00:04 patrickz joined 00:08 patrickb left
Geth ¦ problem-solving: MasterDuke17 assigned to jnthn Issue Expmod returns a default value instead of erroring out on libtommath errors github.com/Raku/problem-solving/issues/150 00:12
00:20 patrickz left 00:42 sena_kun left 00:58 sena_kun joined 01:02 MasterDuke left 01:03 Xliff left 01:14 Xliff joined
Xliff Files=1296, Tests=109781, 174 wallclock secs (27.67 usr 4.33 sys + 3165.51 cusr 159.69 csys = 3357.20 CPU) 01:15
TEST_JOBS=20 ;q
01:15 Xliff left 02:42 sena_kun left 02:51 AlexDaniel left 02:57 sena_kun joined, eaterof joined, eater left 03:03 hungrydonkey joined 03:09 hungrydonkey left 03:10 hungrydonkey joined 03:12 AlexDaniel joined, AlexDaniel left, AlexDaniel joined 03:13 hungrydonkey left 03:14 hungrydonkey joined 03:15 hungrydonkey left, hungrydonkey joined 03:18 hungrydonkey left, hungrydonkey joined 03:19 hungrydonkey left 03:20 hungrydonkey joined 03:21 hungrydonkey left, hungrydonkey joined 03:22 hungrydonkey left, hungrydonkey joined 03:24 hungrydonkey left, hungrydonkey joined 03:28 hungrydonkey left
Geth rakudo: 8c0a5d6b06 | (Ben Davies)++ | src/core.c/Parameter.pm6
Fix Parameter.sigil for anonymous scalar params with default values

This claimed their sigil was \ instead of $. Since anonymous sigilless params can't have default values, use the related flags to differentiate between the two.
04:00
04:42 sena_kun left 04:57 sena_kun joined 06:42 sena_kun left 06:56 sena_kun joined
Geth rakudo: Kaiepi++ created pull request #3430:
Simplify Parameter.raku, add Parameter.prefix and Parameter.suffix
07:05
07:07 domidumont joined 07:11 AlexDani` joined 07:18 AlexDaniel left, tbrowder left, tony-o left 07:26 domidumont left 07:40 domidumont joined 08:05 tbrowder joined, tony-o joined
nine patrickb: you have never answered my question: colabti.org/irclogger/irclogger_lo...-12-10#l39 08:21
tellable6 nine, I'll pass your message to patrickb
lizmat Files=1296, Tests=109769, 210 wallclock secs (28.37 usr 7.97 sys + 2956.88 cusr 249.80 csys = 3243.02 CPU) 08:30
tellable6 2020-01-16T21:20:51Z #raku-dev <vrurg> lizmat We loose lexically declared dynamic variables within react blocks. Similar happens to start, where Promise preserves the outer context. Does Supply code do the same somewhere?
lizmat .tell vrurg since most of Supply depends on dynamic variables, I doubt it, but jnthn would know 08:31
tellable6 lizmat, I'll pass your message to vrurg
08:41 sena_kun left 08:56 sena_kun joined 09:04 eaterof left
nine t/spec/S16-io/watch.t ............................................. Dubious, test returned 1 (wstat 256, 0x100) 09:06
I get this more often than not 09:07
09:07 eater joined
lizmat nine: on what OS ? 09:08
Geth rakudo: 1c46740142 | (Stefan Seifert)++ | src/core.c/CompUnit/PrecompilationRepository.pm6
Fix leaking file handles when loading modules

Since commit efadff2a39641e34a02b32c475b152448d360821 a PrecompilationUnit is marked uninitialized after closing. When we use an accessor like .checksum after closing (by !load-handle-for-path), the precomp file gets reopened and stays that way. When we load multiple modules with a shared dependency, we will detect that ... (5 more lines)
09:12
nine sena_kun: github.com/rakudo/rakudo/commit/1c46740142 is a candidate for the release. It fixes a regression introduced after the previous release. I'd judge it low risk (but not 0), medium benefit especially to people with large code bases. 09:14
lizmat: Linux
lizmat well, is it at all tested on Linux? I only recall writing tests for MacOS 09:15
sena_kun has checked blin output and didn't see anything too dangerous there at the first sight, still some new names there. got the point with cherry-pick, a release is likely 09:17
&afk
lizmat sena_kun++
nine sena_kun++ 09:18
Geth rakudo: 4366980681 | (Elizabeth Mattijsen)++ | 4 files
Merge Routine.rw/yada into a single generic bitmap

This makes each Routine carry one native int less, and opens up a way to add additional meta-information to a Routine without affecting its memory footprint.
This also adds a Routine.set_flag and Routine.get_flag method that take a bit number to set/get.
Not sure what performance effects this has, but needing 8 bytes less per Routine, can't be bad.
09:20
lizmat afk for a few hours& 09:35
09:58 |Tux| left 09:59 |Tux| joined 10:00 |Tux| left 10:14 |Tux| joined 10:36 pyrimidi_ joined 10:39 pyrimidine left 10:42 sena_kun left 10:56 sena_kun joined
Geth roast: Kaiepi++ created pull request #613:
Update tests WRT changes to Parameter.raku
11:11
11:28 brrt joined
|Tux| Rakudo version 2019.11-487-g436698068 - MoarVM version 2020.01-15-g1fa9764bf
csv-ip5xs0.708 - 0.771
csv-ip5xs-206.483 - 6.692
csv-parser22.969 - 23.592
csv-test-xs-200.423 - 0.439
test7.033 - 7.334
test-t1.739 - 1.742
test-t --race0.784 - 0.837
test-t-2030.182 - 30.395
test-t-20 --race8.465 - 8.519
11:33
11:47 brrt left
Geth roast: a4a4798b55 | (Elizabeth Mattijsen)++ | S02-types/compact.t
Remove tests for substr on an int array
12:02
roast: 3e4f35296c | (Elizabeth Mattijsen)++ | 13 files
Batch #19 of roast RT -> GH ticket migration
12:25
12:41 sena_kun left 12:57 sena_kun joined 13:47 TreyHarris left
Geth rakudo: f8724459f3 | (Elizabeth Mattijsen)++ | src/core.c/Supply-factories.pm6
Make Supply.comb(... :match) work

This will first collect all of the emitted strings, combine them into a single string, and emit the result of Str.comb with those parameters.
13:52
roast: 865b465174 | (Elizabeth Mattijsen)++ | S17-supply/comb.t
Adapt tests for Supply.comb(.... :match)
14:17 lucasb joined 14:41 sena_kun left 14:57 sena_kun joined
AlexDani` sena_kun: btw did we get results for the release before the last one too? 15:17
15:18 AlexDani` is now known as AlexDaniel, AlexDaniel left, AlexDaniel joined
AlexDaniel sena_kun: I mean --old=2019.07.1 15:18
16:04 domidumont left
sena_kun AlexDaniel: nope, but can do, I think. It doesn't look like an urgency to me, at the very least. 16:27
I mean, any SUPER SERIOUS things from there up to 2019.11 would be reported already anyway 16:29
AlexDaniel sena_kun: well, not necessarily 16:36
sena_kun: distro packages are still old, star wasn't released
tony-o .tell patrickb i have most of the infrastructure for an ecosystem fleshed out, i need some help testing it - i'm writing on a blog post on the current state rn 16:39
tellable6 tony-o, I'll pass your message to patrickb
AlexDaniel sena_kun: and also we had a lot of modules that are not passing tests but did months ago, it could be related 16:40
sena_kun: how long does it take for Blin to finish?
sena_kun: I suggest we run it overnight and see if there are any Bisected results
16:42 sena_kun left
nine AlexDaniel: the important part for a release is that it does not regress compared to the previous one. This way we ensure steady progress. Our tools let us down, so this was not necessarily true for some releases. But with what we know now we can be sure that 2020.01 is better than 2019.12 16:43
So it meets the criteria for releasing. 16:44
AlexDaniel it doesn't meet mine, but sena_kun can decide either way, I'm just suggesting
nine It doesn't fix all regressions we overlooked in the last release. But it's definitely better than the last release.
AlexDaniel if you look at Stars then some releases don't matter at all 16:45
nine will 2020.01 be a Star?
AlexDaniel no idea 16:46
but there wasn't a 2019.11
and then, I'm talking about just running the thing overnight to see if it brings anything up 16:47
nine That can't hurt in any case 16:48
16:57 sena_kun joined 18:17 Kaiepi left, Kaiepi joined 18:41 sena_kun left
lizmat I would really like to see a Rakudo Star 2020.01 before FOSDEM 18:56
18:57 sena_kun joined 19:58 MasterDuke joined
Geth roast: 4ae11ca905 | (Elizabeth Mattijsen)++ | 6 files
Batch #20 of roast RT -> GH ticket migration
20:01
timotimo this seems like a bug: 20:29
m: say ("42foo", *.chop.&val ... Numeric)
camelia (42foo 42fo 42f 42 4 ...)
timotimo m: say "42".&val.^name
camelia IntStr
timotimo m: say "42".&val.^name; say IntStr ~~ Numeric
camelia IntStr
True
timotimo committable6: releases say ("42foo", *.chop.&val ... Numeric) 20:31
committable6 timotimo, gist.github.com/20f48cddc87add87da...f65e94beb2
timotimo used to work for a while, and was broken in a different way for a little while 20:32
bisectable6: GOOD=2017.06 BAD=2017.07 say ("42foo", *.chop.&val ... Numeric) 20:33
bisectable6 timotimo, Bisecting by output (old=2017.06 new=2017.07) because on both starting points the exit code is 0
timotimo, bisect log: gist.github.com/488ba60002c90b4e56...92701c8ca6
timotimo, (2017-07-05) github.com/rakudo/rakudo/commit/00...270193c2f4
AlexDaniel and how is that relevant? 20:34
c: 001135ca2a^,001135ca2a say ("42foo", *.chop.&val ... Numeric)
committable6 AlexDaniel, ¦001135ca2a^: «(42foo 42fo 42f 42)␤» ¦001135c: «(42foo 42fo 42f 42 4 ...)␤»
AlexDaniel well… I don't know, but it is…
timotimo how is what relevant? 20:39
AlexDaniel timotimo: that commit
timotimo good q
must have changed semantics in some way by accident
we have coverable6, so i can find out what paths it takes at least 20:40
coverable6: 001135ca2a^,001135c say ("42foo", *.chop.&val ... Numeric)
coverable6 timotimo, Coverable only works with one commit
timotimo coverable6: 001135ca2a^ say ("42foo", *.chop.&val ... Numeric)
20:41 sena_kun left
coverable6 timotimo, gist.github.com/6964b9cea305631635...f6486fbfbc 20:41
timotimo coverable6: 001135c say ("42foo", *.chop.&val ... Numeric)
coverable6 timotimo, gist.github.com/6c56d0a83879b7765e...9091aaf86f 20:42
AlexDaniel now tell me honestly, does it help? :D
it's like the first time somebody used this bot in forever 20:43
timotimo m: say Numeric("42")
camelia Cannot invoke this object (REPR: Uninstantiable; Numeric)
in block <unit> at <tmp> line 1
timotimo m: say Int("42")
camelia 42
timotimo not completely certain 20:44
c: 001135ca2a^,001135ca2a say Numeric("42") 20:45
committable6 timotimo, ¦001135ca2a^,001135c: «Cannot invoke this object (REPR: Uninstantiable; Numeric)␤ in block <unit> at /tmp/Hd0ZK6CdIM line 1␤␤ «exit code = 1»»
timotimo fires up moar-remote 20:47
welp, time to debug the debugserver 20:54
actually, punt that to later 20:55
20:57 sena_kun joined
timotimo d'oh, helps to run the code you're actually thinking you're running 20:59
21:21 melezhik joined
AlexDaniel c: all say rand 21:22
committable6 AlexDaniel, gist.github.com/04edc44994f8153223...ce54314ff7 21:23
21:25 melezhik left
dogbert17 lizmat: Unexpected closing bracket: at /home/dogbert/repos/rakudo/t/spec/S02-types/flattening.t:125 21:48
Geth roast: b7561c2a14 | (Elizabeth Mattijsen)++ | S02-types/flattening.t
Oops, lost a curly

  dogbert17++
22:29
22:40 sena_kun left 22:45 lucasb left 22:55 sena_kun joined
lizmat m: say "{"\r".chars} + {"\n".chars} = {("\r" ~ "\n").chars}" # :-) 23:03
camelia 1 + 1 = 1
tony-o just updated from 2019.07 to blead and am getting 'Missing serialize REPR function for REPR VMException (BOOTException)' running nearly anything 23:13
Geth roast: 7082d7970b | (Elizabeth Mattijsen)++ | S17-supply/split.t
Use the needles!

Otherwise there is no point in defining them beforehand.
23:14
rakudo: 5d6c0b6b4e | (Elizabeth Mattijsen)++ | src/core.c/Supply-factories.pm6
Add support for Supply.encode($encoding = "utf8")
23:20
roast: 1810a494b6 | (Elizabeth Mattijsen)++ | S17-supply/encode.t
Add test file for Supply.encode tests
23:21
roast: 62de8f9984 | (Elizabeth Mattijsen)++ | spectest.data
Make sure the Supply.encode tests are run
23:22
lizmat tony-o: you should probably retry with a clean checkout
tony-o yea i nuked everything, figured out it looked like something corrupted, was just happening on two machines 23:28
one osx, one debian 23:29
Geth roast/rakudo_3257: 478ce94729 | (Vadim Belman)++ | 2 files
Consider newly added X::SymbolNotDynamic exception

Make _S32/exceptions/mist.t_ use v6.c explicitly
23:56
roast/rakudo_3257: 23778fabf8 | (Vadim Belman)++ | S02-names/pseudo-6e.t
Additional tests for dynamic pseudo-packages

Support for rakudo/rakudo#3272
roast/rakudo_3257: b9c2cf695c | (Vadim Belman)++ | S02-names/pseudo-6e.t
Add tests for dynamics within react/whenever blocks

When react runs under start it gets own continuation effectively loosing call stack and access to the outer dynamic variables.