#perl6-dev on 16 December 2017 - Raku Programming Language Log

raschipi

Geth: help

00:47

Zoffix

OK. I give up :)

02:24

timotimo: are you still awake?

.tell timotimo the R#1308 release blocker is due to a JIT commit you made. Perhaps you'd know more what the problem is? I tried to fix it but failed. Some of the info on the bug I gathered: github.com/rakudo/rakudo/issues/13...-352154785

02:36

yoleaux

Zoffix: I'll pass your message to timotimo.

synopsebot

R#1308 [open]: github.com/rakudo/rakudo/issues/1308 [SEGV][severe][⚠ blocker ⚠] Frequent-ish SEQV in Perl 5 integration tests

samcv

Zoffix: how to get the number of cpu's with rakudo?

04:17

releasable6: status

04:18

releasable6

samcv, Next release in ≈14 hours. No blockers. 0 out of 125 commits logged

samcv, Details: gist.github.com/56910980b88b800603...7f811f483b

raschipi

m: say $*KERNEL.cpu-cores

04:39

camelia

raschipi

samcv: Is that what you want?

04:41

samcv

yep!

04:43

raschipi

Need a recent Rakudo version.

samcv

Zoffix: there's a bug in Test

05:24

or in prove i guess. if you have a multiple line text test, and you TODO it, then prove doesn't pick up that it's been todo'd

05:25

since the # TODO ends up on a different line from the "ok/not ok"

05:26

m: $*PROGRAM.substr(/".t"$/, "").trans: "-" => "/";

05:43

camelia

No such method '!cursor_start' for invocant of type 'Int'
in block <unit> at <tmp> line 1

samcv

anybody know what's going on here?

anyway i need to go back to sleep. be back in the morning (like 4 hours from now)

05:45

Zoffix

samcv: you wrote "substr" instead of "subst"

06:13

.tell timotimo One thing I noticed in graph.c, category is MVMint32, but in interp.c it's a MVMuint32 and I think I seen it as MVMint16 somewhere too. Dunno if it matters; tried making 'em the same, but that didn't fix nothing

06:21

yoleaux

Zoffix: I'll pass your message to timotimo.

Geth

rakudo: 04eb1da6af | (Zoffix Znet)++ | t/05-messages/02-errors.t
Test C3 linearization mentions line number

RT#126539: rt.perl.org/Ticket/Display.html?id=126539

06:48

synopsebot

RT#126539 [new]: rt.perl.org/Ticket/Display.html?id=126539 [LTA] error message about C3 linearization (no line number)

Zoffix

.tell wander do you have an RT account? You committed tests for a ticket ( github.com/perl6/roast/commit/0368...dc95e1d218 ) but nothing's mentioned on the ticket itself and it's still open: rt.perl.org/Ticket/Display.html?id...et-history

06:54

yoleaux

Zoffix: I'll pass your message to wander.

AlexDaniel

samcv: o/

07:24

ah whoops. Sleep well

07:25

e: say $*KERNEL.cpu-cores

evalable6

AlexDaniel

Zoffix++

07:27

.oO( time to find more blockers before the existing ones are resolved!! )

07:28

.seen Ulti

07:55

yoleaux

I saw Ulti 15 Dec 2017 22:24Z in #perl6: <Ulti> no then they might hire you

AlexDaniel

tbrowder: you here?

09:39

[Tux]

Rakudo version 2017.11-126-g04eb1da6a - MoarVM version 2017.11-43-g93a77dc16

csv-ip5xs	1.086	-	1.093
csv-ip5xs-20	12.989	-	13.136
csv-parser	12.329	-	12.623
csv-test-xs-20	0.456	-	0.483
test	11.592	-	11.986
test-t	2.905	-	2.928
test-t --race	1.253	-	1.280
test-t-20	51.997	-	52.417
test-t-20 --race	18.063	-	18.168

10:22

lizmat

m: class A {}; sub prefix:<++>(A) { say "foo" }; sub postfix:<++>(A) { say "bar" }; A++ # TimToady feels to me the static ++ optimization is too early

10:26

camelia

foo

lizmat

TimToady: expected to see "bar" there

samcv

Zoffix: yeah i know but the error is LTA :(

10:36

m: $*PROGRAM.substr(/".t"$/, "")

10:37

camelia

No such method '!cursor_start' for invocant of type 'Int'
in block <unit> at <tmp> line 1

samcv

m: $*PROGRAM.substr(/".t"$/)

camelia

No such method '!cursor_start' for invocant of type 'Int'
in block <unit> at <tmp> line 1

samcv

i would expect it to give me an error about a signature issue, but it gives a cursor error.

10:39

ah see a github issue open there. thanks!

AlexDaniel

mch: class A {}; sub prefix:<++>(A) { say "foo" }; sub postfix:<++>(A) { say "bar" }; A++

11:06

committable6

AlexDaniel, ¦2015.12: «bar␤» ¦HEAD(04eb1da): «foo␤»

AlexDaniel

damn I'm looking at github.com/rakudo/rakudo/commit/fc52143b

and I remember there was some RT I was working on…

and I got stuck exactly because of that error

err, issue

hmm, I think it was RT#127775

11:09

synopsebot

RT#127775 [new]: rt.perl.org/Ticket/Display.html?id=127775 Declaring enums with Bools, IntStrs and maybe other things ( enum Foo (:Bar(1), :Baz(True) ))

AlexDaniel

but nothing has changed

ok nvm

11:10

tbrowder

AlexDaniel i’m here

11:21

AlexDaniel

oh, cool

e: gist.githubusercontent.com/AlexDan...podregr.p6

11:22

evalable6

AlexDaniel, Successfully fetched the code from the provided URL.

"'Foo and Bar'"

AlexDaniel

tbrowder: that doesn't look quite right, right?

but I'm kinda confused, didn't we fix this already?

tbrowder

no, not yet

11:23

AlexDaniel

tbrowder: can you try running Pod::TreeWalker tests?

11:24

there's one test that is broken now

tbrowder

AlexDaniel

it is related to :caption

tbrowder

noted...

11:25

AlexDaniel

tbrowder: let me know if that's something we should fix in rakudo before the release or if it's just something that needs to be adjusted in Pod::TreeWalker

(or maybe it's something we should ignore altoghether…)

tbrowder

how much time to release?

11:26

AlexDaniel

tbrowder: many-many hours :)

11:27

tbrowder

roger...

AlexDaniel

Zoffix: IIRC wander was using emails to leave their comments, and RT shows them only a few days later

11:28

with unknown success rate…

11:29

.tell wander by any chance, does that sound relevant? irclog.perlgeek.de/perl6/2017-12-15#i_15583699

yoleaux

AlexDaniel: I'll pass your message to wander.

samcv

ok so i have tests now that test lots of codepoints

11:30

tests Name of all codepoints except Hangul's and LineBreak and GeneralCategory

11:31

takes ~14s to run, but if you parallel it then it's faster

tbrowder

ref treewalker, in the test table change “:caption(‘Foo and Bar”) to “:caption<Foo and Bar>” and it should test ok.

samcv

but it uses nqp so it probably needs to go in rakudo test set

AlexDaniel

tbrowder: yeah, but why was it working before and now doesn't?

11:39

samcv: that sounds awesome!

FWIW current blockers: github.com/rakudo/rakudo/labels/%E...0%E2%9A%A0

huggable: no release :is: See current blockers: github.com/rakudo/rakudo/labels/%E...0%E2%9A%A0

11:40

huggable

AlexDaniel, Added no release as See current blockers: github.com/rakudo/rakudo/labels/%E...0%E2%9A%A0

AlexDaniel

huggable: blockers :is: github.com/rakudo/rakudo/labels/%E...0%E2%9A%A0

huggable

AlexDaniel, Added blockers as github.com/rakudo/rakudo/labels/%E...0%E2%9A%A0

lizmat

samcv: yeah, should live under "make test"

11:43

tbrowder

AlexDaniel: when was before? the :caption capability as a property was recently added.

lizmat

samcv: re substr(Regex), what was the issue number ?

m: say Regex ~~ Callable

11:44

camelia

True

lizmat

samcv: underlying issue of the problem

samcv

github.com/rakudo/rakudo/issues/1314

11:50

that makes more sense now :)

AlexDaniel

tbrowder: before (2017-12-11) github.com/rakudo/rakudo/commit/2c...c360623c99

11:52

c: 2cd266fe08^,2cd266fe08 gist.githubusercontent.com/AlexDan...er-regr.p6

11:53

committable6

AlexDaniel, Successfully fetched the code from the provided URL.

AlexDaniel, gist.github.com/ca337853fc2fe8640b...76c82f7969

AlexDaniel

tbrowder: ↑

11:54

lizmat

samcv: still not suer why it selects the candidate that it does

11:55

tbrowder

AlexDaniel: i’m stumped. the caption property fix did nothing i can tell to its representation. how it was working before-hand is a mystery since that string bug has been around since at least 2015.

12:11

Geth

rakudo: samcv++ created pull request #1316:
Add tests for testing almost all Unicode codepoints

synopsebot

RAKUDO#1316 [open]: github.com/rakudo/rakudo/pull/1316 Add tests for testing almost all Unicode codepoints

samcv

hmm ok that pr is messed up somehow

12:12

for some reason it shows tons of other commits in the list even though i forked it straight from master

12:13

oh it's trying to merge into nom

12:14

lizmat

yeah, been quite a few commits since master diverged from nom :-)

samcv

ok cool that's fixed now. takes ~13 seconds single threaded and much faster if it's threaded

and is easy to add tests. all you need to do is copy the test file and change the name based on the file the test is generated from. Line_Break:LineBreak.t tests Line_Break property from LineBreak.txt

12:16

and then add an item to the hash in UnipropCheck.pm6 to tell it if it's a str/int/name

if we start passing more than we are supposed to, a todo will pass. in which case you need to change the number for %expected-loses hash and then the it will then be todo'd and failing (well until which case we get to passing all of them) then it won't be todo'd and won't pass

12:18

which makes sure we get positive confirmation of fixes and negative confirmation of having more failures

12:19

Geth

rakudo: 1da698ab04 | (Elizabeth Mattijsen)++ | src/core/Str.pm
At least temporarily fix GH #1314

  - Regex ~~ Callable is True, that's why it is trying to call it
  - adding a Regex candidate for start did *not* fix the issue
  - I assume that somehow the coercing makes the candidate narrower
  - so fixed by checking for Regex then and there
It should be noted that this reminded me of an issue that the sub version of an action should call the method version, not the other way around.
  "substr" and others are cases where we didn't fix this yet. Should maybe
looked at after the 2017.12 release.

synopsebot

RAKUDO#1314 [open]: github.com/rakudo/rakudo/issues/1314 [LTA] Bad error when mistakenly using `substr` as `subst`

tbrowder

AlexDaniel: i see treewalker passing now???