Perl 6 language and compiler development | Logs at | For toolchain/installation stuff see #perl6-toolchain | For MoarVM see #moarvm
Set by Zoffix on 27 July 2018.
00:00 PufferBot left 00:03 PufferBot joined 00:32 fake_space_whale joined
Geth ¦ rakudo: zoffixznet self-unassigned `$`-sigiled constants don't get constant folded 00:57
¦ rakudo: zoffixznet self-unassigned [RFC] Implement `:handle-win-shell` arg to `run`, `shell`, and `` 00:58
¦ rakudo: zoffixznet self-unassigned Crazy Rats stringify as 0.00000…000Inf
¦ rakudo: zoffixznet self-unassigned Investigate whether $rat**$huge-power can be optimized with math smarts
¦ rakudo: zoffixznet self-unassigned %(), @(), $() are evil/weird
¦ rakudo: zoffixznet self-unassigned Native-typed constants NYI
¦ rakudo: zoffixznet self-unassigned Rational doesn't preserve types of nu/de
¦ rakudo: zoffixznet self-unassigned ZDRs broken with `<=>` and possibly other comparator ops
¦ rakudo: zoffixznet self-unassigned Can't use native type to parametarize a role / typecheck failure when using that type
¦ rakudo: zoffixznet self-unassigned Stringified Rat has traling zeros
¦ rakudo: zoffixznet self-unassigned Rat.Num does not choose the closest representable Num 00:59
¦ rakudo: zoffixznet self-unassigned When coercing to an unknown type, error is "Unable to parse expression in typename; couldn't find final ')'"
¦ rakudo: zoffixznet self-unassigned Type Captures Leaking out Sometimes
¦ rakudo: zoffixznet self-unassigned `my Int constant .= new: 42` doesn't crash
¦ rakudo: zoffixznet self-unassigned `.=` is implemented on constants, but barely functions
¦ rakudo: zoffixznet self-unassigned Some operations with zero-denominator Rationals are broken
¦ rakudo: zoffixznet self-unassigned constant does not support signature literal
01:44 BeastieBot joined 01:58 BeastieBot left 02:01 BeastieBot joined 02:19 ufobat___ joined 02:22 ufobat_ left 02:33 pmurias left
Geth roast: 84e4122481 | (Zoffix Znet)++ | S32-basics/pairup.t
[v6.d REVIEW] Fix up .pairup throwage test

Currently we rely on `dies-ok` sinking the returned Seq. Fix by sinking it ourselves and also test the exception type.
03:13 fake_space_whale left
Geth roast: 5f5c1bedf4 | (Zoffix Znet)++ | S32-array/splice.t
[v6.d REVIEW] Polish Array.splice tests

Toss pointless splits and use more appropriate test routine.
roast: ca11573429 | (Zoffix Znet)++ | S05-modifier/counted-match.t
Remove trailing whitespace
roast: a403c613c1 | (Zoffix Znet)++ | S05-modifier/counted-match.t
[v6.d REVIEW] More comprehensive Junction in :nth tests

  - Test various alternatives give right results
  - Test returned Junction contains all the alternatives
roast: 5742f3ef17 | (Zoffix Znet)++ | S03-operators/eqv.t
Remove trailing whitespcae
roast: 91265cee7a | (Zoffix Znet)++ | S03-operators/eqv.t
[v6.d REVIEW] Improve Setty eqv Setty tests

Spec exact return types.
roast: e542c16e0d | (Zoffix Znet)++ | S03-metaops/zip.t
[v6.d REVIEW] Test exact type in arity-0 Z test

Test claims we test for Seq, but `is-deeply` has a quirk of converting Seqs to lists. Fix by using cmp-ok with eqv instead.
roast: 08a209921d | (Zoffix Znet)++ | S05-metasyntax/angle-brackets.t
Remove trailing whitespace
roast: 74c3096a08 | (Zoffix Znet)++ | S05-metasyntax/angle-brackets.t
[v6.d REVIEW] Test exact types in .comb/match

roast: 98fd4c4119 | (Zoffix Znet)++ | S04-statements/for.t
[v6.d REVIEW] Test exact types in list comprehension

roast: 91ad3febb7 | (Zoffix Znet)++ | S07-iterators/range-iterator.t
[v6.d REVIEW] Remove protocol-violating Range Iterator tests
roast: 3aa9fef795 | (Zoffix Znet)++ | S07-iterators/range-iterator.t
[v6.d REVIEW] Remove pull test in skip-at-least

IMO pulling after a falsy skip is similar to trying to pull after getting an IterationEnd, and thus should be equivalently against the Iterator protocol.
roast: a60c4a94dd | (Zoffix Znet)++ | S06-signature/shape.t
[v6.d REVIEW] Test actual types in .shape test

And don't spec exact .perl
roast: c928f51574 | (Zoffix Znet)++ | S03-operators/gcd.t
[v6.d REVIEW] Test type in gcd tests

roast: ecf46934dc | (Zoffix Znet)++ | S09-multidim/methods.t
Remove trailing whitespace
roast: 04f36d15f3 | (Zoffix Znet)++ | S09-multidim/methods.t
[v6.d REVIEW] Test return types in .rotate/.reverse…

  …on multi dim arrays
roast: 8f39575321 | (Zoffix Znet)++ | S32-num/complex.t
[v6.d REVIEW] Comment on complex**0 reasoning

roast: 1bbd0334c2 | (Zoffix Znet)++ | S32-num/rat.t
[v6.d REVIEW] Loosen Rat.Str test

Here we just test we don't produce a "0.A", so allow either exact or rounded Str as result. Better to place rounding Str tests elsewhere (IIRC later proptest commits have those already)
roast: 344d60d154 | (Zoffix Znet)++ | 2 files
[v6.d REVIEW] Move NoMulti .subst test to appendix

6.d-prep: 86325a24a9 | (Zoffix Znet)++ (committed using GitHub Web editor) | TODO/
Update Zoffix's review commit
07:06 lizmat left 07:32 lizmat joined 07:36 lizmat left
Geth rakudo: JJ++ created pull request #2141:
Adds doc for option that apparently works
08:31 lizmat joined
lizmat Files=1246, Tests=76433, 386 wallclock secs (16.49 usr 5.81 sys + 2708.20 cusr 264.26 csys = 2994.76 CPU) 08:32
Geth rakudo: b5e92a5536 | (JJ Merelo)++ | src/Perl6/Compiler.nqp
Adds doc for option that apparently works
rakudo: 72c8cf68c3 | (Elizabeth Mattijsen)++ (committed using GitHub Web editor) | src/Perl6/Compiler.nqp
Merge pull request #2141 from JJ/master

Adds doc for option that apparently works
09:06 travis-ci joined
travis-ci Rakudo build passed. JJ Merelo 'Adds doc for option that apparently works' 09:06
09:06 travis-ci left
[Tux] Rakudo version 2018.06-259-g72c8cf68c - MoarVM version 2018.06-322-g41108cdaa
csv-ip5xs1.192 - 1.335
csv-ip5xs-207.896 - 8.177
csv-parser24.855 - 24.975
csv-test-xs-200.447 - 0.462
test10.343 - 10.444
test-t2.300 - 2.328
test-t --race1.016 - 1.062
test-t-2039.568 - 40.835
test-t-20 --race13.156 - 13.760
09:21 lizmat left 09:30 lizmat joined
gfldex m: say 5 ∩ (-∞..∞) 12:25
camelia Cannot coerce a lazy list onto a Set
in block <unit> at <tmp> line 1
MasterDuke now that i've converted nqp to generate *_i math ops instead of *_n, is there some way other than manually inspecting all the code to find where num math was actually expected for operators? 12:27
e.g., $a / $b, if those are ever and always ints, then it's fine. but if they are/could be nums, then it needs to be changed to nqp::div_n($a, $b) 12:28
12:32 ufobat___ left 13:48 lizmat left 13:51 stmuk_ joined 13:54 stmuk left 13:56 x[LGWs4x4i]uG2N0 left 14:00 x[LGWs4x4i]uG2N0 joined 14:30 x[LGWs4x4i]uG2N0 left 14:31 fake_space_whale joined 14:34 x[LGWs4x4i]uG2N0 joined 15:09 Ven`` joined 15:22 lizmat joined
Geth roast: ba9f5233c8 | (Zoffix Znet)++ | S32-num/complex.t
[v6.d REVIEW] Document how we came up with 1+0i for power
16:28 Kaiepi left 16:30 PufferBot left 16:52 fake_space_whale left 16:58 Ven`` left
Geth 6.d-prep: 014a14b244 | (Zoffix Znet)++ | 2 files
Defer deprecation/spec-errata policies to 6.e

IMO our current process is working sufficiently well to not require immediate implementations of these policies.
Defer creation of the policies until later language versions, until we get more data more informed.
17:07 pmurias joined 17:18 ufobat___ joined
Geth roast: cb721cb626 | (Zoffix Znet)++ | 2 files
[v6.d REVIEW] Improve .printf tests

  - Move no args case into appendices
  - Don't block `%z` for later use
roast: 8ad9599e25 | (Zoffix Znet)++ | S02-literals/radix.t
[v6.d REVIEW] Test :۳<12>'s return value too

roast: db0cbd1247 | (Zoffix Znet)++ | S17-procasync/basic.t
[v6.d REVIEW] Reword Proc::Async empty output test

Simplify code and test for any emission, not just those that are empty
roast: 6dc5661e62 | (Zoffix Znet)++ | S04-exceptions/fail.t
[v6.d REVIEW] Polish sink + Failure test

Explicitly set no fatal and don't rely on return value handling in the test routine.
19:13 PufferBot joined
Geth rakudo: 72831b9177 | (Aleks-Daniel Jakimenko-Aleksejev)++ | docs/release_guide.pod
2018.07 release is cancelled

After assessing all of the options this seems to be the best idea for this release.
rakudo: 0ec4f5f283 | (Aleks-Daniel Jakimenko-Aleksejev)++ | docs/release_guide.pod
Add more info to the release guide

Mention Toaster and say that cancelling is an option sometimes.
AlexDaniel Added some minimal info to the release guide 19:24
so “too long; didn't write” for the justification is: cutting from HEAD isn't a great idea because of all the issues; recutting 2018.06 without any commits is just somewhat wasteful; doing 2018.06 + a bunch of cherry-picked changes will require a full-blown release effort with testing which will potentially uncover new issues only happening on the release branch and we better put all that effort into 2018.08 19:34
stmuk_: ↑ so basically you can wait for 2018.08 or cut from 2018.06 19:36
stmuk_ I think I'll use 2018.06 because it would be good to get the Windows fixes out 19:38
so probably R* 2018.06 RC tomorrow 19:39
AlexDaniel cool 19:40
lizmat stmuk_++ 19:45
19:52 pmurias left
Geth rakudo: jstuder-gh++ created pull request #2142:
Add Callable, Whatever, and zen indexing candidates for CArray
19:57 Kaiepi joined
Geth roast: a43547db13 | (Zoffix Znet)++ | S32-list/rotor.t
[v6.d REVIEW] Also test :partial in empty-list rotor

roast: afe5a73980 | (Zoffix Znet)++ | S32-num/negative-zero.t
[v6.d REVIEW] Don't spec exact .perl of Complex

¦ rakudo: lizmat self-unassigned Constants slower than literals in range creation 21:55
rakudo/master: 4 commits pushed by (Jeremy Studer)++, (Zoffix Znet)++ 21:56
22:13 fake_space_whale joined 22:39 lizmat left 22:45 ufobat___ left 23:00 lizmat joined 23:12 Zoffix joined
Zoffix Talk Slides and Recording: "Faster Perl 6 Programs": 23:13
23:23 travis-ci joined
travis-ci Rakudo build failed. Zoffix Znet 'Merge pull request #2142 from jstuder-gh/carray_indexing_candidates 23:24
23:24 travis-ci left
buggable [travis build above] ☠ Did not recognize some failures. Check results manually. 23:24
Zoffix One job failed test t/09-moar/General_Category__UnicodeData__2.t 23:27