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.
01:21 Kaypie joined, Kaiepi left 02:28 BenGoldberg joined 02:37 ufobat_ joined 02:40 ufobat___ left 02:42 lizmat left 03:05 BenGoldberg left 03:07 BenGoldberg joined
Geth rakudo: MasterDuke17++ created pull request #2054:
Make infix:<!=> for Ints consistent with...		 03:19
04:21 Ven` joined 04:26 Ven` left 04:42 brrt joined 05:24 ufobat_ left 05:49 robertle left
[Tux] Interesting. I::P5 now works with XS, but not with PP on my CSV tests 06:03
gist.github.com/Tux/e0d94d164dc192...c06f021fcc
gist.github.com/Tux/0418582c161750...57158164b9 06:08
Rakudo version 2018.06-151-g5682ee1f1 - MoarVM version 2018.06-146-g1b075ec1a
csv-ip5xs0.982 - 0.993
csv-ip5xs-208.747 - 9.301
csv-parser25.181 - 26.883
csv-test-xs-200.445 - 0.453
test9.221 - 9.267
test-t2.397 - 2.427
test-t --race1.012 - 1.034
test-t-2043.463 - 43.998
test-t-20 --race14.686 - 18.231
 06:09
06:18 BenGoldberg left 06:52 Ven` joined 06:59 [TuxCM] left 07:31 brrt left 08:04 robertle joined 08:17 lizmat joined
lizmat Files=1242, Tests=76401, 328 wallclock secs (15.74 usr 5.60 sys + 2279.96 cusr 231.75 csys = 2533.05 CPU) 08:18
08:24 [TuxCM] joined
Geth rakudo: 46e7640fef | (Daniel Green)++ | src/core/Int.pm6
Make infix:<!=> for Ints consistent with...

the other infix operators.		 08:39
rakudo: eb6656b876 | (Elizabeth Mattijsen)++ (committed using GitHub Web editor) | src/core/Int.pm6
Merge pull request #2054 from MasterDuke17/make_not_equal_for_Ints_consistent_with_other_operators

Make infix:<!=> for Ints consistent with...
08:54 brrt joined 09:16 pmurias joined
pmurias MasterDuke: yes, nqp-truffle doesn't have any loops yet 09:16
MasterDuke: I have been focusing on getting the bytecode/serialized-ast runner to work on SubstrateVM recently (which should offer a super fast startup) 09:19
cognominal m say("wrong?") if "a a" ~ / b :s || a a / 09:28
m, say("wrong?") if "a a" ~ / b :s || a a / 09:29
m say("test") 09:30
jkramer m: say "test"
camelia test
cognominal I forgot
m: say("wrong?") if "a a" ~ / b :s || a a /
camelia Regex object coerced to string (please use .gist or .perl to do that)
wrong?
in block <unit> at <tmp> line 1
cognominal oops
jkramer You need ~~ 09:31
cognominal m: say("wrong?") if "a a" ~~ / b :s || a a /
camelia wrong?
cognominal m: say("wrong?") if "a a" ~~ / [ b :s ] || a a /
camelia Potential difficulties:
Space is not significant here; please use quotes or :s (:sigspace) modifier (or, to suppress this warning, omit the space, or otherwise change the spacing)
at <tmp>:1
------> 3("wrong?") if "a a" ~~ / [ b :ā€¦
cognominal I expected the :s not to apply right of the || 09:32
The work around is easy though 09:33
jkramer > Regex adverbs can come before or inside a regex declaration and only affect the part of the regex that comes afterwards, lexically. 09:41
Sounds to me like it would apply to everything after :s 09:42
jnthn They're bounded by [...] and (...)
But not other non-grouping constructs
Geth rakudo: fa073679bc | (Jonathan Worthington)++ | 5 files
Provide p6box_* as a desugar using box_*

Rather than the extops that we have today. For MoarVM we totally get rid of them. For JVM, the binder being written in Java means we don't do go quite so far as we still need the code.
This will simplify implementing box/unbox elimination optimization in MoarVM spesh. Further, it aids our medium-term goal of eliminating the use of MoarVM extops.		 09:43
09:44 ExtraCrispy joined
cognominal thx 09:45
Geth rakudo: e2bbe3f0a9 | (Elizabeth Mattijsen)++ | 2 files
Make trying to assign to a Map give a less LTA error

There was however one catch: PseudoStash is a Map, but assumes the values in it are assignable. Fixed by giving PseudoStash its own ASSIGN-KEY method.
Given: ... (9 more lines)		 10:28
10:35 stmuk_ joined 10:36 stmuk left
Geth nqp/truffle: 4 commits pushed by (Paweł Murias)++ 11:00
11:11 Ven` left 11:55 Ven` joined 11:57 pmurias left 12:03 pmurias joined 12:07 pmurias left 12:22 ExtraCrispy left 12:27 ExtraCrispy joined
lizmat so I'm thinking of making this die: 12:40
m: my %h is Map = a => 42; %h = b => 666; dd %h 12:41
camelia Map.new((:b(666)))
lizmat in other words, a Map can only be initialized once, and from there on out, cannot be initialized again
one way of doing this, would be to give Hash its own STORE method 12:42
which would remove the need for having a STORE_AT_KEY semi-official interface method
jnthn moritz timotimo Zoffix ^^^ suggestions / comments ?
jnthn huh, I thought Map already wasn't assignable? 12:44
jnthn is surprised
lizmat it's not assignable but it is initializable
yeah, just found out about it
m: my %h is Map = a => 42; %h<a> = 666
camelia Cannot change key 'a' in an immutable Map
in block <unit> at <tmp> line 1
lizmat m: my %h is Map = a => 42; %h<c> = 666
camelia Cannot add key 'c' to an immutable Map
in block <unit> at <tmp> line 1
lizmat that all stops you from changing, but:
m: my %h is Map = a => 42; %h = b => 666; dd %h # just works :-( 12:45
camelia Map.new((:b(666)))
lizmat jnthn: anyways, so you think the current behaviour is a bug
jnthn Yeah, htat looks bad 12:46
lizmat ok, will make an issue and start fixing it 12:47
Geth rakudo: lizmat self-assigned You should not be able to assign more than once to a Map github.com/rakudo/rakudo/issues/2055
35fec8e09a | (Elizabeth Mattijsen)++ | src/core/IO/Path.pm6
 12:51
roast: e689a085b8 | (Zoffix Znet)++ | S02-types/list.t
[v6.d REVIEW] Test actual value in regexp first/grep

Instead of its stringification
Orig: github.com/perl6/roast/commit/4317d5826		 13:27
nqp: fad8b7cdaf | (Jonathan Worthington)++ | src/vm/moar/QAST/QASTOperationsMAST.nqp
Emit invoke_v instruction for void context calls

This is less useless work dealing with return values, and also can lead to better results if such calls are inlined. If the called code returns a native, but we call it with invoke_o, then we force a box just to throw that away. Now we don't bother.		 13:28
timotimo hooray! 13:29
i didn't know we don't do that
jnthn Me either, was looking at where a pointless box was coming from and then noticed :) 13:30
13:44 brrt left 13:48 travis-ci joined
travis-ci NQP build passed. Jonathan Worthington 'Emit invoke_v instruction for void context calls 13:48
travis-ci.org/perl6/nqp/builds/403117472 github.com/perl6/nqp/compare/2e053...d8b7cdaf2e
13:48 travis-ci left 14:08 skids joined 14:39 Ven` left 14:40 Ven` joined 14:44 brrt joined
Geth rakudo: ee800adaf1 | (Elizabeth Mattijsen)++ | src/core/Hash.pm6
Give Hash its own .STORE

For now, just a copy of Map.STORE, in preparation for fixing R#2055		 14:47
synopsebot R#2055 [open]: github.com/rakudo/rakudo/issues/2055 You should not be able to assign more than once to a Map
15:04 fake_space_whale joined 15:07 robertle left 15:09 brrt left 15:20 ufobat_ joined
Geth rakudo: fffe89c5e9 | (Elizabeth Mattijsen)++ | src/core/Map.pm6
Make it impossible to assign to a Map

This fixes R#2055, but in the most nihilistic way.		 15:23
synopsebot R#2055 [open]: github.com/rakudo/rakudo/issues/2055 You should not be able to assign more than once to a Map
Geth roast: 05b2ca9594 | (Elizabeth Mattijsen)++ | S32-hash/map.t
Add test for R#2055
 15:26
synopsebot R#2055 [open]: github.com/rakudo/rakudo/issues/2055 You should not be able to assign more than once to a Map
15:33 ufobat_ left 15:39 [TuxCM] left
AlexDaniel jnthn: ping :) github.com/rakudo/rakudo/issues/2056 15:39
jnthn Hmm, that looks vaguely familiar, to the degree that I'm wondering if it's one of the things I already solved in MoarVM HEAD... :) 15:40
AlexDaniel alright, 14 more modules to go
oh
AlexDaniel upgrades
ehhhhā€¦ longest 2 minutes of the day 15:43
jnthn: no, not fixed 15:44
jnthn OK, thanks for checking
AlexDaniel actually, why didn't I just ask committable 15:45
c: d3c5381b74^,d3c5381b74,HEAD gist.github.com/AlexDaniel/f5eb6ef...415fc7a639
committable6 AlexDaniel, gist.github.com/c1fa0489e3439e3343...b92f512729
AlexDaniel ā†‘ proof :)
15:46 travis-ci joined
travis-ci Rakudo build errored. Elizabeth Mattijsen 'Give Hash its own .STORE 15:46
travis-ci.org/rakudo/rakudo/builds/403153375 github.com/rakudo/rakudo/compare/3...800adaf188
15:46 travis-ci left
buggable [travis build above] ā˜  Did not recognize some failures. Check results manually. 15:46
Geth rakudo: jnthn self-assigned ā€œMVMHash representation requires MVMString keysā€ rescalar regression github.com/rakudo/rakudo/issues/2056
8f090821a7 | (Elizabeth Mattijsen)++ | src/core/Map.pm6

We already get native string from the iterator, so use that directly.		 15:48
AlexDaniel Geth: ver github.com/rakudo/rakudo/commit/c2...356201cc21 15:56
Geth AlexDaniel, version bump brought in these changes: github.com/perl6/nqp/compare/2018....6-g39f1057
AlexDaniel Geth: ver github.com/perl6/nqp/commit/39f105...5b2ce19e8b
Geth AlexDaniel, version bump brought in these changes: github.com/MoarVM/MoarVM/compare/2...1-gcae4731
AlexDaniel everything in one bumpā€¦ not cool 15:57
hm cool, clearly another (or maybe the same) spesh bug 16:09
16:15 fake_space_whale left
AlexDaniel jnthn: R#2058 also 16:20
synopsebot R#2058 [open]: github.com/rakudo/rakudo/issues/2058 [regression][āš  blocker āš ] Rescalar regression found in P5quotemeta module (goes away with MVM_SPESH_DISABLE=1)
AlexDaniel .tell lizmat github.com/lizmat/P5substr/pull/1 16:25
yoleaux AlexDaniel: I'll pass your message to lizmat.
16:30 Ven` left
AlexDaniel cool, one module was fixed by the author before I managed to get to it 16:31
jnthn: fwiw there's another module failure that was bisected to the same commit: github.com/rakudo/rakudo/issues/20...-404575027 16:41
17:08 [TuxCM] joined 17:50 lizmat left 18:32 ExtraCrispy left 18:43 undersightable6 joined 19:20 dogbert17 joined 19:33 robertle joined
robertle AlexDaniel: regarding github.com/robertlemmen/perl6-cach...c/issues/1 19:35
you seem to be looking at this for some reason that I do not quite understand, but I do not think that there is any change in rakudo or moar that could have changed the behavior here significantly, I think the test has always been flappy and bad in general... 19:36
19:54 j3nnn1 joined 19:56 stmuk joined 19:59 stmuk_ left 20:05 brrt joined 20:08 lizmat joined 20:16 brrt left
20:30 benchable6 left 20:32 benchable6 joined
lizmat mail.python.org/pipermail/python-c...05664.html # wow 20:40
yoleaux 16:25Z <AlexDaniel> lizmat: github.com/lizmat/P5substr/pull/1
lizmat AlexDaniel++
20:59 japhb_ joined 21:01 japhb left 21:04 skids left
Geth rakudo: lizmat self-assigned `my @foo is List` doesn't work and is reassignable github.com/rakudo/rakudo/issues/2060
1956df2552 | (Elizabeth Mattijsen)++ | 2 files

Partial fix to R#2052		 21:05
21:16 stmuk_ joined
synopsebot R#2052 [open]: github.com/rakudo/rakudo/issues/2052 [LTA] LTA error in construct declaring `sub class`
21:18 stmuk left
Geth nqp/truffle: 8b64366406 | (Daniel Green)++ | README.pod
[truffle] Some style fixes to readme
 21:21
rakudo: 21eb168ca4 | (Elizabeth Mattijsen)++ | src/core/traits.pm6
Fix "class A is B{ ... }" LTA error

Another partial fix to R#2052		 21:33
synopsebot R#2052 [open]: github.com/rakudo/rakudo/issues/2052 [LTA] "no compile time value" errors lack location of the problematic code
Geth rakudo/revert-2038-issue-2022: a6df35af67 | (Nick Logan)++ (committed using GitHub Web editor) | 3 files
Revert "Include source file name in module-not-found error message."
 21:39
lizmat ugexe++
Geth rakudo: ugexe++ created pull request #2061:
Revert "Include source file name in module-not-found error message."		 21:42
rakudo: lizmat self-assigned Initializing a Hash/Map from a list of a object hashes shows object hash internals github.com/rakudo/rakudo/issues/2062
a6df35af67 | (Nick Logan)++ (committed using GitHub Web editor) | 3 files
 21:49
rakudo: 612d071b83 | (Elizabeth Mattijsen)++ (committed using GitHub Web editor) | 3 files
Merge pull request #2061 from rakudo/revert-2038-issue-2022

Revert "Include source file name in module-not-found error message."
nqp/truffle: 9e59b16619 | (Daniel Green)++ | 2 files
[truffle] Implement nqp::codes
 22:01
nqp: 60f07ed644 | MasterDuke17++ (committed using GitHub Web editor) | docs/ops.markdown
Add nqp::codes to the docs
 22:05
rakudo: ugexe++ created pull request #2063:
Remove unspecced/used pre-6.c jvm modules		 22:39
22:55 stmuk joined 22:57 stmuk_ left 23:22 lizmat left 23:24 j3nnn1 left 23:25 lizmat joined

Powered by Cro and the RakuĀ® Programming Language.

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

Thank you!