposedly editorial) refactoring of the entity part of the parser section |
| 07:05 | <annevk> | If that's what you mean |
| 07:06 | <hsivonen> | It looks like I might have said something that could be read to be in favor of the ambiguous ampersand error tweak. 😬 |
| 07:06 | <hsivonen> | If that's what you mean |
| 07:07 | <annevk> | hsivonen: you mean https://lists.w3.org/Archives/Public/public-whatwg-archive/2007Jun/0281.html? :-) |
| 07:07 | <annevk> | hsivonen: oh, I think that was all done in that time, not later (probably as part of those changes) |
| 07:07 | <annevk> | That is, in June 2007 |
| 07:09 | <hsivonen> | hsivonen: you mean https://lists.w3.org/Archives/Public/public-whatwg-archive/2007Jun/0281.html? :-) |
| 07:11 | <hsivonen> | I believe the conformance change was around 2011. I'd like see what example cases were given so that I could land a test case with those. |
| 07:16 | <annevk> | hsivonen: oh okay, I guess I did go through git blame by first looking for the introduction of ambiguous ampersand and then seeing what was changed to the entity part of the parser section... So if there was some kind of conformance change later I would have missed it |
| 07:21 | <annevk> | hsivonen: https://github.com/whatwg/html/commit/259c0608d4bd89528eb6a823bb468836daa50176 maybe? https://github.com/whatwg/html/commit/1eb194f229a6e481f313320a396b9da99b9f0706 is the most recent normative change to entity parsing afaict (2013). |
| 07:22 | <annevk> | Methodology: git blame, search for "Consume the maximum number of characters possible", and then look at changes while skipping over editorial work |
| 07:25 | <hsivonen> | annevk: Thanks! I think https://github.com/whatwg/html/commit/259c0608d4bd89528eb6a823bb468836daa50176 is it. I'm surprised that the corresponding bug was filed by Maciej. I expected to find a W3C Bugzilla bug with zcorpan arguing in favor and me arguing against. |
| 08:00 | <hsivonen> | sideshowbarker: Experimentally, I'm now rather convinced that your ambiguous ampersand patch works if the review comments are addressed. I still don't understand why it works, and I think it would be good if I understood why. (Particularly, how the unusual reading of the next character in the AMBIGUOUS_AMPERSAND state and the transition to DATA works if the input buffer ends right there vs. doesn't end right there. I guess I'll step through it in Pernosco, to see how it works.) |
| 09:39 | <Ms2ger 💉💉> | Does JS define the order of iteration for the declaratively defined properties on prototypes? |
| 10:38 | <hsivonen> | sideshowbarker: Experimentally, I'm now rather convinced that your ambiguous ampersand patch works if the review comments are addressed. I still don't understand why it works, and I think it would be good if I understood why. (Particularly, how the unusual reading of the next character in the |
| 12:13 | <Sam Sneddon [:gsnedders]> | Does JS define the order of iteration for the declaratively defined properties on prototypes? |
| 12:14 | <Ms2ger 💉💉> | Object.getOwnPropertyNames(Array.prototype), say |
| 12:16 | <hsivonen> | sideshowbarker: So, I'm looking at the AMBIGUOUS_AMPERSAND state and thinking "do we need this line"? And it's looking a lot like we could get rid of the whole state by inlining the error-if-semicolon check to all places that transition to AMBIGUOUS_AMPERSAND. What am I missing? |
| 12:16 | <hsivonen> | It quite possible that I'm missing something |
| 12:26 | <hsivonen> | Wow. After manually inlining stuff back and looking at the diff, it all makes sense! I'll clean this up a bit and will post an alternative PR. |
| 12:44 | <hsivonen> | Wow. After manually inlining stuff back and looking at the diff, it all makes sense! I'll clean this up a bit and will post an alternative PR. |
| 16:24 | <Domenic> | Does JS define the order of iteration for the declaratively defined properties on prototypes? |
| 16:24 | <Domenic> | Folks in #tc39-general:matrix.org might know more. |
| 21:00 | <zcorpan> | hsivonen: yeah I think I was always against allowing unescaped ampersands. I still think it's not great |
| 21:24 | <zcorpan> | sideshowbarker: oh no. https://www.w3.org/html/wg/spec/ exists and people link to it on twitter |
| 21:25 | <miketayl_r> | 2010 was a great vintage for HTML |
| 21:31 | <zcorpan> | those were good times certainly |
| 21:32 | <zcorpan> | remember Google Gears? |