[Coke] AlexDaniel++ thanks. in the middle of a different docuthing at the moment, but will circle back to that when I can 00:26
Geth 6.d-prep: ce20847d94 | (Samantha McVey)++ | 2 files
Update status of Ornate Parenthesis Removal

This has been resolved now.		 00:52
6.d-prep: 3a8c4587fe | (Samantha McVey)++ | TODO/FEATURES.md
Add proposal for strict encode/decode of one byte encodings
tbrowder .tell AlexDaniel see my updated comment in nqp PR #401. i need expert help to get the nqp::spawnprocasync code in it working. if that can work, then we can have a good test. 01:36
yoleaux tbrowder: I'll pass your message to AlexDaniel.
Geth rakudo: b3689e63e0 | (Zoffix Znet)++ | t/08-performance/02-qast-rewrites.t
Cover a couple of `for ...` opts
 02:28
rakudo: 59556c70b4 | (Zoffix Znet)++ | src/Perl6/Actions.nqp
Addititonal fixes to slurpies on `if`

For RT#105872: rt.perl.org/Ticket/Display.html?id=105872
In the original fix[^1] we bail out of $count counting if we encounter a named arg or named slurpy. However, these seem to be allowed before named slurpies, so the `if 42 -> *%, *@a {}` ... (5 more lines)		 02:33
synopsebot RT#105872 [resolved]: rt.perl.org/Ticket/Display.html?id=105872 [BUG] 42, 44, 22 -> *@a { say @a.perl }
roast: 9084d7be1a | (Zoffix Znet)++ | S04-statements/if.t
Cover more cases of `if` slurpies

Rakudo fix: github.com/rakudo/rakudo/commit/59556c70b4		 02:34
rakudo: c6b1880b4c | (Zoffix Znet)++ | 8 files
Use saner filenames for Unicode tests

Closes github.com/rakudo/rakudo/issues/1511
Although, previous commit[^1] remove OS-restricted characters, the `;` is still a special shell char and makes quoting required for these filenames, which I see no reason to require.
  [1] github.com/rakudo/rakudo/commit/80...8a3a7adec6		 03:07
Zoffix dogbert17: is the `m: say "ab" ~~ /^ a ?% (b) $/` thing from some ticket? 03:12
I'm assuming it should throw about `?` quantifier being useless on `%`
RT#125618 03:14
synopsebot RT#125618 [new]: rt.perl.org/Ticket/Display.html?id=125618 [REGEX] Positional capture of separator on ?% operator causes compiler error
Geth nqp: a3f4410709 | (Tom Browder)++ | src/NQP/Grammar.nqp
improve pod panic msg
 03:18
nqp: 0292b5a0d5 | (Aleks-Daniel Jakimenko-Aleksejev)++ (committed using GitHub Web editor) | src/NQP/Grammar.nqp
Merge pull request #401 from tbrowder/mod-pod-panic-msg

improve pod panic msg
Zoffix .ask moritz Currently this crashes `m: say "ab" ~~ /^ a ?% (b) $/`. I was thinking instead of trying to fix it, we should just throw some "Useless use of ? with %" if the `%` thing is used with `??` or `? quantifiers. Does that make sense? 03:23
yoleaux Zoffix: I'll pass your message to moritz.
Zoffix .tell tbrowder just use the asyncspawn I showed you the other week. You can stick it as a sub here and use it in tests: github.com/perl6/nqp/blob/master/s...esting.nqp 03:31
yoleaux Zoffix: I'll pass your message to tbrowder.
Zoffix gosh why is our sprintf code so complex 04:09
And looks like the impl is all over the shop. A `"%d"` can take any numeric but `"%f"` can't take an Int 04:17
m: dd sprintf "%d", 2i+0 04:20
camelia Directive d not applicable for type Complex
Zoffix Surely it *is* applicable:
m: dd sprintf "%d", <0+2i>
camelia Directive d not applicable for type Complex
Zoffix m: dd sprintf "%d", <2+0i>
camelia "2"
Zoffix m: dd sprintf "%s", class {}.new 04:22
camelia "<anon|65689024><66496368>"
Geth roast: c37c7010c8 | (Zoffix Znet)++ | S32-exceptions/misc.t
Fix overly-specific tests

They're not part of 6.c.
It's probably overly specific to spec that exceptions must have a precise number of bracktrace frames up in 'em.		 06:39
nqp: 51bc089f03 | (Zoffix Znet)++ | src/HLL/sprintf.nqp
Remove unused variable
 06:40
nqp: 2c51764e93 | (Zoffix Znet)++ | src/HLL/sprintf.nqp
Fix sprintf with type objects

Fixes RT#132846: rt.perl.org/Ticket/Display.html?id=132846
  - For %c, use the same intify() routine we use for other ints
  - For floats, add a floatify routine which will go to
   SprintfHandler and ask it to handle floats. It will leave
   Numeric:Ds as is, but everything else, it'll try to .Numeric		 06:42
synopsebot RT#132846 [new]: rt.perl.org/Ticket/Display.html?id=132846 [Bug] Directive d not applicable for type Int
rakudo: d419afe462 | (Zoffix Znet)++ | tools/build/NQP_REVISION
[NQP Bump] Brings 4 commits

NQP bump brought: github.com/perl6/nqp/compare/2018....0-g2c51764 2c51764 Fix sprintf with type objects 51bc089 Remove unused variable 0292b5a Merge pull request #401 from tbrowder/mod-pod-panic-msg a3f4410 improve pod panic msg		 06:43
rakudo: version bump brought these changes: github.com/perl6/nqp/compare/2018....0-g2c51764
4ac67e7384 | (Zoffix Znet)++ | src/core/Rakudo/Internals.pm

Fixes RT#132846: rt.perl.org/Ticket/Display.html?id=132846
  - Make :Us go through .Numeric first, since that's the method that
   does the whole "undef is 0" thing, whereas .Int just coercers
   to Int and it leaves our Int:U as Int:U
  - Implement float handler: return Numeric:D as-is, for the rest,
   return result of .Numeric
rakudo: 684b99eaa8 | (Zoffix Znet)++ | src/core/Rakudo/Internals.pm
Fix potential issue in Sprintf exception handler

First, in the code we expect to throw the return value of R::I.HANDLE-NQP-SPRINTF-ERRORS, but it already throws. Second, if the payload doesn't mention any of the keys, we'll end up returning an `Empty` and trying to call `.throw` on it.
Fix by having R::I.HNSE return an Exception object: either the updated sprintf exception or the original one, and then we'll throw it at the point of call.		 06:47
roast: 81f0a53452 | (Zoffix Znet)++ | S32-exceptions/misc.t
Revert "Fix overly-specific tests"

This reverts commit c37c7010c8651322baaecfdd93997c6e5ddf6829. Or maybe they're fine...		 07:15
travis-ci Rakudo build failed. Zoffix Znet '[NQP Bump] Brings 4 commits 07:20
travis-ci.org/rakudo/rakudo/builds/340056411 github.com/rakudo/rakudo/compare/c...19afe46206
buggable [travis build above] ☠ Did not recognize some failures. Check results manually.
Geth rakudo: 13406517ad | (Zoffix Znet)++ | 2 files
Fix regression in backtrace raker

Fixes RT#132846: rt.perl.org/Ticket/Display.html?id=132846
When we encounter NQP guts in backtraces, we switch into trying to see whether we got any more HLL frames afterwards (from things like `sprintf` errors). Currently, we look for `.setting` annotation file ending, but awhile back we switched these to start with
  `SETTING::` instead. Fix by looking for that instead.
synopsebot RT#132846 [new]: rt.perl.org/Ticket/Display.html?id=132846 [Bug] Directive d not applicable for type Int
roast: ceed491adb | (Zoffix Znet)++ | S32-str/sprintf.t
Test sprintf handling of Str/Numeric type objects

RT#132846: rt.perl.org/Ticket/Display.html?id=132846 Rakudo fix: github.com/perl6/nqp/commit/2c51764e93
   github.com/rakudo/rakudo/commit/d419afe462
   github.com/rakudo/rakudo/commit/4ac67e7384
   github.com/rakudo/rakudo/commit/684b99eaa8
   github.com/rakudo/rakudo/commit/13406517ad		 07:23
moritz ZofBot: +1 on outlawing the ? regex quantifier in combination with % for a delimiter 08:06
yoleaux 03:23Z <Zoffix> moritz: Currently this crashes `m: say "ab" ~~ /^ a ?% (b) $/`. I was thinking instead of trying to fix it, we should just throw some "Useless use of ? with %" if the `%` thing is used with `??` or `? quantifiers. Does that make sense?
ZofBot moritz, ORG #SUPERBOWL IS HAVING A SUPERBOWL WATCH PARTY ON FEB
moritz how *fascinating* 08:07
Geth rakudo: 57a1aa7a87 | (Zoffix Znet)++ | src/Perl6/Actions.nqp
Fix unwanted curries with HyperWhatever type object

Fixes RT#128880: rt.perl.org/Ticket/Display.html?id=128880 Phixes R#1489: github.com/rakudo/rakudo/issues/1489
When we're trying to figure out whether to make a curry, we forget to check whether the HyperWhatever we encountered is an instance. Fix by adding that check.		 08:20
synopsebot RT#128880 [open]: rt.perl.org/Ticket/Display.html?id=128880 [BUG] Dead code in HyperWhatever, ｢.new｣ is not supposed to work but it does (HyperWhatever.new)
R#1489 [open]: github.com/rakudo/rakudo/issues/1489 Smartmatch with HyperWhatever:U RHS curries instead of type-matching
roast: dbe43e171d | (Zoffix Znet)++ | S02-types/hyperwhatever.t
Test we don't curry HyperWhatever:Us

RT#128880: rt.perl.org/Ticket/Display.html?id=128880 R#1489: github.com/rakudo/rakudo/issues/1489 Rakudo fix: github.com/rakudo/rakudo/commit/57a1aa7a87		 08:22
rakudo: 835bb35081 | (Zoffix Znet)++ | t/spectest.data
Add S02-types/hyperwhatever.t to list of tests to run

Closes github.com/rakudo/rakudo/issues/1489
travis-ci Rakudo build passed. Zoffix Znet 'Fix regression in backtrace raker 08:41
travis-ci.org/rakudo/rakudo/builds/340061853 github.com/rakudo/rakudo/compare/6...406517ad0f
Zoffix moritz: heh, so I made the fix but there's a single failing 6.c spec file.... Guess who committed it? :) github.com/perl6/roast/commit/aaa7...8121fabR16 09:56
Just gonna shove the stuff into a branch and let AlexDaniel`++ figure out whether tests can be changed or not. 09:57
Geth nqp/limit-quant-on-sep-RT125618: ac3d5307f3 | (Zoffix Znet)++ | src/QRegex/P6Regex/Actions.nqp
Make `?` and `??` quantifiers on `%`/`%%` throw…

  …a message that points out they're useless and suggests the user
may wanted to quote the constructs instead.
Fixes RT#125618: rt.perl.org/Ticket/Display.html?id=125618		 09:59
synopsebot RT#125618 [new]: rt.perl.org/Ticket/Display.html?id=125618 [REGEX] Positional capture of separator on ?% operator causes compiler error
Geth roast/limit-quant-on-sep-RT125618: f2b07afbba | (Zoffix Znet)++ | S05-metasyntax/regex.t
Test `?` and `??` quantifiers on `%`/`%%` throw

RT#125618: rt.perl.org/Ticket/Display.html?id=125618 NQP Fix: github.com/perl6/nqp/commit/ac3d5307f3		 10:00
moritz Zoffix: heh
travis-ci Rakudo build failed. Zoffix Znet 'Add S02-types/hyperwhatever.t to list of tests to run 10:01
travis-ci.org/rakudo/rakudo/builds/340069147 github.com/rakudo/rakudo/compare/5...5bb350815b
buggable [travis build above] ☠ All failures are due to: failed make test (1 failure). Across all jobs, only t/02-rakudo/repl.t test file failed.
Zoffix .tell FWIW, there's a potential fix for RT#125618 in branches in github.com/perl6/nqp/commit/ac3d5307f3 and github.com/perl6/roast/commit/f2b07afbba It's blocked by 2 lines of 6.c tests that indirectly use this combination: github.com/perl6/roast/commit/aaa7...8121fabR16 leaving it up to you whether to merge it or to 10:02
yoleaux Zoffix: What kind of a name is "FWIW,"?!
Zoffix delete the branches.
.tell AlexDaniel FWIW, there's a potential fix for RT#125618 in branches in github.com/perl6/nqp/commit/ac3d5307f3 and github.com/perl6/roast/commit/f2b07afbba It's blocked by 2 lines of 6.c tests that indirectly use this combination: github.com/perl6/roast/commit/aaa7...8121fabR16 leaving it up to you whether to merge
yoleaux Zoffix: I'll pass your message to AlexDaniel.
Zoffix it or to delete the branches.
.tell AlexDaniel it or to delete the branches.
yoleaux Zoffix: I'll pass your message to AlexDaniel.
[Tux] Rakudo version 2018.01-183-g835bb3508 - MoarVM version 2018.01-82-g296620e86
csv-ip5xs0.809 - 0.864
csv-ip5xs-207.575 - 8.046
csv-parser11.973 - 12.373
csv-test-xs-200.428 - 0.433
test9.209 - 9.533
test-t2.648 - 2.656
test-t --race1.085 - 1.123
test-t-2047.964 - 50.267
test-t-20 --race18.450 - 18.532
 10:48
but I see a lot of fluctuation
Geth rakudo: 4b7113fa06 | (Zoffix Znet)++ | src/Perl6/Actions.nqp
Fix thinko from earlier commit

The earlier fix[^1] was meant to continue looking for arifying args after nameds, but instead it tossed the entire line, causing named being counted for arity.
  [1] github.com/rakudo/rakudo/commit/59...09c6d70eba		 12:15
rakudo: 00af9ce272 | (Zoffix Znet)++ | t/08-performance/02-qast-rewrites.t
Cover bug in QAST::Block arity calculation

Covers github.com/rakudo/rakudo/commit/4b7113fa06
AlexDaniel releasable6: next 12:41
yoleaux 10:02Z <Zoffix> AlexDaniel: FWIW, there's a potential fix for RT#125618 in branches in github.com/perl6/nqp/commit/ac3d5307f3 and github.com/perl6/roast/commit/f2b07afbba It's blocked by 2 lines of 6.c tests that indirectly use this combination: github.com/perl6/roast/commit/aaa7...8121fabR16 leaving it up to you whether to merge
10:02Z <Zoffix> AlexDaniel: it or to delete the branches.
releasable6 AlexDaniel, Next release in 6 days and ≈6 hours. Blockers: github.com/rakudo/rakudo/issues?q=...%9A%A0%22. Changelog for this release was not started yet
synopsebot RT#125618 [open]: rt.perl.org/Ticket/Display.html?id=125618 [REGEX] Positional capture of separator on ?% operator causes compiler error
releasable6 AlexDaniel, Details: gist.github.com/22cf7587f4d6415a03...5c6c2ecf56
AlexDaniel let's see… 12:43
well, it's not useless with %%, as far as I can see 12:53
m: say ‘af’|‘a’|‘f’ ~~ /a? %% f/ 12:54
camelia ｢af｣
｢a｣
｢f｣
AlexDaniel the equivalent to that is just a?f? though 12:55
m: say ‘af’|‘a’|‘f’|‘’ ~~ /a? %% f/
camelia ｢af｣
｢a｣
｢f｣
｢｣
AlexDaniel m: say ‘af’|‘a’|‘f’|‘’ ~~ /a?f?/
camelia ｢af｣
｢a｣
｢f｣
｢｣
AlexDaniel hmm
travis-ci Rakudo build failed. Zoffix Znet 'Fix thinko from earlier commit 13:02
travis-ci.org/rakudo/rakudo/builds/340111378 github.com/rakudo/rakudo/compare/8...7113fa065f
buggable [travis build above] ☠ All failures are due to: failed make test (1 failure). Across all jobs, only t/04-nativecall/09-nativecast.t test file failed.
Zoffix AlexDaniel: then just let's delete those branches. 13:17
I didn't realize there was any difference that %% brought in those cases, but there is.
m: say ‘af’|‘a’|‘f’ ~~ /a? %% f/ 13:18
camelia ｢af｣
｢a｣
｢f｣
Zoffix In this one
AlexDaniel okay
Zoffix + save a couple of CPU ticks with not having to check whether this used was used.
AlexDaniel Zoffix: done, and left a comment on that ticket. FWIW links to these commits will work for a while, so deleting these branches doesn't really wipe out the information completely 13:23
(which is good) 13:24
Zoffix cool 13:29
AlexDaniel so using prebuilt rakudo for doc tests only made it faster by 6 minutes, that's not a big difference really… 13:31
like 25% faster, not really groundbreaking 13:32
dogbert17 Zoffix: yeah, it was ticket RT #125618 13:41
synopsebot RT#125618 [open]: rt.perl.org/Ticket/Display.html?id=125618 [REGEX] Positional capture of separator on ?% operator causes compiler error
travis-ci Rakudo build passed. Zoffix Znet 'Cover bug in QAST::Block arity calculation 13:47
travis-ci.org/rakudo/rakudo/builds/340111506 github.com/rakudo/rakudo/compare/4...af9ce272fc
AlexDaniel MasterDuke: here? 15:20
MasterDuke for maybe 5 min 15:21
AlexDaniel ah well, nvm then. I have an interesting example that leaks 15:22
# use Telemetry; use nqp; loop { nqp::force_gc(); say T<max-rss> }
but IIRC Telemetry itself shells out to get max-rss, or something like that 15:23
so maybe that's just the Proc::Async doing its leaky job
or maybe I'm mistaking it with Kernel stuff 15:24
lizmat has come down bit a bad cold and doesn't really feel up to doing the P6W already 15:25
AlexDaniel weekly: (Juan Julián Merelo Guervós)++ worked hard on the doc repo (a lot of old issues updated or resolved, faster travis test times) 15:33
notable6 AlexDaniel, Noted!
AlexDaniel notable6: weekly 15:34
notable6 AlexDaniel, 4 notes: gist.github.com/8db62cab3d39a58254...a9550ecdc9
AlexDaniel weekly: Camelia in the wild (check out the top of the Marshall amp on the right hand side): www.facebook.com/139508053330390/p...mp;theater 15:37
notable6 AlexDaniel, Noted!
lizmat AlexDaniel: Telemetry doesn't shell out for max-rss, just depends on what nqp::getrusage gives 16:05
AlexDaniel ah k
I was mistaking it with this I think: github.com/rakudo/rakudo/blob/mast...nel.pm#L14 16:06
lizmat yup, that does :-) 16:07
AlexDaniel anyway, I created a ticket for that: github.com/rakudo/rakudo/issues/1513 16:43
dogbert17 bisect: await do for ^10 -> $i { start { "A".match: /(“A”) { say "BOOM: $i) ", $0 if not $0 }/ } } 19:42
bisectable6 dogbert17, On both starting points (old=2015.12 new=00af9ce) the exit code is 0 and the output is identical as well
dogbert17, Output on both points: «»
dogbert17 bisect: await do for ^1000 -> $i { start { "A".match: /(“A”) { say "BOOM: $i) ", $0 if not $0 }/ } } 19:43
bisectable6 dogbert17, On both starting points (old=2015.12 new=00af9ce) the exit code is 0 and the output is identical as well
dogbert17, Output on both points: «»
AlexDaniel 6c: await do for ^1000 -> $i { start { "A".match: /(“A”) { say "BOOM: $i) ", $0 if not $0 }/ } } 19:47
committable6 AlexDaniel, gist.github.com/8ee07c502b5eb301bf...da68f6778f 19:48
AlexDaniel dogbert17: ↑ maybe this gives a better picture
dogbert17 yes, still looks broken 19:49
samcv i'm researching about AutoFDO which could be a way to make MoarVM faster. basically it creates a profile when you run your code (using CPU branch result instructions to return the result of branches) and then you compile it again with this collected data 20:10
and it can optimize based on how code actually executes, rather than just static code analysis
AlexDaniel MasterDuke++ # github.com/sergot/http-useragent/pull/187 21:06
1500MiB → 219MiB maxrss is a noticeable improvement I think! :)
jnthn Not announced yet, but today I got news that news.perlfoundation.org/2018/01/gra...l-6-1.html is approved \o/ 21:08
I...might just spend some of the time on hunting memory leaks ;)
MasterDuke ++jnthn 21:09
AlexDaniel fully supports that idea :) 21:10
benchable6: uptime
benchable6 AlexDaniel, 23 hours, 22 minutes, and 32 seconds, 1140.160156MiB maxrss. This is Rakudo version 2018.01-174-gef1d22f4c built on MoarVM version 2018.01-82-g296620e86 implementing Perl 6.c.
samcv jnthn++ \o/ 21:11
AlexDaniel say 1140 × 16
evalable6 18240
AlexDaniel does not have that much memory on his server 21:12
Geth nqp/master: 5 commits pushed by pmurias++ 21:20
japhb Congratulations ++jnthn!! 21:22
tbrowder Well deserved ++jnthn! 22:51
yoleaux 03:31Z <Zoffix> tbrowder: just use the asyncspawn I showed you the other week. You can stick it as a sub here and use it in tests: github.com/perl6/nqp/blob/master/s...esting.nqp
tbrowder Zoffix: i’ll try it, but i have been studying the moar c code it touches and may have lots of questions—be forewarned! 22:53
AlexDaniel reportable6: 2018-02-05T00:00:00Z 2018-02-12T00:00:00Z 23:04
reportable6 AlexDaniel, Couldn't find a snapshot for 2018-02-12T00:00:00Z (try ｢list｣ command to see what's available).
AlexDaniel ah… we're not there *yet*!
dogbert17 Congrats ++jnthn 23:21

Powered by Cro and the Raku® Programming Language.

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

Thank you!