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.
02:29 epony left
Kaiepi "it's not stale precomp" 02:46
"there's no way it's stale precomp!"
...
"it was stale precomp"
03:14 epony joined 05:10 mst left
nine Kaiepi: how did you end up with stale precomp? 07:44
Kaiepi not sure 07:45
this time in particular it happened with a couple test files in rakudo's test suite, but it happens every now and then 07:46
AlexDaniel fwiw I'm seeing precomp issues all the time, and it was like this since the beginning… 08:10
that said, I never really ticketed anything because nuking always helps, and I have no idea how to reproduce any of these issues 08:11
:S
lizmat AlexDaniel: just here for a moment 09:21
re cached Seq's
your example uses a reified list (1,2,3)
in that case the sequence uses the reified list as its cachew
afk again&
problem something for the problem solving repo: 10:11
do we want to keep our open / all bugs from RT around in some form?
in the even that RT is shut down 10:12
Geth rakudo: 0c8f960e87 | (Elizabeth Mattijsen)++ | 2 files
Revert "Do correct assignment to SetHash"

This reverts commit f1512e9f976b4f57f39db809bf69bc068a70e7e5.
  - changes current semantics on which code may rely
   - so should probably be version based
  - doesn't fix the same issue that Set.STORE has
  - doesn't fix more generally issue that QuantHash.STORE(QuantHash) may have
  - is for a ticket that is more than 1.5 years old
   - so there is no urgency to fix this now just before a release
   that is already 2 months overdue
11:19
lizmat ticket in question: R#1203
synopsebot_ R#1203 [open]: github.com/rakudo/rakudo/issues/1203 [data types] Would be nice to have an easy way to work with SetHashes and its kin
11:24 Kaiepi joined
Geth roast: 0f5ace21a2 | (Elizabeth Mattijsen)++ | S03-operators/assign.t
Revert "Test for SetHash assignment"

This reverts commit 12298da5a6169d441a28141430cf328217bbfc93.
Reverting, because the associated Rakudo commit also got reverted.
12:21
12:23 travis-ci joined
travis-ci Rakudo build failed. Elizabeth Mattijsen 'Revert "Do correct assignment to SetHash" 12:23
travis-ci.org/rakudo/rakudo/builds/549069736 github.com/rakudo/rakudo/compare/e...8f960e87f8
12:23 travis-ci left
lizmat appears to be unrelated 12:24
restarted the one failing job, to see if it's a flapper
github.com/perl6/problem-solving/issues/47 12:26
Metaop semantics with QuantHashes
12:42 dogbert17 joined
lizmat Travis all green now 12:42
13:53 lucasb joined 14:11 mst joined
vrurg lizmat: I think it was preliminary to close 1203. The right answer was with my patch which you reverted. 14:14
lizmat reopened 1203 14:15
Please see github.com/perl6/problem-solving/issues/47 14:16
vrurg I saw your problem-solving ticket, that's what I was about to ask for. ;) Need to checkout from the hotel now, but gonna have plenty of time in the airport to answer.
lizmat vrurg: ok 14:17
MasterDuke lizmat: can you reliably recreate that odd profile you showed me a while ago? 14:18
lizmat I'm afraid not :-( 14:19
MasterDuke ugh, not here either
AlexDaniel lizmat: reportable6 has a backup of all RT tickets, as far as I know 14:23
lizmat AlexDaniel: good to know 14:24
MasterDuke timotimo: btw, any idea why `<anon> gen/moar/stage2/NQPHLL.nqp:1825` is now showing up at the top of profiles? i thought i managed to reset profiling state after calculating the overhead
Geth ¦ problem-solving: AlexDaniel assigned to jnthn Issue Metaop semantics with QuantHashes github.com/perl6/problem-solving/issues/47
AlexDaniel lizmat: given the amount of activity on RT, we can write it to a USB drive and then frame it on the wall 14:25
lizmat well, RT will be shut off pretty soon... so if we want that USB drive, we're going to need to fill that now
AlexDaniel “All RT tickets in perl6 queue, up-to-date.”
lizmat: where can I read about that? 14:26
lizmat about RT being shut down ?
AlexDaniel yeah
lizmat looks in her mail 14:27
MasterDuke random question, any idea why `say (^8).permutations>>.sort(:as(*.Str)).elems` would have Mu.pm6's `proto method Bool()` at the top by exclusive time?
lizmat AlexDaniel: saw it in a private mail, referring to github.com/rspier/rt-static 14:28
AlexDaniel lizmat: ok, so there will be a static version 14:29
lizmat well... fsvo version
MasterDuke: hmmm... it looks like *every* profile I do now, has gen/moar/stage2/NQPHLL.nqp:1825 taking about 50% of the timer 14:33
which is, afaics, the code doing the profile
timotimo: ^^^
MasterDuke lizmat: yep. that was not intentional
14:34 vrurg left
timotimo hmm, tbh i also don't know how confprog and that work together 14:44
MasterDuke timotimo: when did you merge confprog? 14:47
timotimo only to moarvm so far, and to nqp i guess? 14:49
let me have another look
ah. not to nqp yet 14:50
MasterDuke there was a rakudo bump yesterday i think
timotimo because moar hasn't been bumped
hadn't been bumped, to be exact
there we go
Geth nqp: 1bf35cbed2 | (Timo Paulssen)++ | 2 files
Revert "back out installconfprog until it's in moarvm master"

This reverts commit cdc9e773acbdab01e4e8f59bc4591806c25e4d2a.
MasterDuke ?
my tools/templates/MOAR_REVISION matches `git describe` in moarvm 14:51
timotimo it hadn't beem bumped the last time i looked :) 14:52
MasterDuke ah, you're bringing installconfprog back in
does that also need something in rakudo?
timotimo nope 14:53
only an nqp bump would be required to activate it 14:55
lizmat afk for several hours& 14:58
MasterDuke timotimo: fwiw, after rebuilding nqp and rakudo, that <anon> is still at the top of a profile 15:14
timotimo yes, if you don't activate confprog, nothing happens (hopefully) 15:33
MasterDuke timotimo: it must be something i did(n't) do in github.com/MoarVM/MoarVM/pull/1111 15:39
timotimo do we perhaps want to revert it for the release? 15:43
MasterDuke unless we can find a fix, yeah 15:45
maybe there is something i should do with tc->instance->instrumentation_level ?
timotimo release is already being pushed far past its intended release date :)
AlexDaniel and you can help with that a lot! 16:01
:)
16:07 vrurg joined 16:12 vrurg left 16:14 vrurg joined
[Tux] Rakudo version 2019.03.1-645-g0c8f960e8 - MoarVM version 2019.05-91-g81e5cbf2a
csv-ip5xs1.153 - 1.190
csv-ip5xs-2059.552 - 60.550
csv-parser23.423 - 23.520
csv-test-xs-200.437 - 0.455
test6.706 - 6.874
test-t1.722 - 1.728
test-t --race0.828 - 0.830
test-t-2027.576 - 29.315
test-t-20 --race8.957 - 9.465
16:16
16:58 jnthn left
vrurg .tell AlexDaniel looks like R#3012 might need conceptual consideration. I don't see a direct solution but to prohibit binding of exports whatsoever for now. 18:01
yoleaux vrurg: I'll pass your message to AlexDaniel.
synopsebot_ R#3012 [open]: github.com/rakudo/rakudo/issues/3012 [BLOCKER][regression] Some issue after “Decontainerize non-scalar symbols upon import” commit
AlexDaniel .
yoleaux 18:01Z <vrurg> AlexDaniel: looks like R#3012 might need conceptual consideration. I don't see a direct solution but to prohibit binding of exports whatsoever for now.
AlexDaniel vrurg: well… there's at least one module that is broken…
on the other hand, only one that we were able to detect 18:02
vrurg: is it greppable?
vrurg AlexDaniel: my decont commit is not related, the problem is reproducible on 2019.03. It just reveals what previously was a damn sheer luck.
Try 'is export .* :=' 18:03
AlexDaniel greppable6: is export .* := 18:04
greppable6 AlexDaniel, Found nothing!
vrurg I'm still trying to find a way to either apply binding before traits or make 'is export' a lazy thing.
Hm... 18:08
AlexDaniel Geth: ver github.com/rakudo/rakudo/commit/23...5bc0141031
Geth AlexDaniel, version bump brought in these changes: github.com/perl6/nqp/compare/2019....6-gf68c2ba
AlexDaniel Geth: ver github.com/perl6/nqp/commit/f68c2b...755c59b29b
Geth AlexDaniel, version bump brought in these changes: github.com/MoarVM/MoarVM/compare/2...7-g8fef199
AlexDaniel timotimo: in these commits, what could cause `elems requires a concrete object (got a NativeCall::Types::CArray[uint8] type object instead)` ? 18:09
timotimo well, is there a backtrace? 18:27
AlexDaniel timotimo: sorta: github.com/rakudo/rakudo/issues/3015
looking at the code, it's probably indeed a type object 18:30
timotimo where do i have to look? 18:33
AlexDaniel timotimo: github.com/rakudo/rakudo/blob/0c8f...s.pm6#L223 18:35
timotimo: github.com/salortiz/p6-LMDB/blob/3...DB.pm6#L98
timotimo: github.com/salortiz/p6-LMDB/blob/3...B.pm6#L452
vrurg AlexDaniel: if spectests are passed then I have the a for 3012. 19:05
timotimo yeah, that seems wrong 19:11
19:13 Kaiepi left 19:25 Kaiepi joined
Geth rakudo: vrurg++ created pull request #3016:
Implement more fine-grained deconting of exported symbols
19:27
rakudo: 0b9762642c | (Elizabeth Mattijsen)++ | 2 files
Revert "Don't allow junctions as keys for Hash/Map initializations"

This reverts commit 358d59fdae47edbe788fc8fea9afc27148896330.
This appears to be needing a little more thought on maybe a deprecation cycle / version dependency.
19:33
lizmat AlexDaniel: reverted the commit causing the regression, left the ticket open, removed the blocker label 19:35
AlexDaniel lizmat: alright 19:36
lizmat: thank you! Also, I just left a comment here: github.com/rakudo/rakudo/issues/30...-504692735
c: HEAD my @a = <foo bar baz>; my Seq $s = @a.Seq; say $s.first; @a[0] = 20; say $s.first
committable6 AlexDaniel, ¦HEAD(0c8f960): «foo␤20␤»
AlexDaniel c: 2019.03 my @a = <foo bar baz>; my Seq $s = @a.Seq; say $s.first; @a[0] = 20; say $s.first
committable6 AlexDaniel, ¦2019.03: «foo␤The iterator of this Seq is already in use/consumed by another Seq␤(you might solve this by adding .cache on usages of the Seq, or␤by assigning the Seq into an array)␤ in block <unit> at /tmp/EsZHTk2BvV line 1␤␤ «exit code = 1»»
lizmat AlexDaniel: not sure why you use .first instead of .head 19:45
bisectable6: my @a = <foo bar baz>; my Seq $s = @a.Seq; say $s.first; @a[0] = 20; say $s.first
bisectable6 lizmat, Bisecting by exit code (old=2015.12 new=0b97626). Old exit code: 1
AlexDaniel lizmat: because maybe, just maybe, the naming is a bit confusing and not exactly explicit 19:46
bisectable6 lizmat, bisect log: gist.github.com/87199c5433eeccb67a...c9441988d9
lizmat, (2019-06-12) github.com/rakudo/rakudo/commit/94...ee2650b018
AlexDaniel lizmat: :)
but yes, I meant head
greppable6: \.first;
greppable6 AlexDaniel, 68 lines, 36 modules: gist.github.com/9293e27affd6d158e8...0fa9584b02
AlexDaniel and 36 modules agree! 19:47
oh my xD
I also like this
greppable6: grep.*\.first;
greppable6 AlexDaniel, 18 lines, 13 modules: gist.github.com/b8cc1f7fdbce9163c0...b5a9467710
lizmat looks like me should make the distinction between .first / .head clearer 19:49
*we
AlexDaniel lizmat: is there any downside? 19:52
MasterDuke at least there isn't much of a performance difference
lizmat .grep(foo).first is the same as .first(foo)
AlexDaniel lizmat: yes yes, that's the point :)
m: say <foo bar baz>.head(3) 19:53
camelia (foo bar baz)
AlexDaniel m: say <foo bar baz>.head(2)
camelia (foo bar)
AlexDaniel m: say <foo bar baz>.first(2)
camelia Nil
AlexDaniel m: say <foo 2 baz>.first(2)
camelia 2
lizmat .grep(foo).head is *not* the same as .head(foo)
yoleaux lizmat: Sorry, that command (.grep(foo).head) crashed.
AlexDaniel shrugs 20:05
github.com/perl6/problem-solving/issues/48
lizmat AlexDaniel: In process of reverting all 3 commits re R#3014 and spectesting... 20:37
synopsebot_ R#3014 [open]: github.com/rakudo/rakudo/issues/3014 [BLOCKER][regression] Unwanted caching of Seqs
AlexDaniel lizmat++
Geth rakudo: 090f3f1b45 | (Elizabeth Mattijsen)++ | 4 files
Revert 7af0fb1 94ba19f 9107215

We need to look into this deeper.
20:39
vrurg lizmat: You have a 'Hallo' from Theo. We're sitting together in PIT 20:40
lizmat waves at Theo
Geth roast: d3fa62ae99 | (Elizabeth Mattijsen)++ | S02-types/hash.t
Revert "Add tests for R#2865"

This reverts commit 40c3f305d92cdb9e142323fcddb28a1d2bee2f1d.
We're allowing junctions again for the moment
20:41
synopsebot_ R#2865 [open]: github.com/rakudo/rakudo/issues/2865 All Junctions behave the same when used to define a hash key
lizmat AlexDaniel: I'm out of tuits for this evening... 20:48
21:24 vrurg left
Geth rakudo: 8652386974 | (Elizabeth Mattijsen)++ | src/core/Any-iterable-methods.pm6
Handle Junctions in Bool context within grep

Fixes R#2975
21:56
synopsebot_ R#2975 [open]: github.com/rakudo/rakudo/issues/2975 [BLOCKER][regression] Regression in grep with Junctions
AlexDaniel m: (when %()<>:exists { ... }) 21:58
camelia ===SORRY!===
internal problem: parser did not give circumfix an EXPR
AlexDaniel bisect: (when %()<>:exists { ... })
bisectable6 AlexDaniel, Bisecting by exit code (old=2015.12 new=090f3f1). Old exit code: 0
AlexDaniel, bisect log: gist.github.com/a9528bec096226d369...03a43654ab 21:59
AlexDaniel, (2018-05-22) github.com/rakudo/rakudo/commit/91...98dc273bec
lizmat AlexDaniel: really out of tuits now :-) 22:03
22:24 jnthn joined 22:27 jnthn left, jnthn joined 23:03 lucasb left
Geth ¦ problem-solving: AlexDaniel assigned to JJ Issue .head and .first github.com/perl6/problem-solving/issues/48 23:04