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.
ugexe i just realized if we hadn't already put json::fast code into rakudo we would have had to do *something* to for sorted-keys for reproducibility 02:38
s/to for/too for/ 02:39
nine I'm glad I didn't have to :) 06:35
brrt \o 07:24
tyil sorted-keys in JSON::Fast are neat tbh 11:30
hope YAMLish can get something like it soon-ish 11:31
lizmat commute to Amsterdam.PM& 12:47
|Tux| Rakudo version 2019.03.1-284-gb0dd44b00 - MoarVM version 2019.05
csv-ip5xs0.693 - 0.742
csv-ip5xs-205.794 - 5.871
csv-parser22.208 - 22.543
csv-test-xs-200.429 - 0.432
test6.564 - 6.786
test-t1.674 - 1.713
test-t --race0.781 - 0.808
test-t-2028.654 - 28.810
test-t-20 --race8.824 - 8.971
13:27
lizmat m: sub a($a,$b) { $a + $b }; for ^10_000_000 { a 42,666 }; say now - ENTER now 15:28
camelia 2.0588669
lizmat m: sub a(\a,\b) { a + b }; for ^10_000_000 { a 42,666 }; say now - ENTER now
camelia 0.314658
lizmat m: sub a($a,$b) { $a + $b }; for ^10_000_000 { a 42,666 }; say now - ENTER now
camelia 2.031367
lizmat m: sub a(\a,\b) { a + b }; for ^10_000_000 { a 42,666 }; say now - ENTER now
camelia 1.651616
lizmat m: sub a($a,$b) { $a + $b }; for ^10_000_000 { a 42,666 }; say now - ENTER now
camelia 2.01352
lizmat m: sub a(\a,\b) { a + b }; for ^10_000_000 { a 42,666 }; say now - ENTER now 15:29
camelia 1.56884157
lizmat the \a vs $a difference is caused by the scalar container that apparently is allocated for each parameter / call 15:30
for the $a case (and not for the \a case)
I seem to remember that this was not such a huge difference ? 15:31
did we lose an optimization along the way somehow? 15:32
discord6 <timotimo> The question is why doesn't scalar replacement make both identical? 16:24
<timotimo> Don't have a computer hooked up to look at it in spesh 16:50
timotimo ok my computer is back up 17:23
timotimo yeah, with the dollar it does create the scalar objects 17:37
Kaiepi is github.com/rakudo/rakudo/pull/2819 something that can be reviewed and merged before the next release? the roast tests and docs updates are already merged, so i don't want spec tests to be broken for too long 20:35
timotimo Kaiepi: you never need to check for Nil for return values 21:00
oh, ok, Seconds is used for a sub parameter too
but for the return value it can be dropped to become just Numeric
the code is a bit too long for my current attention levels to review properly 21:02