Camelia, the Perl 6 bug

IRC log for #november-wiki, 2009-04-22

| Channels | #november-wiki index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:53 sri_kraih joined #november-wiki
01:01 sri_kraih__ joined #november-wiki
01:42 sri_kraih joined #november-wiki
03:26 ihrd joined #november-wiki
03:26 ihrd left #november-wiki
09:03 ihrd joined #november-wiki
09:03 ihrd left #november-wiki
11:35 ruoso joined #november-wiki
14:41 masak joined #november-wiki
14:51 masak I almost have a working prototype of the Genshi clone.
14:52 masak but I'm a bit busy right now. if someone who likes Perl 6 grammars wants to go the final few yards, I could make a gist paste and link to it here.
14:53 masak right now, I've settled on the name "Hitomi", meaning "doubly beautiful".
14:54 moritz_ feel free, but I won't get to it in the next 2 or 3 hours (though maybe later)
14:55 masak great. it'll be interesting to discuss, at least. I'm mainly being hampered by insufficient introspection into Match objects. (and I got too tired to code last night.)
14:55 * masak goes to make the gist
14:57 moritz_ masak: I fixed that last known issues with Match.perl today
14:57 masak sweet.
14:57 * masak svn-ups Rakudo
14:57 masak moritz_++
14:57 masak karma moritz_
14:57 zarah moritz_ has a karma of 10
14:58 moritz_ svn up?
14:58 moritz_ where are you living? 2007?
14:58 masak er.
14:59 masak my thought patterns seem to have survived the move to git...
14:59 masak I abstract away all the piping with huffmanized aliases :)
15:17 masak http://gist.github.com/99851 # simple hitomi prototype
15:18 moritz_ do closing tags allow attributes?
15:19 moritz_ and it seems you don't make sure that opening and closing tags match
15:19 moritz_ ie you'd parse something like <a></b>, no?
15:20 masak oops. removing the attrs from the closing tags.
15:20 masak yes, I don't check that opening and closing match. that's for later.
15:20 moritz_ you could, you know?
15:20 masak I just want to prototype something right now.
15:20 moritz_ with backreferences
15:20 masak moritz_: patches welcome. :)
15:21 masak but I'd rather make the filtering work right now.
15:21 moritz_ ok
15:21 moritz_ as I said, I'll look into it a bit later
15:21 masak excellent.
15:22 masak the objective is to filter the pe:if things and only include the element in question iff the code evals to true.
15:22 masak s/the code/the code in the attribute value/
15:28 moritz_ masak: what's your motivation not to use action methods?
15:29 masak moritz_: also the prototyping.
15:29 moritz_ ah
15:29 masak I saw in the beginning that action methods would be good here.
15:29 masak but I'm not familiar enough with them yet.
15:29 masak I even tend to forget the syntax for tying them in.
15:30 masak in the long run, it would surely be a Good Thing, especially if the separation meant that the grammar could be kept pure XML while the action methods did the Hitomi part.
15:59 Tene masak: I got rakudo using other HLLs properly, except for some case issues mentioned on the list.
15:59 masak Tene: cool!
17:04 masak ok, will be heading home soon.
17:04 masak I might be hacking more on the Hitomi prototype, but if others want to have a go at it, feel free.
17:04 masak I expect to be able to blog my week 5 of Web.pm tomorrow.
17:38 Tene masak: posted hitomi anywhere?
17:38 Tene besides the paste?
17:38 masak Tene: no. maybe I should push it to WEb.pm/drafts
17:38 Tene Yeah
17:38 masak s/E/e/
17:38 * masak does so
17:38 Tene thanks. :)
17:39 Tene I'll be working on moving rakudo to its own HLL tonight
17:39 Tene well, getting it as far as I can before bugs, which allison will then look at
17:40 masak sounds great.
17:40 masak I love the progress we're seeing right now.
17:40 Tene hitomi is a PCT language?
17:40 Tene right?
17:40 masak ideally, yes.
17:40 Tene right now?
17:40 masak right now it's a prototyped Perl 6 script.
17:40 Tene ah
17:40 masak I've pushed now.
17:59 szabgab joined #november-wiki
18:43 moritz_ @tell masak re xml parsing gist, I've put a 'say $chunk.perl' before the offending line - $chunk has only one named capture, that is named 'content' - no <opening>, so that's where your failure comes from
18:43 zarah Consider it noted.
18:44 moritz_ @tell masak that is you have a tag that's followed by text, not other tags
18:45 zarah Consider it noted.
19:27 moritz_ @tell masak I tried to re-do the stuff with action methods, but hit a segfault pretty soon :/ - so recursion seems to be the way to go, still
19:27 zarah Consider it noted.
20:48 masak joined #november-wiki
20:48 masak @messages
20:48 zarah masak: 2 h 4 m 21 s ago, moritz_ said re xml parsing gist, I've put a 'say $chunk.perl' before the offending line - $chunk has only one named capture, that is named 'content' - no <opening>, so that's where your failure comes from
20:48 zarah masak: 2 h 3 m 17 s ago, moritz_ said that is you have a tag that's followed by text, not other tags
20:48 zarah masak: 1 h 20 m 57 s ago, moritz_ said I tried to re-do the stuff with action methods, but hit a segfault pretty soon :/ - so recursion seems to be the way to go, still
20:48 masak moritz_: nice to know that our efforts keep digging up things to do. :) moritz_++
20:48 masak moritz_: also, thanks for the investigation.
20:49 moritz_ masak: you're welcome
20:49 masak I tried earlier to devise a test that doesn't trigger an 'undefined value' warning, and that does exactly what I want.
20:49 masak haven't succeeded yet.
20:49 masak will give it another go now with this new knowledge.
20:52 masak grr, .WHAT on Matches still give the name of the grammars that matched them.
20:52 masak I thought pmichaud fixed that.
20:53 moritz_ anyway, use the power of .perl
20:53 * masak just did
20:54 moritz_ I spent 1 1/2 hours fixing that yesterday night
20:54 masak mberends and I will soon unleash a yet-to-be-written little web app called "context", which will help in diagnosing grammars, showing how and what they match.
20:54 masak moritz_++
20:55 masak will there be a way to bunch $h<several><hash><indexing><steps><together> ?
20:56 PerlJam masak: [{}] $h, <several hash indexing steps together>   #  ;-)
20:56 masak PerlJam: wow.
20:57 PerlJam I doubt that works though.
20:57 masak quick, let's hack it into the spec!
20:57 PerlJam I remember dicussing this with TimToady at some point in the past though
20:58 PerlJam Having such a mutable spec is quite nice.  We can grow or shrink the language as we see fit   :)
20:58 masak is there a precedent to the specification strategy of Perl 6?
21:02 PerlJam What's hitomi all about?
21:02 masak PerlJam: it's a clone of Python's Genshi.
21:02 PerlJam ah
21:02 PerlJam not yet functional?
21:02 masak PerlJam: I want a clone of a popular templating system in Web.pm, and Genshi seems very well likes.
21:03 masak PerlJam: not at all functional. just a broken prototype.
21:03 masak I like the name, though :)
21:14 masak ok, dropping below the radar again. g'night.

| Channels | #november-wiki index | Today | | Search | Google Search | Plain-Text | summary