Perl 6 language and compiler development | Logs at colabti.org/irclogger/irclogger_log/perl6-dev | For toolchain/installation stuff see #perl6-toolchain | For MoarVM see #moarvm
Set by Zoffix on 27 July 2018.
02:31 leont left 04:12 Kaypie left 06:23 Kaypie joined 07:08 gfldex left 07:09 gfldex joined
nine lizmat: migrating tickets sounds like a lot of work with little reward. It's also pretty much the last thing the Parrot project really did... 07:18
07:18 AlexDaniel joined 07:51 robertle_ joined
Geth rakudo: JJ++ created pull request #2654:
Change to .gitignore		 08:22
08:42 Kaypie left 08:56 patrickb joined
patrickb Hi everyone! 09:00
AlexDaniel, MasterDuke: In your GSoCproposals you are referring to a list of perl6 mentors. What does that refer to? (I would actually prefer a hard list of people, so students directly know who to contact.) 09:01
AlexDaniel yes yes I agree 09:02
I don't think there's a list atm
patrickb A distinct list for each idea I mean. Otherwise it's very difficult to tell who feels responsible for which idea. 09:04
AlexDaniel In fact I'd prefer these references to be replaced with specific people
yeah
patrickb ATM The ideas repo feels more like a place people dump their ideas on and just hope for someone else to mentor. But we'll have to remove all the ideas without a mentor before the deadline. 09:06
AlexDaniel yeah 09:12
hmm… I can mentor some, but only for the first half of the summer… 09:41
patrickb I'm not entirely sure sharing mentorship is ok for Google (but I might have actually read something like that, I can look that up), but seems like a good idea to me. 09:53
Geth rakudo: 2cd96705c1 | (Juan Julián Merelo Guervós)++ (committed using GitHub Web editor) | 2 files
Change to .gitignore

And added two plausible file extensions, and a note to avoid deletion.		 10:04
rakudo: 88b3517342 | (Jonathan Worthington)++ (committed using GitHub Web editor) | 2 files
Merge pull request #2654 from JJ/patch-4

Change to .gitignore
10:18 leont joined
lizmat Files=1266, Tests=88056, 379 wallclock secs (20.36 usr 5.79 sys + 2712.91 cusr 230.67 csys = 2969.73 CPU) 10:34
10:38 Kaypie joined 11:25 ufobat joined 11:27 |Tux| left 11:33 |Tux| joined 11:44 nebg joined
nebg hello all guys, how can i contirbute to the perl6 language ? 11:44
AlexDaniel nebg: hello! Anything specific that you're interested in? 11:49
lizmat nebg o/ 12:03
nebg AlexDaniel, no just programming 12:11
i mean i would like to work the programming part and not on the translation/docs part
lizmat nebg: github.com/rakudo/rakudo/issues contains the currently open issues of Rakudo Perl 6 12:17
and Pull Requests to fix bugs are always welcome :-)
AlexDaniel let us know if you'd prefer to work on MoarVM, maybe jnthn or brrt can introduce you to that if you have that specific interest 12:18
maybe something from that list is a good start? github.com/rakudo/rakudo/issues?q=...t+issue%22 12:19
also: rakudo.party/post/Newcomer-Guide-t...ore-Perl-6 12:20
Geth rakudo: 7133b81e4f | (Elizabeth Mattijsen)++ | 2 files
Make Buf[Int].new(1,2,3) error less cryptic

Basically indicate that anything but a native int type is not allowed
  (and will most likely not ever be implemented, so remove the NYI also).
Fixes R#2655
synopsebot R#2655 [open]: github.com/rakudo/rakudo/issues/2655 Cryptic error when parametrizing Bufs with non-native types
12:34 Kaypie left 13:18 brrt joined 13:50 lucasb joined
jnthn AlexDaniel: Can I take your <3 on my last post on github.com/rakudo/rakudo/issues/2608 as approval of that branch/erratum? 13:54
Also, if anyone else has views on that, please comment :) 13:55
AlexDaniel jnthn: uhh, it's hard. Only Blin knows for sure. Like there could be tens of modules depending on the original behavior in some unknown ways 13:58
whateverable does not follow branches, unfortunately
which needs to be fixed…
or at least Blin should know how to build on demand… 13:59
jnthn AlexDaniel: I find it unlikely that merging the branch would implicate many more, if any more, modules than the current state does; it largely just makes the behavior change more sensible.
AlexDaniel jnthn: let's try it 14:00
jnthn And gets rid of some previous, odd, semantics
AlexDaniel I can run Blin right after the merge and we'll be able to confirm that
jnthn OK. Well, first I need to make it behave sensibly for 6.c's sake
14:03 brrt left
jnthn I also need to look at the Cro::WebSocket one. Also... 14:03
It turns out that lowering away lexicals into locals and not retaining any hints about what we did makes the remote debug stuff in MoarVM way less useful (in fact, it was already a lot less useful than it could be for NQP for that reason). 14:04
I'm wondering about fixing that quickly, especially since this release is the one that'll be R*'d, meaning it'll be the Windows MSI, and I don't really want to doom Windows users to poorer debug experience for 3 months :)
14:05 nebg left
jnthn Since it's just a change to pass a little extra info down to the VM, it feels very low risk. 14:05
timotimo it'll still be difficult since locals have lifetimes and such
and registers change what they are "for" during spesh
jnthn timotimo: The first is only true if they are declared in a QAST::Stmt, which isn't the case for the ones lexicals are lowered in to 14:06
timotimo oh? OK, that makes it better
jnthn And yeah, you're right about spesh, but if you add an unconditional deopt usage for such locals when running in debug mode, well, you lose a few set eliminate opt possibilities, but retain them 14:07
Or maybe it's enough to make a breakpoint we marked as a deopt point, actually
*be marked
That's no extra lines of code, just a change in oplist :)
timotimo hm, that's fair. that'll change inline behaviour, but ehhh 14:08
less bad in debug than in profile
jnthn Yeah, though we should probably make a list of ops that are not counted for the sake of inline size
Probably as a new annotation in oplist
And then subtract them when considering inlines 14:09
timotimo aye, i've got a PR that'll do the calculations
jnthn Ah, neat :)
timotimo but it depends on the opcode number being past the first prof_ one
jnthn With an oplist annotation?
Hm, I think it'd be more robust with the annotaiton :)
timotimo nine already asked for that specific change :D
jnthn hah :)
Geth rakudo/regex-literal-experiment: 8ef7c15ed5 | (Jonathan Worthington)++ | 3 files
Make regex literals capture $/ and $!

These are then used when the regex is evaluated in boolean context, rather than using dynamic scope, which is not reliable given that `$_` is no longer dynamic as of 6.d. Previously, we would not lower away `$_` if we saw a regex literally present in that scope, however this cannot handle the at-a-distance uses. ... (20 more lines)		 14:11
jnthn Just a rebase, so I won't get spurious test fails 14:12
14:23 Guest92871 joined
timotimo jnthn: anything beside prof_, ctw_*, coverage_log, and breakpoint requires :ignoresize? 14:23
Guest92871 jnthn: I have tried to bisect R#2644 for you. (dogbert17 in disguise) 14:24
synopsebot R#2644 [open]: github.com/rakudo/rakudo/issues/2644 [regression][⚠ blocker ⚠] Some regression with Cro::WebSocket
jnthn Guest92871: Yeah, I saw that...doesn't give me much more clues, alas, but still helpful to know :) 14:25
timotimo: I don't think so
timotimo good 14:26
14:29 brrt joined 14:30 Kaypie joined
timotimo bleh, first a bunch of conflicts, now it won't compile any more 14:36
jnthn Hurrah, think I have the 6.c semantic retention for that regex-literal-experiment branch working 14:37
timotimo how did MVM_SWITCHENDIAN get lost? 14:41
Geth rakudo/regex-literal-experiment: 3d581c8d23 | (Jonathan Worthington)++ | 3 files
Preserve 6.c Regex.Bool behavior
 14:42
lizmat notable6: weekly reset 14:43
notable6 lizmat, Moved existing notes to “weekly_2019-01-28T14:43:52Z”
Geth roast: b8a8f3fe81 | (Jonathan Worthington)++ | S05-metasyntax/interpolating-closure.t
Update test for refined Regex.Bool semantics
 14:45
AlexDaniel lizmat: maybe a mention of GSoC repo? 14:46
lizmat yeah, in there already, in fact main article :-)
Geth roast/6.d-errata: 01e8830458 | (Jonathan Worthington)++ | S05-metasyntax/interpolating-closure.t
Update test for refined Regex.Bool semantics
AlexDaniel alright :)
Geth rakudo: 8ef7c15ed5 | (Jonathan Worthington)++ | 3 files
Make regex literals capture $/ and $!

These are then used when the regex is evaluated in boolean context, rather than using dynamic scope, which is not reliable given that `$_` is no longer dynamic as of 6.d. Previously, we would not lower away `$_` if we saw a regex literally present in that scope, however this cannot handle the at-a-distance uses. ... (20 more lines)		 14:47
rakudo: 3d581c8d23 | (Jonathan Worthington)++ | 3 files
Preserve 6.c Regex.Bool behavior
14:52 MasterDuke left
patrickb lizmat: Do you still have ties to p5 people? I have the feeling we need some more advertisement of GSoC among the p5ers. I already tried to reach two high profile p5ers but I no replies so far and I fear we are running out of time... 14:54
leont is a p5 person, but p5 is organized different than p6 14:57
lizmat and another Perl 6 Weekly hits the Net: p6weekly.wordpress.com/2019/01/28/...r-of-code/ 14:58
patrickb leont: I'd suspect the core devs and large module authors are candidates to be gsoc mentors.
lizmat patrickb: in the past year I've learned the hard way that I'm not the person to ask the Perl 5 community anything :-( 14:59
15:10 Guest92871 left
Geth nqp: 342e7934d6 | (Jonathan Worthington)++ | src/QAST/Block.nqp
Add a local debug map to QAST::Block

This will be used in order to retain the names of symbols that once existed as lexicals, but have been lowered to locals, so that those debugging programs using, for example, the MoarVM remote debugger, will be able to find the variables.		 15:14
|Tux| Rakudo version 2018.12-266-g3d581c8d2 - MoarVM version 2018.12-37-g631cfdf6c
csv-ip5xs0.763 - 0.782
csv-ip5xs-206.536 - 6.573
csv-parser22.311 - 23.375
csv-test-xs-200.438 - 0.445
test7.299 - 7.811
test-t1.893 - 1.901
test-t --race0.838 - 0.853
test-t-2031.556 - 33.839
test-t-20 --race9.836 - 10.107
 15:37
when I ran it locally while working on that box this morning, it was more like 1.91 -2.30 :( 15:38
16:11 robertle_ left
AlexDaniel patrickb: colabti.org/irclogger/irclogger_lo...01-28#l643 16:28
(question about gsoc)
16:39 ExtraCrispy joined 16:43 brrt left 16:45 ExtraCrispy left 16:56 patrickb left 17:04 patrickb joined 17:56 sivoais left 18:17 brrt joined 18:23 lucasb left 18:29 brrt left 20:27 gfldex left, dct joined 20:28 gfldex joined 22:30 MasterDuke joined, MasterDuke left, MasterDuke joined 23:28 dct left 23:30 dct joined 23:40 patrickb left

Powered by Cro and the Raku® Programming Language.

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

Thank you!