japhb ditto, but we will get there again 00:28
timotimo i'm not making much progress with that :( 00:29
00:38 lizmat joined 01:15 TimToady joined 01:48 ilbot3 joined 02:10 FROGGS joined 03:57 vendethiel joined 07:25 vendethiel joined
dalek arVM/ignoremark: 3ea7470 | FROGGS++ | / (10 files):
implement ordbaseat and eqatim_s for ignoremark
 08:11
arVM/ignoremark: e40b21c | FROGGS++ | src/core/interp.c:
chars are not equal beyond end of string
 08:19
08:20 FROGGS joined
FROGGS I'll gist my nqp changes in a bit, to prevent a stage0 update in case I did silly things 08:21
08:28 tgt joined 08:31 tgt joined
FROGGS jnthn: gist.github.com/FROGGS/1a0cb3448b7f014511d8 - it passes all tests from S05-modifier/ignoremark.t, though I probably need to implement ignorecase+ignoremark 08:33
tadzik japhb: what's the problem? 09:11
09:12 lizmat joined 11:00 lizmat joined
jnthn FROGGS: Seems reasonable. 11:07
11:36 FROGGS joined
FROGGS jnthn: feel free to tweak the MVM_string_ord_basechar_at function :o) 11:38
jnthn *nod* 11:39
FROGGS so, my plan is to make ignorecase+ignoremark work, then pushing a stage0 11:40
ohh, and I have to think about jvm and parrot
nwc10 in the same way that we're thinking about NFG for parrot? 11:41
and, for that matter, NFG for JVM?
(different from parrot NFG, to be fractionally less obtuse)
FROGGS and I had spotted an :mm modifier, and also have seen samemark tests (perhaps :mm is samemark)... ahh, and there we also sameaccent tests, I wonder if that is the same
hmpf, ranges in character classes also need to be handled to work under ignoremark 11:50
jnthn urgh 11:51
FROGGS which is a problem... 11:53
or... hmmm
dunno if I can translate that to bound checks
12:15 FROGGS joined
FROGGS I think I am going for an QAST::Regex subtype called 'ignorecase+ignoremark', and I really really hope that we do not have any other unimplemented modifiers of this kind 12:17
ahh, samecase and samemark seems to be for substitutions only... 12:20
so that "aBcDe".subst(/<[b..d]>/, 'xxx', :ii) might end up as "eXxXe" 12:22
TimToady: is that correct? 12:23
and "aöúe".subst(/\w+/, 'a', :mm) results in "aäáa" 12:25
(the ii example lacks a quantifier but besides that...)
jnthn: +1 to QAST::Regex subtype 'ignorecase+ignoremark ? 12:29
jnthn FROGGS: I...hmm 12:32
I don't have a beter idea.
FROGGS beter... :P 12:33
that reminds me of weter.com
the funny thing is that this page is unmodified since at least 10 years 12:34
14:32 lizmat joined
japhb tadzik: The one where panda didn't know to treat .so/.dll specially, and people had to install them "by hand" for e.g. linenoise. I just hadn't noticed whether that was already done, or still NYI. 14:51
tadzik let me see 14:53
I wasn't aware ofit
yep, it's broken
and now my perl6 is broken
fuck
but it does Copying blib/lib/liblinenoise.so to /home/tadzik/.rakudobrew/moar-nom/install/share/perl6/site/lib/liblinenoise.so
so is that place wrong?
japhb tadzik: I wonder if there is something that needs to be done so the dlopen (or Windows equivalent) can find it. The collective memory of #perl6 might know it. 14:55
(Or if what really needs to be fixed is NativeCall ...)
tadzik right now I don't know how to make it work on linux either :)
japhb Come to think of it (from the outside), I would think it's the local install and NativeCall library lookup code paths that probably need thought. Whether panda has to be directly involved in that, I dunno. 14:59
lizmat fwiw, we had a temporary solution for that in CURLI 15:00
but the migration of @*INC to strings broke that
japhb (It occurs to me that everyone had been referring to it as a "panda v. DLL" problem, but it's not actually a foregone conclusion that it's panda's problem beyond just getting the DLL to the proper install directory for that particular auth/version of the module.)
lizmat I hope to have time again tomorrow to look at this 15:01
japhb lizmat: Ah, gotcha.
tadzik is there a reson why we shouldn't put site/lib in nativecall's search path?
and/or: why isn't it there already :)
oh, versioning..
japhb Yeah 15:02
tadzik so CUR needs to deal with .so (does it already?) and NativeCall has to be aware of it
lizmat CURLI does, but CURLI doesn't work with panda
:-(
at least, it didn't for me... :-( 15:03
tadzik is it panda's fault? 15:04
lizmat well, fault is a big word 15:08
it is at least incompatible atm
tadzik hm 15:10
15:13 FROGGS joined 15:41 TimToady joined
TimToady FROGGS: re irclog.perlgeek.de/moarvm/2015-05-09#i_10575280 I think that would merely result in 'a'. You'd have to replace with aaaa to get the result you postulated 15:45
or just use \w and :g 15:46
FROGGS TimToady: I see
19:12 zakharyas joined 19:13 brrt joined
brrt \o 19:14
timotimo o/ 19:15
brrt just beat mario kart :-) 19:25
19:43 vendethiel joined 20:11 lizmat joined
jnthn tries to beat tutorial slides :) 20:18
lizmat waved at jnthn, but he and pmichaud didn't look :-(
:-) 20:19
jnthn lizmat: Didn't see you at all :( Such is my eyesight 20:23
lizmat you were in a deep discussion, so even with good eyesight, you would not have seen us :-) 20:24
jnthn ah :)
lizmat we were inside, you were outside passing
jnthn Ah, in l'irish pub? 20:25
lizmat yup 20:27
jnthn I'd totally join if I didn't have to finish this tutorial :) 20:28
lizmat we were just in back from "Taste Of China"
jnthn Also I should rest my voice given I have to talk for 4 hours tomorrow. :)
lizmat I completely understand... I was just teasing :-)
jnthn Figured :-) 20:29
Well, tomorrow the talks are finished and I can relax with a beer :)
lizmat :-) 20:30
20:56 zakharyas joined 23:27 TimToady joined

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!