Perl 6 language and compiler development | Logs at colabti.org/irclogger/irclogger_logs/perl6-dev | For toolchain/installation stuff see #perl6-toolchain | For MoarVM see #moarvm
Set by AlexDaniel on 12 June 2018.
MasterDuke nqp-j: nqp::say(nqp::sec_n(3.4)) 00:19
00:55 BenGoldberg joined
Geth rakudo: 513577c3dc | (Tom Browder)++ (committed using GitHub Web editor) | src/Perl6/Pod.nqp
add TODO for deprecation warning for version 6.d
 01:13
samcv do we want a warning if someone tries to insert into a hash while it's still in an iterator? 02:01
AlexDaniel .tell lizmat how can I get nth element of a Seq without caching it? 02:04
yoleaux AlexDaniel: I'll pass your message to lizmat.
samcv AlexDaniel: any thoughts on this? perl 5 does this 02:05
02:05 Zoffix joined
Zoffix AlexDaniel: there's .skip method 02:05
AlexDaniel samcv: how does that look like in code? 02:06
samcv AlexDaniel: uh well i just implemented it in moarvm on a local branch but. do you mean in moarvm? 02:07
AlexDaniel Zoffix: ah right
Zoffix: thanks
samcv also for now i just added a print to stderr, but idk if there's any way for moarvm to give a warning to perl6 that isn't an exception
AlexDaniel .tell lizmat nevermind, forgot about .skip
yoleaux AlexDaniel: I'll pass your message to lizmat.
samcv but it works gist.github.com/1dff1d2acb7e32ba7a...a0231ac01f pretty nicely 02:08
but it piggy backs off the new insertion randomization code which determines if a key is inserted at the first place in the linked list or the second 02:09
BUCKET -> [a] # before insertion 02:10
well: BUCKET -> [b] -> [a] # then we try and insert [c] and get BUCKET -> [c] -> [b] -> [a] so it's predictable 02:11
AlexDaniel samcv: I was thinking about user code that would trigger the warning, or am I misunderstanding something? 02:12
samcv with the change it'll randomly choose to do what we do now, or insert like BUCKET -> [b] -> [c] -> [a], so they get randomized within a bucket. the goal is to randomize order of iteration
see the gist
AlexDaniel ah, looking at the gist now
samcv m: my @list; sub add-to (@list, $item) { if 2.rand.Int && @list { splice @list, 1, 0, $item } else { unshift @list, $item }}; for ^10 {add-to(@list, $_)}; say @list
camelia [7 9 8 5 6 0 4 3 2 1]
AlexDaniel ohhh now I understand
samcv yeah
Geth nqp/truffle: afa9494b2a | Unknown++ (committed by Daniel Green) | 14 files
[truffle] Implement a bunch of _n trigonometric ops

nqp::<asec asin acos atan atan2 cos cosh sin sinh sec sech tan tanh>		 02:14
nqp/truffle: 079c024884 | Unknown++ (committed by Daniel Green) | 2 files
[truffle] Implement nqp::isnanorinf
nqp/truffle: e7f48ecb20 | (Daniel Green)++ | nqp-truffle.nqp
[truffle] Return AST to coerce VOID to CALL_ARG
samcv now it'll *usually* give you all already existing keys, and *some* of the ones you've added
02:19 Zoffix left
MasterDuke .tell pmurias with my latest commits t/nqp/083-math.t should pass, but right now it dies with `java.lang.RuntimeException: Can't find lexical: &ok` because of the call to ok() inside is_approx() 02:22
yoleaux MasterDuke: I'll pass your message to pmurias.
AlexDaniel samcv: personally I wouldn't mind a proper exception there
02:23 lizmat left
samcv when we grab HASH_ITER_FIRST_ITEMit we set ht->bucket_rand_last = ht->bucket_rand; 02:24
every time a key is added it does ht->bucket_rand++, and so when we do HASH_ITER_NEXT_ITEM it compares the two and we can nicely figure out
AlexDaniel shouldn't we try to do that on rakudo level? 02:25
samcv no i don't think we should. it's a vm operation. and the vm handles the hash impl 02:26
also how would that work if you're adding items from another thread? 02:27
would it still detect it properly? it seems like a pain. i mean it's really the vm that needs to worry about it
AlexDaniel samcv: maybe file a ticket and we'll all discuss? 02:28
a warning that comes from VM is perhaps LTA because how do you silence that? 02:29
samcv well that is the question
AlexDaniel I know that we like warnings that cannot be turned off… looking at you, “potential difficulties” :) 02:30
samcv hah
but my point is that it's a deficiency with the hash implementation, and there could be others that are designed to insert while iterating 02:31
but that's just my opinion.
and in the vm it'll basically be free
AlexDaniel I'd say open a ticket and we'll figure it out :) 02:34
samcv and i'm not sure how we'd do it in perl6. we do nqp::shift_s on an iterator to get the next key 02:35
but i guess we can figure it out. i think it's a good feature
AlexDaniel: i still think there should be some way for the VM to send a warning to rakudo, though i'm not sure how that'd look 03:21
03:36 BenGoldberg left 04:43 Ven` joined 04:48 Ven` left 04:51 skids left
samcv looks like there's some roast tests that assume every hash will have the same key order. so gonna correct these 05:04
05:37 ufobat joined 06:36 robertle joined
[TuxCM] Rakudo version 2018.06-62-g513577c3d - MoarVM version 2018.06-98-g7547dd800
csv-ip5xs0.918 - 0.918
csv-ip5xs-207.600 - 7.686
csv-parser24.414 - 25.812
csv-test-xs-200.457 - 0.460
test8.884 - 8.887
test-t2.288 - 2.361
test-t --race0.978 - 1.049
test-t-2040.423 - 42.571
test-t-20 --race14.321 - 14.744
 06:39
07:47 pmurias joined
pmurias MasterDuke: the &ok not being seen inside subs is caused by lack of closures 07:48
yoleaux 02:22Z <MasterDuke> pmurias: with my latest commits t/nqp/083-math.t should pass, but right now it dies with `java.lang.RuntimeException: Can't find lexical: &ok` because of the call to ok() inside is_approx()
08:02 pmurias left 08:13 dogbert17 joined 08:14 robertle left, robertle joined, pmurias joined, [Coke] joined 08:15 [Coke] left, [Coke] joined 08:32 Ven` joined 08:35 lizmat joined 09:03 pmurias left, pmurias joined 09:14 lizmat left 10:23 MasterDuke left 10:51 Ven` left 11:08 gfldex left 11:25 brrt joined
tbrowder_ g’day #perl6-dev! 11:51
Zoffix: i’m working on deprecation of defining a pod table caption in the %config hash. i see no reason to not start warning of such deprecation now instead of waiting till 6.d is released. what say you? 11:53
the use will generate an exception in version 6.e. 11:54
.tell Zoffix see my blurb at 0600 ish (CDT, Z-5) 12:12
yoleaux tbrowder_: I'll pass your message to Zoffix.
12:12 brrt left
Geth star: 16eb5edd99 | (Steve Mynott)++ | tools/build/Makefile.in
single quote PATH in order to fix #114
 12:28
12:37 brrt joined 12:40 Ven` joined 12:50 j3nnn1 joined 12:51 Ven` left 12:53 Ven` joined 13:45 brrt left 13:51 scovit joined 14:07 Zoffix joined 14:09 brrt joined
Zoffix .tell tbrowder_ I was wanting to barn on it now implies no one uses it. Is there any reason for the rush? 14:09
yoleaux Zoffix: I'll pass your message to tbrowder_.
brrt \o 14:12
yoleaux 14:08Z <Zoffix> brrt: yeah, that's fine. I figured out Good Enough instructions for WSL for now and no longer need the execstack option
brrt Zoffix++
14:30 Zoffix left 14:32 skids joined
Geth rakudo/rescalar: b5a5d5bc76 | (Jonathan Worthington)++ | 2 files
Add a spesh plugin for return value decont

Which both reduces the generated bytecode size for this operation, as well as allowing us to provide optimized variants for a number of the different cases that show up.		 14:40
15:12 gfldex joined
tbrowder_ Zoffix: i think (i’ll check again) that Pod::To::* and docs htmlify may use the old-style. 15:22
yoleaux 14:09Z <Zoffix> tbrowder_: I was wanting to barn on it now implies no one uses it. Is there any reason for the rush?
Geth rakudo/rescalar: adb00ba7d3 | (Jonathan Worthington)++ | src/vm/moar/spesh-plugins.nqp
Only look up Iterable once
 15:33
tbrowder_ yes, Pod::To::Text expects the old-style caption only. i haven’t looked at any actual pod tables other than my own. 15:37
15:47 pmurias left 16:18 robertle left 16:44 MasterDuke joined
MasterDuke . 16:44
.tell pmurias i suspected so. at this point i don't know enough to contribute toward implementing closures, so good luck! 16:45
yoleaux MasterDuke: I'll pass your message to pmurias.
16:46 lizmat joined
lizmat . 17:10
yoleaux 09:17Z <sarna> lizmat: oh, it's even your post. I meant a link to IRC logs, here: colabti.org/irclogger/irclogger_log...07-10#l134
09:19Z <sarna> lizmat: and why didn't I privmsg them - I didn't want to bother them if they were done with contributing. I shouldn't really bring this up in the first place, my bad
lizmat .tell sarna that's all water under the bridge 17:20
yoleaux lizmat: I'll pass your message to sarna.
17:21 brrt left
tbrowder_ Zoffix: i’ll finish my PR and ask for review. PR will fix Pod::To::Text, too. 17:39
18:31 brrt joined, brrt left 18:51 Ven` left 18:58 kanopis joined 19:16 kanopis left 19:29 lizmat left 19:39 lizmat joined 19:46 lizmat left 20:54 AlexDaniel joined 21:10 skids left 21:11 DrForr joined 21:16 Kaiepi joined
AlexDaniel squashable6: next 21:16
squashable6 AlexDaniel, ⚠🍕 Next SQUASHathon in 2 days and ≈12 hours (2018-07-07 UTC-12⌁UTC+14). See github.com/rakudo/rakudo/wiki/Mont...Squash-Day
Geth rakudo/ugexe-patch-1: 57f8c61b40 | (Nick Logan)++ (committed using GitHub Web editor) | src/Perl6/Grammar.nqp
Fix potential race condition in EVAL
 21:31
rakudo: ugexe++ created pull request #2015:
Fix potential race condition in EVAL
rakudo/ugexe-patch-1: 0bbd7bc69a | (Nick Logan)++ (committed using GitHub Web editor) | src/Perl6/Grammar.nqp
Use NQPLock
 22:01
22:45 cog joined
Geth rakudo: 80bbfbb36a | (Nick Logan)++ (committed using GitHub Web editor) | src/Perl6/Grammar.nqp
Fix potential race condition in EVAL (#2015)

Fix potential race condition in EVAL		 22:46
23:11 cog left 23:12 cog joined
Geth nqp: cec76ff79e | (Daniel Green)++ | tools/build/MOAR_REVISION
Bump MoarVM

af455397f Refactor inline logging; detect missed inlines 7ac061092 Correct a mis-placed `break` statement 9745f9216 Add facts to attributes in spesh plugin guards bd8361e2f [JIT] Fix ARG6 for windows better 2592774c5 Make minor alterations to some exprjit templates ... (64 more lines)		 23:19
¦ nqp: version bump brought these changes: github.com/MoarVM/MoarVM/compare/2...gaf455397f
23:19 Kaiepi left 23:21 Kaiepi joined
Geth rakudo: b659a9c2f8 | (Daniel Green)++ | tools/build/NQP_REVISION
Bump NQP

cec76ff79 Bump MoarVM dbee32dc1 Just typo fixes in comments 2d5d8868c Merge pull request #478 from perl6/W4anD0eR96-patch-2 28908b2d5 Update 114-pod-panic.t 0d797a89f Merge pull request #480 from perl6/W4anD0eR96-patch-1 ... (7 more lines)		 23:21
¦ rakudo: version bump brought these changes: github.com/perl6/nqp/compare/2018....gcec76ff79
japhb Wheeee, lots of commits in those bumps 23:58

Powered by Cro and the Raku® Programming Language.

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

Thank you!