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 m: say "test" 10:26
camelia sudo: effective uid is not 0, is /usr/bin/sudo on a file system with the 'nosuid' option set or an NFS file system without root privileges? 10:27
nine m: say "test" 10:37
camelia test 10:38
lizmat m: say "hello world"
camelia hello world
lizmat yeah, probably was precomping
[Tux] Rakudo version 2019.03.1-134-g507c3c37e - MoarVM version 2019.03-61-g645a70c0d
csv-ip5xs0.739 - 0.773
csv-ip5xs-206.388 - 6.400
csv-parser22.387 - 22.860
csv-test-xs-200.436 - 0.439
test6.871 - 7.451
test-t1.759 - 1.834
test-t --race0.822 - 0.922
test-t-2028.813 - 30.070
test-t-20 --race9.651 - 9.856
 11:18
Geth rakudo: 47832fe1ca | (Elizabeth Mattijsen)++ | src/RESTRICTED.setting
Remove no longer existing internal sub wrappers
 11:44
lizmat afk& 11:45
bartolin in src/core/Buf.pm6 there are a few calls to nqp::getattr for a Range object that don't nqp::decont that object: github.com/rakudo/rakudo/blob/adba...f.pm6#L387 15:15
that doesn't work on the jvm backend (java.lang.RuntimeException: No such attribute '$!is-int' for this object)
timotimo good catch 15:16
i think oftentimes there'd be a variable that holds a deconted version of some object, but i'm not sure if that's still good for performance or if putting deconts everywhere is more efficient or something
bartolin so, there should be a nqp::decont, right? (like here: github.com/rakudo/rakudo/blob/adba....pm6#L693)
bartolin wonders what would be more efficient, too. 15:17
timotimo probably depends entirely upon how spesh deals with it
and the jit fater spesh, too
bartolin in this particular cases it would need five additional calls to nqp::decont. I'm reluctant to add those. (it would make the code work on the JVM backend, though :) 15:19
lizmat bartolin: please add the deconts 17:06
even more generally: if you find something doesn't work on the JVM backend because of a missing decont. please add it 17:07
bartolin thanks, will do. 17:42
bartolin I've also just found a missing (from the perspective of the JVM backend) null check for %!hash in src/core/Capture.pm6: github.com/rakudo/rakudo/blob/adba...re.pm6#L37 I think I'll add that, too. (We have similiar checks for null or definedness in that area of code.) 17:44
. o O ( there is always git revert )
Geth rakudo: a5b22387a3 | usev6++ | src/core/Buf.pm6
Decont Range before calling nqp::getattr on it

Makes the method work on the JVM backend. Calling the method did fail with a RuntimeException "No such attribute '6298is-int' for this object" before.		 18:49
bartolin oops, that $! was interpreted by the shell :)
Geth rakudo: f26fbd374a | usev6++ | src/core/Capture.pm6
Add null/defined checks for %!hash

The first check avoids a NullPointerException in S06-other/main-refactored.t on the JVM backend. The failing code was:
   sub MAIN("NEVER MATCHES") { 42 }		 19:30
hankache Rakudo Star 2019.03 has been released (tar.gz & msi) The dmg will follow soon 21:51
AlexDaniel
.oO( that could've been said right after the v6.d release… )
 23:40

Powered by Cro and the Raku® Programming Language.

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

Thank you!