Perl 6 language and compiler development | Logs at colabti.org/irclogger/irclogger_log/perl6-dev | For toolchain/installation stuff see #perl6-toolchain | For MoarVM see #moarvm
Set by Zoffix on 27 July 2018.
AlexDaniel jnthn: maybe this is an interesting issue: github.com/rakudo/rakudo/issues/2538 00:02
Geth rakudo: bca05ae5e8 | (Aleks-Daniel Jakimenko-Aleksejev)++ | src/Perl6/Actions.nqp
Revert "Avoid repeated PseudoStash creation in regexes"

This reverts commit e747b19a792d7c610845135bd2ef6b6ca1de9407.
  See rakudo issue #2544.
00:03
synopsebot RAKUDO#2544 [open]: github.com/rakudo/rakudo/issues/2544 [⚠ blocker ⚠] Missing serialize REPR function for REPR MVMContext (BOOTContext)
jnthn AlexDaniel: It is; it's on my list of things to look at 00:04
(Though I won't get to it before next week; I've been really exhausted this week, and need to spend the weekend resting rather than debugging tricky stuffs...)
japhb Proper prioritization, that 00:11
AlexDaniel c: 2018.11,e747b19a79^,e747b19a79,HEAD gist.github.com/AlexDaniel/815d139...3fcf07d64b 00:15
committable6 AlexDaniel, Using file “a.p6” as a main file, other files are placed in “sandbox/”
AlexDaniel, ¦2018.11,e747b19a79^,HEAD(bca05ae): «» ¦e747b19: «===SORRY!===␤Missing serialize REPR function for REPR MVMContext (BOOTContext)␤ «exit code = 1»» 00:16
AlexDaniel ok cool
sooo hmm 00:19
nowadays we're expecting both 6.c-errata and 6.d-errata to pass, right?
00:20 cognominal joined 00:21 p6bannerbot sets mode: +v cognominal
jnthn AlexDaniel: Think so 00:23
00:34 dct left 00:47 MasterDuke joined, p6bannerbot sets mode: +v MasterDuke, MasterDuke left, MasterDuke joined, herbert.freenode.net sets mode: +v MasterDuke, p6bannerbot sets mode: +v MasterDuke
AlexDaniel just wrote `while without $x { }` and expected it to work 00:49
00:53 cognominal left
Geth rakudo: a87c27e4de | (Aleks-Daniel Jakimenko-Aleksejev)++ | docs/release_guide.pod
Mention 6.d-errata in the release guide
01:06
rakudo: 19238b870e | (Aleks-Daniel Jakimenko-Aleksejev)++ | tools/release-dates.p6
Add 2019 checkpoint to release-dates.p6

Also make release-dates.p6 more readable (it was hard to follow the math in the previous version).
rakudo: 0f3c370a4a | (Aleks-Daniel Jakimenko-Aleksejev)++ | docs/release_guide.pod
Populate release dates for 2019
AlexDaniel btw t/spec/S03-buf/read-int.t does not pass on 6.c-errata
01:24 lucasb left 01:56 lizmat left 02:51 cognominal joined 02:52 p6bannerbot sets mode: +v cognominal 03:20 ufobat___ joined 03:21 p6bannerbot sets mode: +v ufobat___ 03:23 ufobat_ left 03:24 cognominal left 04:13 MasterDuke left 05:21 cognominal joined 05:22 p6bannerbot sets mode: +v cognominal 05:55 cognominal left
Geth rakudo: 370310b35e | (Nick Logan)++ (committed using GitHub Web editor) | src/core/Distribution.pm6
Fix resource content lookup for Distribution::Path

Fixes content lookup on a Distribution::Path with a parameter starting with 'resources/libraries/` since the platform specific name will be known to some Distributions.
06:48
roast/cur-candidates: 738 commits pushed by 30 authors
review: github.com/perl6/roast/compare/d2a...91aafa4c3e
06:52
roast/cur-candidates: 15399b07e7 | (Nick Logan)++ | packages/Test-Helpers/lib/Test/Util.pm6
Add make-test-dist test utility

Generates a non-installation on-disk representation of a distribution from meta data.
07:00
roast/cur-candidates: 57cea78684 | (Nick Logan)++ | S11-repository/cur-candidates.t
Add module lookup / identification tests

  * Adds the tests for CUR's .candidates, and by extension much
of the CUR interface ( uninstall, resolve, etc )
  * Adds some basic tests for both CURFS and CURI
  * Adds basic %?RESOURCES test
roast/cur-candidates: 15399b07e7 | (Nick Logan)++ | packages/Test-Helpers/lib/Test/Util.pm6
Add make-test-dist test utility

Generates a non-installation on-disk representation of a distribution from meta data.
07:07
roast/cur-candidates: 57cea78684 | (Nick Logan)++ | S11-repository/cur-candidates.t
Add module lookup / identification tests

  * Adds the tests for CUR's .candidates, and by extension much
of the CUR interface ( uninstall, resolve, etc )
  * Adds some basic tests for both CURFS and CURI
  * Adds basic %?RESOURCES test
07:11 AlexDaniel left 07:52 cognominal joined, p6bannerbot sets mode: +v cognominal 08:24 cognominal left 08:28 lizmat joined, p6bannerbot sets mode: +v lizmat
lizmat bisectable6: dd blob8.new(42,127).read-int8(-1) 09:18
bisectable6 lizmat, Bisecting by exit code (old=2015.12 new=370310b). Old exit code: 1
lizmat, bisect log: gist.github.com/81cbbedf9f6d48ed51...92887e720b
lizmat, (2018-12-15) github.com/rakudo/rakudo/commit/92...93492635f4
lizmat .ask nine looks like NQP 3fe17a683 suddenly gave end-of-buffer semantics to negative indexes on read-intX, was that intentional? If so, I guess I need to move the test for negative indices to the setting 09:20
yoleaux lizmat: I'll pass your message to nine.
[Tux] Rakudo version 2018.11-68-g370310b35 - MoarVM version 2018.11-37-g30088ed43
csv-ip5xs0.912 - 0.940
csv-ip5xs-206.969 - 7.138
csv-parser22.688 - 23.009
csv-test-xs-200.431 - 0.439
test6.821 - 7.003
test-t1.779 - 1.809
test-t --race0.810 - 0.831
test-t-2030.239 - 31.067
test-t-20 --race9.436 - 10.072
09:58
lizmat Files=1257, Tests=83578, 341 wallclock secs (15.43 usr 5.58 sys + 2419.93 cusr 215.70 csys = 2656.64 CPU)
10:21 cognominal joined 10:22 p6bannerbot sets mode: +v cognominal 10:41 robertle joined 10:42 p6bannerbot sets mode: +v robertle 10:55 cognominal left 11:22 lucasb joined, p6bannerbot sets mode: +v lucasb 12:20 cognominal joined 12:21 p6bannerbot sets mode: +v cognominal 12:24 cognominal left
lizmat .ask AlexDaniel is it still possible to commit for the release or not ? 12:39
yoleaux lizmat: I'll pass your message to AlexDaniel.
lizmat .tell AlexDaniel asking because of github.com/MoarVM/MoarVM/issues/1024
yoleaux lizmat: I'll pass your message to AlexDaniel.
timotimo github.com/MoarVM/MoarVM/commit/7df8596eb 13:13
^- lizmat, "Other array ops treat negative indices as counting from the end of the array.
Do the same for the bytestream handling ops.
"
lizmat I guess that answers that question
feels to me that is propagating incorrect baheviour 13:14
the index < 0 handling is *sooo* Perl 5
and it causes quite a lot of overhead on the HLL side :-(
slowing down *all* indexed accesses 13:15
lizmat goes off adding checks in the settinng
13:31 cognominal joined 13:32 p6bannerbot sets mode: +v cognominal
timotimo do we actually use those semantics anywhere at all? 13:36
should i try to add an exception to moarvm's code and run a spectest to see?
lizmat timotimo: that sounds like an excellent idea :-) 13:37
timotimo building it now 13:39
MVMArray: Index out of bounds 13:40
at gen/moar/stage2/QRegex.nqp:1569 (src/vm/moar/stage0/QRegex.moarvm:!cursor_push_cstack)
should be able to fix code as it comes up, though, right?
lizmat think so... 13:41
timotimo will need a rebootstrap, though
OK, looks like we do actually generate access to negative indices in regex code 13:46
lizmat I guess another legacy from the Parrot days :-( 13:47
timotimo: if you could generate a gist of places where that is happening, I could have a go at it as well 13:48
timotimo i should be able, actually 13:49
i now have a cat on my lap 13:51
lizmat
.oO( priorities priorities :-)
13:52
13:59 cognominal left 14:00 cognominal joined 14:01 p6bannerbot sets mode: +v cognominal
timotimo OK 14:02
here's a compressed text file, it's about 500 megabytes, but i didn't deduplicate at all yet
hack.p6c.org/~timo/nqp_all_usages_o...s.txt.lzma
oh
720 megabytes
uncompressed, that is
compressed it's only 400 kilobytes 14:03
jnthn Yeah, the regex engine will doubtless have been taking advantage of that to get smaller code :)
Not everything will win.
timotimo add an op that does the calculation itself :) :) 14:04
14:04 cognominal left
lizmat hmmm... macOS doesn't know about .lzma : ( 14:06
timotimo oh
do yo uhave a tool "7za"?
i can recompress it on hack
haha 14:07
hack doesn't have 7za either :D
Geth rakudo: 5f3a955aed | (Elizabeth Mattijsen)++ | src/core/Buf.pm6
Add negative offset checks to read-int/write-int

To at least make sure we don't release a version of these that do allow negative offsets.
14:09
lizmat I'm looking forward to be able to revert that one ^^^ 14:10
timotimo hack.p6c.org/~timo/nqp_all_usages_o...es.txt.bz2 - not finished compressing yet
hum. bzip2 isn't extremely fast, that's for sure 14:12
lizmat it's asymmetrical, I seem to recall, yeah 14:13
timotimo so is 7za, at least when i used -9 :)
why did i use -9 for bzip2 as well?
done! 14:14
1.4 megabytes, wow
lizmat 404 on that link ? 14:16
timotimo oh? 14:20
hold on
ah, yes
//hack.p6c.org/~timo/nqp_all_usages...es.txt.bz7
i typod that
lizmat: gist.github.com/timo/6f88a234a72a1...b592364ee9 - i already had this patch applied when i generated that output 14:28
14:37 cognominal joined 14:38 p6bannerbot sets mode: +v cognominal
Geth rakudo: c00aef29f9 | (Stefan Seifert)++ | src/core/Buf.pm6
Revert "Add negative offset checks to read-int/write-int"

This reverts commit 5f3a955aed468836a6e504995b66a877e912dfb9.
The VM now catches these erroneous cases for us.
15:00
nqp: e4c40b5987 | (Elizabeth Mattijsen)++ | tools/build/MOAR_REVISION
Bump NQP for read/write-int fixes, nine++
15:07
¦ nqp: version bump brought these changes: github.com/MoarVM/MoarVM/compare/2...8-gfa34110
15:10 cognominal left
nine jnthn: do you have a hunch about the invoke_v optimization issue? github.com/rakudo/rakudo/issues/2531 15:10
Geth rakudo: 79dd1c8e40 | (Elizabeth Mattijsen)++ | tools/build/NQP_REVISION
Bump NQP to get a MoarVM with read-int/write-int fixes
15:28
¦ rakudo: version bump brought these changes: github.com/perl6/nqp/compare/2018....6-ge4c40b5
15:28 travis-ci joined, p6bannerbot sets mode: +v travis-ci
travis-ci Rakudo build failed. Stefan Seifert 'Revert "Add negative offset checks to read-int/write-int" 15:28
travis-ci.org/rakudo/rakudo/builds/468417736 github.com/rakudo/rakudo/compare/5...0aef29f9b0
15:28 travis-ci left
lizmat afk& 15:28
buggable [travis build above] ☠ Did not recognize some failures. Check results manually.
nine m: sub a(uint $a) { dd $a }; a -1 15:43
camelia -1
nine That's a curious one. I actually don't see anything in the compiled code that would prevent this.
15:44 patrickb joined 15:45 p6bannerbot sets mode: +v patrickb 15:57 travis-ci joined, p6bannerbot sets mode: +v travis-ci
travis-ci Rakudo build passed. Elizabeth Mattijsen 'Bump NQP to get a MoarVM with read-int/write-int fixes' 15:57
travis-ci.org/rakudo/rakudo/builds/468424426 github.com/rakudo/rakudo/compare/c...dd1c8e4075
15:57 travis-ci left 15:59 cognominal joined 16:00 p6bannerbot sets mode: +v cognominal 16:02 AlexDaniel joined, p6bannerbot sets mode: +v AlexDaniel
AlexDaniel . 16:03
yoleaux 09:35Z <jmerelo> AlexDaniel: (see conversation above). Is it OK if we make the next Squashathon a "testneeded" one?
12:39Z <lizmat> AlexDaniel: is it still possible to commit for the release or not ?
12:39Z <lizmat> AlexDaniel: asking because of github.com/MoarVM/MoarVM/issues/1024
16:03 cognominal left
Geth rakudo/cur-candidates: 34 commits pushed by (Ben Davies)++, (Aleks-Daniel Jakimenko-Aleksejev)++, (Elizabeth Mattijsen)++, (Jonathan Worthington)++, (Nick Logan)++, (Stefan Seifert)++
review: github.com/rakudo/rakudo/compare/d...c884307da4
16:05
AlexDaniel lizmat: yeah it should be ok (I guess something is already committed?)
lizmat yeah, already committed 16:19
and reverted and etc. :-)
timotimo aaah why do you write "and etc"? :D 16:24
16:26 [Tux] left 16:31 [Tux] joined 16:32 p6bannerbot sets mode: +v [Tux] 16:42 cognominal joined 16:43 p6bannerbot sets mode: +v cognominal 16:53 patrickz joined 16:54 p6bannerbot sets mode: +v patrickz 16:56 patrickb left 17:06 AlexDaniel left 17:24 cognominal left 17:25 cognominal joined 17:26 p6bannerbot sets mode: +v cognominal 17:30 cognominal left 18:03 cognominal joined 18:04 p6bannerbot sets mode: +v cognominal 18:29 AlexDaniel joined, p6bannerbot sets mode: +v AlexDaniel 18:42 cognominal left 18:43 cognominal joined 18:44 p6bannerbot sets mode: +v cognominal
Geth rakudo: ceaf7218fc | (Elizabeth Mattijsen)++ | src/core/Buf.pm6
Add initial implementation of write-int128

Which I didn't have time yet to do.
18:44
18:47 cognominal left 19:01 patrickz left
lizmat afk for the rest of the day& 19:13
19:16 cognominal joined 19:17 p6bannerbot sets mode: +v cognominal 19:23 jmerelo joined 19:24 p6bannerbot sets mode: +v jmerelo
jmerelo .tell masak we just panic a bit if we don't see the article scheduled the previous day... 19:24
yoleaux jmerelo: I'll pass your message to masak.
19:50 cognominal left 19:57 cognominal joined, p6bannerbot sets mode: +v cognominal
lucasb Can Perl 6 have string constructors Str.new(obj) and Str.new(a, b, c), for uniformity with other type constructors? Should evaluate to obj.Str and (a, b, c).Str. 20:09
20:16 jmerelo left 20:18 dct joined, p6bannerbot sets mode: +v dct
AlexDaniel releasable6: status 20:45
releasable6 AlexDaniel, Next release will happen when it's ready. 3 blockers. 62 out of 72 commits logged (⚠ 1 warnings)
AlexDaniel, Details: gist.github.com/5dbbbaaccecc80d406...04b88699ec
20:49 lizmat left 20:50 cognominal left 20:51 cognominal joined 20:52 p6bannerbot sets mode: +v cognominal
Geth nqp: 2fb17c58b1 | (Aleks-Daniel Jakimenko-Aleksejev)++ | src/vm/moar/QAST/QASTOperationsMAST.nqp
Revert "Implement VOID context optimization for method calls"

This reverts commit a20a0a14a8c12ae0cabb883f10026c883edf6a91.
  See rakudo/rakudo#2531
20:56
synopsebot RAKUDO#2531 [open]: github.com/rakudo/rakudo/issues/2531 [regression][⚠ blocker ⚠] No such method 'payload' for invocant of type 'X::Parameter::InvalidConcreteness'
20:58 lizmat joined, p6bannerbot sets mode: +v lizmat 21:01 cognominal left, cognominal joined 21:02 p6bannerbot sets mode: +v cognominal 21:03 lizmat left 21:06 cognominal left 21:10 dct left
Geth rakudo: 203487f3e2 | (Aleks-Daniel Jakimenko-Aleksejev)++ | tools/build/NQP_REVISION
[NQP Bump] 2fb17c58b Revert "Implement VOID co […]

NQP bump brought: github.com/perl6/nqp/compare/2018....g2fb17c58b
21:19
rakudo: version bump brought these changes: github.com/perl6/nqp/compare/2018....g2fb17c58b
41d2da0958 | (Aleks-Daniel Jakimenko-Aleksejev)++ | docs/release_guide.pod
AlexDaniel releasable6: status 21:26
releasable6 AlexDaniel, Next release will happen when it's ready. 2 blockers. 62 out of 73 commits logged (⚠ 1 warnings)
AlexDaniel, Details: gist.github.com/9c56f1ef2c081ea414...8071f2b0ec
AlexDaniel alright down to 2…
now dat 6.c-errata failure… 21:27
21:35 cognominal joined, p6bannerbot sets mode: +v cognominal
AlexDaniel ah it's gone? 21:54
22:08 cognominal left 22:12 cognominal joined 22:13 p6bannerbot sets mode: +v cognominal 22:26 cognominal left 22:27 cognominal joined 22:28 p6bannerbot sets mode: +v cognominal 22:31 ExtraCrispy left 22:39 shareable6 left, evalable6 left, committable6 left, bisectable6 left, greppable6 left, quotable6 left, committable6 joined 22:40 shareable6 joined, ChanServ sets mode: +v shareable6, p6bannerbot sets mode: +v committable6, p6bannerbot sets mode: +v shareable6 22:41 evalable6 joined, ChanServ sets mode: +v evalable6, bisectable6 joined 22:42 p6bannerbot sets mode: +v evalable6, p6bannerbot sets mode: +v bisectable6, cognominal left, quotable6 joined, cognominal joined 22:43 p6bannerbot sets mode: +v quotable6, greppable6 joined, ChanServ sets mode: +v greppable6, cognomin_ joined, p6bannerbot sets mode: +v cognominal 22:44 p6bannerbot sets mode: +v greppable6, p6bannerbot sets mode: +v cognomin_ 22:47 cognominal left 22:48 cognomin_ left 22:51 AlexDani` joined, p6bannerbot sets mode: +v AlexDani` 22:52 AlexDaniel left 22:53 AlexDani` is now known as AlexDaniel 23:22 lucasb left 23:25 cognominal joined 23:26 p6bannerbot sets mode: +v cognominal 23:35 cognominal left
timotimo .tell lucasb why would it be (a, b, c).Str and not (a, b, c).join("")? 23:42
yoleaux timotimo: I'll pass your message to lucasb.