wander afternoon #perl6-dev 04:45
could anyone explain this line at NQP/Grammar.nqp for me?
token infix:sym<||> { <sym> <O(|%tight_or, :op<unless>)> } 04:46
I cannot find its action, and think maybe something done by syntax <O(..)>, which I don't know what it means
it coverts a || b into 04:47
- QAST::Op(unless &infix:<||>) ||
- QAST::Var(lexical $bar)
- QAST::IVal(200)
oh, $bar || 200
but - QAST::Var(lexical $bar) is a condition, if we write down $foo = $bar || 200, some coercion needed might not be done 04:49
nqp: my $foo := 42; my int $bar := $foo || 200; say($bar) 04:51
camelia 0
Geth rakudo: 11d90cac59 | (Tom Browder)++ (committed by Samantha McVey) | 3 files
Allow escaping POD column sep. Fixes rakudo/rakudo#1282

Add code to fix rakudo issue #1282; allows escaping column separators to use as cell data; Tests to be added with github.com/perl6/roast/pull/361
synopsebot RAKUDO#1282 [closed]: github.com/rakudo/rakudo/issues/1282 [POD] need to handle table cells with char column separators as data
wander here, since the result_kind is native int, $foo should be coerced, but because it's condition, no coercion happens
that's why I think $bar || 200 should be 04:52
- QAST::Op(unless &infix:<||>) ||
- QAST::Var(lexical $bar) 04:53
- QAST::IVal(200)
but it might lose efficient, any advise?
wait 04:56
we *do* coercion, but it's not right. 04:58
Geth nqp/patch-1: 335ac1cf0b | (Alex Chen)++ | 2 files
Fix op_unless coercion

Sample code:
  ```
my $foo := 42; my int $bar := $foo || 200; say($bar); ... (14 more lines)
06:10
nqp: W4anD0eR96++ created pull request #383:
Fix op_unless coercion
06:12
synopsebot NQP#383 [open]: github.com/perl6/nqp/pull/383 Fix op_unless coercion
wander Last time I make a new nqp but run `nqp` instead of `./nqp`, thus I lost in "wtf" :\ 06:15
Does rakudo use the same LTM strategy as NQP? I means, do we implement it in NQP and them implement it again in rakudo? 06:36
Geth nqp/patch-1: 59be3324d4 | (Alex Chen)++ | 2 files
Resolve all 2-arg conditions
08:25
nqp/patch-1: b4fdfa7333 | (Alex Chen)++ | main.nqp
Remove debug file
08:41
lizmat Files=1229, Tests=75890, 317 wallclock secs (14.77 usr 5.35 sys + 2187.72 cusr 217.00 csys = 2424.84 CPU) 09:07
wander: sorry, can't answer that question 09:08
pmurias wander: the QAST is correct, it's the moarvm generating code that is wrong 09:16
wander: I have reviewed your PR, if something is unclear I can clarify 09:49
wander: thank you for finding the cause of the bug btw
tbrowder .tell samcv Thank you Sam I am! 10:07
yoleaux tbrowder: I'll pass your message to samcv.
Geth rakudo: eed733e277 | (Martin Barth)++ | src/core/IO/Path.pm
Fixed resolve of IO::Path for Windows

Pathes with Windows volumes do no longer start with a seperator character.
IO::Path.resolve: Added $volume to CWD
11:57
rakudo: f70e20bdf5 | (Zoffix Znet)++ (committed using GitHub Web editor) | src/core/IO/Path.pm
Merge pull request #1249 from ufobat/path_3rd

RFC && 3rd attempt: Fixed resolve of IO::Path for Windows
synopsebot RAKUDO#1249 [closed]: github.com/rakudo/rakudo/pull/1249 RFC && 3rd attempt: Fixed resolve of IO::Path for Windows
nqp/patch-1: 05eb3097a5 | (Alex Chen)++ | src/vm/moar/QAST/QASTOperationsMAST.nqp
Improve impl, pmurias++
12:40
wander pmurias: yes, I don't know our QAST can handle that then. 12:41
Thanks for your review, maybe my last commit do what you means :)
I'm not quite sure how our registers alloc, too. Try to work it out 12:42
Geth nqp: 68439403ec | pmurias++ | src/vm/js/Utils.nqp
[js] Remove extra whitespace
13:38
nqp: 8a06289378 | pmurias++ | src/vm/moar/QAST/QASTOperationsMAST.nqp
Fix wrong coercion in if on moar

my int $bar := $foo || 200; should now work correctly
nqp: 0a5e03a21e | pmurias++ | t/nqp/012-logical.t
Add test for fixed bug
pmurias wander: the bug was in a different place, I fixed that
wander oh. thank you 13:39
|Tux| Rakudo version 2017.11-76-gf70e20bdf - MoarVM version 2017.11-28-gba1fb9542
csv-ip5xs1.064 - 1.086
csv-ip5xs-2012.696 - 12.775
csv-parser11.977 - 11.979
csv-test-xs-200.426 - 0.450
test11.427 - 11.883
test-t2.838 - 2.891
test-t --race1.152 - 1.217
test-t-2051.526 - 52.518
test-t-20 --race17.425 - 17.660
14:12
Geth nqp: 00660af056 | (Alex Chen)++ | t/nqp/012-logical.t
Add test of logical ops

This is one bad case of an incorrect fixing of `||` logic. It says 0, which should be 200.
Add this test to prevent regressions from happening.
17:45
travis-ci NQP build passed. Alex Chen 'Add test of logical ops 18:05
travis-ci.org/perl6/nqp/builds/313080859 github.com/perl6/nqp/compare/0a5e0...660af056f0
AlexDaniel releasable6: next 22:09
releasable6 AlexDaniel, Next release in 8 days and ≈20 hours. No blockers. 0 out of 76 commits logged
AlexDaniel, Details: gist.github.com/b10bb6a66d4948ff03...31d82a9d4a