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.
nine .tell pmurias just waiting for after release 06:33
yoleaux nine: I'll pass your message to pmurias.
20 Aug 2018 01:41Z <AlexDaniel> nine: fwiw here's a module that perhaps wants to be fixed :) github.com/tony-o/perl6-module-does/issues/3
|Tux| Did something change in the install procedure of zef from git checkout? This line now fails: 08:43
rakudo/zef 513 > perl6 -Ilib bin/zef install --install-to=/pro/3gl/CPAN/rakudo/install
Rakudo version 2018.06-463-gf965e26aa - MoarVM version 2018.06-430-g02de10c42
csv-ip5xs0.983 - 1.013
csv-ip5xs-207.582 - 7.732
csv-parser23.200 - 23.990
csv-test-xs-200.433 - 0.470
test9.459 - 9.504
test-t2.035 - 2.080
test-t --race0.891 - 0.930
test-t-2036.467 - 37.176
test-t-20 --race11.653 - 11.849
 08:44
Yep, a new low 08:45
Please bring it under 2.000 so I can Twitter about it :) :) 08:49
perl6 -I. -Ilib bin/zef install --install-to=/pro/3gl/CPAN/rakudo/install . 08:50
-I. and a trailing .
samcv good * 08:57
yoleaux 20 Aug 2018 01:54Z <AlexDaniel> samcv: current status: colabti.org/irclogger/irclogger_log...08-20#l152
jnthn |Tux|: I have a whole branch of opts to merge after the next release that have helped various things and might well do it :) 09:44
Geth rakudo/master: 4 commits pushed by (Elizabeth Mattijsen)++ 10:12
Geth rakudo: f4f8cbb432 | (Elizabeth Mattijsen)++ | src/core/Iterable.pm6
Don't close over anonymous classes in Iterable
 10:30
rakudo: 558fa31e7d | (Aleks-Daniel Jakimenko-Aleksejev)++ | t/04-nativecall/02-simple-args.t
Skip test that fails with clang

It was already skipped on other platforms.
  See R#2124 github.com/rakudo/rakudo/issues/2124		 10:31
synopsebot_ R#2124 [open]: github.com/rakudo/rakudo/issues/2124 [⚠ blocker ⚠] Failure in t/04-nativecall/02-simple-args.t with clang
Geth rakudo: 3117cfd9b7 | (Elizabeth Mattijsen)++ | src/core/Any-iterable-methods.pm6
Don't close over anonymous classes in Any-iterable-methods
 10:50
travis-ci Rakudo build canceled. Aleks-Daniel Jakimenko-Aleksejev 'Skip test that fails with clang 10:51
travis-ci.org/rakudo/rakudo/builds/418633624 github.com/rakudo/rakudo/compare/f...8fa31e7d68
buggable [travis build above] ☠ Did not recognize some failures. Check results manually. 10:51
AlexDaniel this is not cool 10:52
so I was talking here and turns out my messages were not being sent lol 10:53
AlexDaniel timotimo: so I'm looking at github.com/rakudo/rakudo/pull/1868...-413386084 11:32
timotimo: why can't we just change the order in the PR?
timotimo: I mean here: github.com/rakudo/rakudo/pull/1868...53281R2292
AlexDaniel oh nevermind 11:33
AlexDaniel is being dumb
AlexDaniel timotimo: but why not change just this? github.com/rakudo/rakudo/blob/mast....pm6#L2309 11:43
like, does it have to be a separate exception? 11:46
maybe, I'm not sure
but if there's a proto without any multis, what's more important? That the arguments don't match the proto or that there are no multis 11:47
?
c: 2018.06,HEAD proto sub foo(Str) {*}; foo 42 11:48
committable6 AlexDaniel, gist.github.com/ffeda7383fd01bddf6...4d83329361
AlexDaniel I can't tell which one should come first, and that test is part of errata… 11:51
TimToady safely back home from vacation/TPC 15:13
timotimo \o/ 15:25
jnthn TimToady: Hope vacation was the appropriate amount of fun/relax :) 15:35
TimToady hopes so too :)
Geth rakudo/revert-1868-master: 397e230b12 | (Aleks-Daniel Jakimenko-Aleksejev)++ (committed using GitHub Web editor) | 2 files
Revert "Throws X::Multi::NoMatch if a proto routine is defined without multis."
 16:18
Geth rakudo: 4a44dcd571 | (Aleks-Daniel Jakimenko-Aleksejev)++ | 2 files
Revert "Throws X::Multi::NoMatch if a proto routine is defined without multis."

This reverts commit d0a8a7af0e231a5b31796890c9da390c68433f4d.
The commit breaks an errata spectest. The test can be tweaked or removed if there's a good justification for doing that, but at the moment none was presented, so I believe the PR was merged by mistake. The change can (and likely should) be reintroduced later, please follow R#2227 for further discussion and progress.
  github.com/rakudo/rakudo/issues/2227		 16:24
synopsebot_ R#2227 [open]: github.com/rakudo/rakudo/issues/2227 [⚠ blocker ⚠] t/spec/S32-exceptions/misc.rakudo.moar is failing
Geth nqp/decodelocaltime: cb7d342ca9 | (Paweł Murias)++ | src/vm/js/nqp-runtime/package-lock.json
[js] Remove package log with outdated vulnerable package
 19:39
nqp/decodelocaltime: 52e58abe06 | (Paweł Murias)++ | src/vm/moar/QAST/QASTOperationsMAST.nqp
[moar] Add a nqp::decodelocaltime op
Geth rakudo: 9cd0e9f2c2 | (Elizabeth Mattijsen)++ | 2 files
Don't close over anonymous classes in Seq

Also: move the actual class to Rakudo::Iterator, as we don't want to pollute the Setting outer lexical namespace further.
Also, now that this lives in Rakudo::Iterator, we should be able to codegen the call to Rakudo::Iterator.Gather directly, and remove sub GATHER. But that's for another time.		 20:05
Geth rakudo: 7d598988c1 | (Elizabeth Mattijsen)++ | src/core/Range.pm6
Don't close over anonymous classes in Range
 21:29
AlexDaniel oh wow that's a lot of errata failures… 21:41
Geth rakudo: 5262301f3d | (Elizabeth Mattijsen)++ | src/core/Buf.pm6
Don't close over anonymous classes in Buf
 22:06
rakudo: c0f713b2af | (Elizabeth Mattijsen)++ | src/core/Rakudo/Internals.pm6
Don't close over anonymous classes in Rakudo::Internals
AlexDaniel lizmat: what's the meaning of github.com/perl6/roast/commit/a291...55fca444be ? It says “todo” without a comment, so that is transformed to `todo();` which doesn't compile at all because there's no multi like that. A fix to that is `#?rakudo todo "some description here"`, but why is the TODO left there at all? 22:07
is that because you wanted older rakudos to pass these tests? Has anybody even checked that lately? :) 22:09
lizmat: then there's t/spec/S03-operators/bag.t that fails a bunch of tests. Basically it does ($baghash ∪ $baghash) and expects it to be Bag but on HEAD it's BagHash. I guess these tests can be fixed to check for Baggy instead? 22:16
lizmat: same goes for t/spec/S03-operators/mix.t and t/spec/S03-operators/set.t 22:17
jnthn m: class C { }; await start { for ^10_000 { C.^set_name('B') } } xx 4; 22:34
camelia (signal ABRT)*** Error in `/home/camelia/rakudo-m-inst-2/bin/moar': double free or corruption (fasttop): 0x00007f015c085940 ***
======= Backtrace: =========
/lib64/libc.so.6(+0x740ef)[0x7f016f0be0ef]
/lib64/libc.so.6(+0x79646)[0x7f016f0c3646]
…
AlexDaniel c: b243a966327^,b243a966327 class C { }; await start { for ^10_000 { C.^set_name('B') } } xx 4; 22:36
committable6 AlexDaniel, ¦b243a966327^: «» ¦b243a96: «free(): invalid pointer␤free(): invalid pointer␤ «exit signal = SIGABRT (6)»»
AlexDaniel (2016-03-15) github.com/rakudo/rakudo/commit/b2...9556d2177e
jnthn: so it bisects to that ↑, but what is that snippet? 22:37
jnthn So far as I can tell, it's always been like that
AlexDaniel: It's a "golf" (actually constructed based on ASAN output) of the zef SEGV 22:38
AlexDaniel jnthn: oh, ok. But what do you mean by saying that it has always been like that
6c: class C { }; await start { for ^10_000 { C.^set_name('B') } } xx 4;
jnthn ugexe++ managed to get it to make ASAN blow up and it gave a nicely precise stack trace
AlexDaniel ugexe++ indeed
committable6 AlexDaniel, gist.github.com/2516d6162568e4e04e...a348bec94c 22:39
AlexDaniel c: 2016.02,2016.02,2016.02,2016.02,2016.02,2016.02,2016.02,2016.02 class C { }; await start { for ^10_000 { C.^set_name('B') } } xx 4;
committable6 AlexDaniel, ¦2016.02,2016.02,2016.02,2016.02,2016.02,2016.02,2016.02: «» ¦2016.02: «Cannot invoke this object (REPR: Null)␤ in block <unit> at /tmp/IEzVUztEfR line 1␤␤ «exit code = 1»»
AlexDaniel c: 2016.02,2016.02,2016.02,2016.02,2016.02,2016.02,2016.02,2016.02 class C { }; await start { for ^10_0000 { C.^set_name('B') } } xx 4;
committable6 AlexDaniel, ¦2016.02,2016.02,2016.02,2016.02,2016.02,2016.02,2016.02,2016.02: «»
AlexDaniel c: 2016.02,2016.02,2016.02,2016.02,2016.02,2016.02,2016.02,2016.02 class C { }; await start { for ^10_0000 { C.^set_name('B') } } xx 4;
jnthn I think at one point it might have leaked copious amounts of memory instead.
committable6 AlexDaniel, ¦2016.02,2016.02,2016.02,2016.02,2016.02,2016.02,2016.02,2016.02: «»
AlexDaniel c: 2016.03,2016.03,2016.03,2016.03,2016.03,2016.03,2016.03,2016.03 class C { }; await start { for ^10_0000 { C.^set_name('B') } } xx 4; 22:40
committable6 AlexDaniel, gist.github.com/f2f08824ea0bdf8c11...a33b1a66de
lizmat AlexDaniel: re github.com/perl6/roast/commit/a291...55fca444be , what todo are you talking about?
AlexDaniel lizmat: the commit adds `#?rakudo todo`. When the test file is fudged, that todo is transformed into `todo();` 22:41
lizmat so a reason MUST be added ?
AlexDaniel lizmat: yes
lizmat: you can check that by checking out 6.c-errata in t/spec, and then running this `perl t/spec/fudgeandrun S02-literals/quoting.t`
lizmat ah... hmmm... somehow I had it in my mind that that was ok 22:42
AlexDaniel jnthn: so what's the meaning of this commit then: (2016-03-15) github.com/rakudo/rakudo/commit/b2...9556d2177e
jnthn AlexDaniel: fwiw, I got a MoarVM patch already for this; I was testing the above here to make sure my test case blows up prior to the fix
lizmat as the same is done for the niecza case
AlexDaniel ah alright
jnthn++ 22:43
lizmat: yeah, I've noticed that too. But maybe `todo` was different in niecza? I don't know.
jnthn That patch was just moving the setting of the debug name to apply to all types, not just classes
AlexDaniel jnthn: by the way, have you seen the fix for R#2231 ? 22:44
synopsebot_ R#2231 [open]: github.com/rakudo/rakudo/issues/2231 [SEGV][regression][⚠ blocker ⚠] SEGV in Audio::Sndfile
AlexDaniel I mean, the way the module was fixed
jnthn ah no, not yet 22:46
I saw that the author was going to look 22:47
Then went for dinner, homework, etc. :)
Hmm...interesting 22:55
I wonder if that's "valgrind is happy but it leaks" 22:56
Geth roast/6.c-errata: 6d8d40eb02 | (Zoffix Znet)++ (committed by Aleks-Daniel Jakimenko-Aleksejev) | S04-declarations/constant.t
[CaR Grant] Fix incorrect typed constant test

The test declares a constant of type `Num` but tries to assign a value of type `Int` to it. The only reason this test didn't explode in the past was because type constraints on constants weren't plugged into anything and were entirely ignored.
Fix test by using correct type of the value. Extra 6.d tests in S04-declarations/constant-6.d.t (master branch) also test the typecheck failure when types do not match.
nqp: f7c733dcb2 | (Jonathan Worthington)++ | tools/build/MOAR_REVISION
Bump MOAR_REVISION for setdebugtypename fix
 23:01
¦ nqp: version bump brought these changes: github.com/MoarVM/MoarVM/compare/2...1-g919e78a
rakudo: 562f60c2cf | (Jonathan Worthington)++ | tools/build/NQP_REVISION
Bump for MoarVM with setdebugtypename fix
rakudo: version bump brought these changes: github.com/perl6/nqp/compare/2018....5-gf7c733d
0ec0de6a3a | (Jonathan Worthington)++ | t/09-moar/00-misc.t
Geth roast/6.c-errata: ac36728c76 | (Samantha McVey)++ (committed by Aleks-Daniel Jakimenko-Aleksejev) | S32-hash/delete-adverb.t
Fix S32-hash/delete-adverb.t if hash keys iter randomized

The implementation is allowed to randomize the order of the iterator. You can't rely on the order unless you don't modify the hash, in that case order is consistent.		 23:11
Geth rakudo: 37a30ab9bf | (Elizabeth Mattijsen)++ | src/core/Rakudo/Iterator.pm6
Don't close over anonymous classes in Rakudo::Iterator, part 1
 23:32
lizmat and that concludes my hacking for today& 23:36

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!