jnthn samcv: Depends if anything in the bootstrapped version depends on it, I guess... 00:01
samcv: I mean, it's a lot cleaner to add a new op
A lot less disruptive at least
Note that for the streaming decoder, we pass in a config hash
So for that we don't need to do anything
Geth rakudo: 4a989a5e4f | (Samantha McVey)++ | tools/build/NQP_REVISION
Bump NQP: MVM revert throw when windows1252 errors. JVM add 1251

Changes: 2018.01-33-g99493b0..2018.01-44-ga860fd310 a860fd310 Bump MoarVM to revert throwing on windows-1252/1 decode errors d32d84d4b [jvm] Add windows-1251 as an available encoding b348d45cf [js] Check for known encoding in codecs 273850777 [js] Update dep ... (7 more lines)		 00:03
rakudo: version bump brought these changes: github.com/perl6/nqp/compare/2018....ga860fd310
f473216452 | (Zoffix Znet)++ | src/Perl6/Actions.nqp

We actually go through whatever curry on each curryeable thing like a method call or some op. Avoiding the dynvar and caching the lookup of Whatever, WhateverCode, and HyperWhatever in lexical vars comes out about 10x faster.
Makes compilation of 3000-line file ~6% faster. Not a lot but... Whatever
travis-ci NQP build failed. Samantha McVey 'Bump MoarVM to revert throwing on windows-1252/1 decode errors 00:16
travis-ci.org/perl6/nqp/builds/336782862 github.com/perl6/nqp/compare/d32d8...60fd3109e7
Rakudo build failed. Samantha McVey 'Bump NQP: MVM revert throw when windows1252 errors. JVM add 1251 00:41
travis-ci.org/rakudo/rakudo/builds/336784134 github.com/rakudo/rakudo/compare/9...989a5e4f5b
buggable [travis build above] ā˜  Did not recognize some failures. Check results manually.
MasterDuke is there a mic drop emoji proposed for Unicode 10? 00:42
Geth rakudo: b93656d82d | (Zoffix Znet)++ | src/Perl6/Actions.nqp
Revert "Elevate statevars to enclosing block when wrapped"

This reverts commit 299e85264b50104ad82411e9c3999dc727154091.
Travis is unhappy. Try to figure out what it is that makes it blow up.		 00:51
rakudo: 42af2a35b6 | (Zoffix Znet)++ | src/Perl6/Actions.nqp
Revert "Revert "Elevate statevars to enclosing block when wrapped""

This reverts commit b93656d82d81a5f93780f11bed055114d2095fc7.
Taht ain't it		 01:06
rakudo: 01ce5632bc | (Zoffix Znet)++ | src/core/VM.pm
Revert "Use a more descriptive desc for $*VM"

This reverts commit 86e926c76d6b3c47a2a9ffa11d43e34744493ac7.
Travis fix attempt #2		 01:07
rakudo: 07f30e4bab | (Zoffix Znet)++ | src/core/VM.pm
Revert "Revert "Use a more descriptive desc for $*VM""

This reverts commit 01ce5632bc9a4359870fa9c68347b9217d949995.
That ain't it either. I'm stumped :|		 01:16
Zoffix No clue why Travis fails compilation. I tried reverting all the commits around the area where it started failing, but even with reverts it's failing :/ travis-ci.org/rakudo/rakudo/builds 01:17
Geth rakudo: 7d42e33637 | (Zoffix Znet)++ | src/core/Exception.pm
Try to make travis error say what's actually wrong
 01:20
rakudo: 4929651929 | (Zoffix Znet)++ | src/core/Exception.pm
Coaxe travis some more
 01:34
Zoffix ZofBot: we'll get h' 01:35
ZofBot Zoffix, Though there had been times during the game when he had howled, for the most part he had watched in silence so hungrily tense that a less experienced observer than Mr Birdsey might have attributed his immobility to boredom
Zoffix So Travis is trolling us just 'cause he's bored? That's just mean.
geekosaur <Grunt> Travis can't be bargained with. Travis can't be reasoned with. Travis doesn't feel pity, or remorse, or fear. And Travis absolutely will not stop, ever, until your project is dead. 01:37
Zoffix :)
yey, progress. "Associative indexing implementation missing from type Hash" 01:42
wonder what's that about
MasterDuke src/Perl6/World.nqp:3864 is something about BEGIN phasers. could it be nine's recent $*PERL commit? 01:47
Zoffix There's a day worth of passing travises since then. But I don't got a better idea and I do see a PROCESS::<PERL> in it that's related to associative indexing 01:50
Geth rakudo: 85ae1d1ee6 | (Zoffix Znet)++ | 5 files
Revert "Initialize $*PERL during setting compilation to avoid runtime cost"

This reverts commit 671c3d60161053809415a2eb85097b0714417b56.
Try to unbust travis		 01:51
MasterDuke looks like it's passing 02:00
Zoffix oh yeah. MasterDuke++
Geth rakudo: 2339d8dd00 | (Zoffix Znet)++ | src/core/Exception.pm
Travis unbusted; restore X::Comp::BeginTime to former glory
 02:01
Zoffix .tell nine FYI we had to revert your $*PERL modifications in github.com/rakudo/rakudo/commit/85ae1d1ee6 For some reason a day after your commit travis started failing to compile (no failures in my local builds). After some modifications to X::Comp::BeginTime, we managed to coax the failure out of it and it says it's about associative indexing, perhaps due to PROCESS::<PERL> thing. Here's one of the 02:03
yoleaux Zoffix: I'll pass your message to nine.
Zoffix failures: travis-ci.org/rakudo/rakudo/jobs/3...2798#L1449
.tell nine failures: travis-ci.org/rakudo/rakudo/jobs/3...2798#L1449
yoleaux Zoffix: I'll pass your message to nine.
Zoffix m: 1 < *.flip < 52 02:28
camelia ===SORRY!===
QAST::Block with cuid 1 has not appeared
MasterDuke that's one sneaky block 02:29
Zoffix IIRC TimToady said ^ fixing that would need to be doing something in EXPR parser, but looking at QAST going through whatever_curry, I think it's possible to fix it up right there. Gonna try that
rewriting it to `my $temp := $arg.flip; 1 < $arg < 52` 02:30
but maybe something smarter is also possible
oh wait, no I think just 1 < $arg.flip < 52 is good. 02:31
Zoffix grabs more coffee and tries 02:32
travis-ci Rakudo build passed. Zoffix Znet 'Revert "Initialize $*PERL during setting compilation to avoid runtime cost" 02:33
travis-ci.org/rakudo/rakudo/builds/336805568 github.com/rakudo/rakudo/compare/4...ae1d1ee682
Zoffix m: 1 < *.flip < 52 < *.flip < 90
camelia ===SORRY!===
QAST::Block with cuid 1 has not appeared
Zoffix hehe that's fun
travis-ci Rakudo build passed. Zoffix Znet 'Travis unbusted; restore X::Comp::BeginTime to former glory' 03:05
travis-ci.org/rakudo/rakudo/builds/336807094 github.com/rakudo/rakudo/compare/8...39d8dd0091
Geth rakudo: 2e65de2999 | (Zoffix Znet)++ | src/core/Rakudo/Sorting.pm
Fix broken 2-arity sort of 2-el list

Phixes github.com/rakudo/rakudo/issues/1472
When we have a 2-el list, we need to swap only if Order::More is received from the comparator, but the test was checking whether to leave it alone if Order::Less is received.
Fix by leaving it alone for Order::Same as well.		 06:48
roast: 6fd123a1bf | (Zoffix Znet)++ | S32-list/sort.t
Remove trailing whitespace
roast: 5dc25f13e6 | (Zoffix Znet)++ | S32-list/sort.t
Cover degenerate cases of `sort`

Closes github.com/rakudo/rakudo/issues/1472 Rakudo fix: github.com/rakudo/rakudo/commit/2e65de2999		 06:49
nine Zoffix: how did you get that error message in the first place? Better BEGIN time error handling would help in many cases :) 09:54
Zoffix: I think the error message is still misleading as there simply isn't any Perl 6 hash access in any of the code that should now run at BEGIN time 11:12
Really, if any part of the setting at all, src/core/Exception.pm should be written in nqp. 11:51
lizmat boarding& 13:36
Geth rakudo/udp_receive_hostname_port: 3b35f64c64 | (Timo Paulssen)++ | src/core/IO/Socket/Async.pm
first draft of "datagram"-oriented API for udp sockets

ask the .Supply to give you :datagrams and you'll receive instances of IO::Socket::Async::Datagram with
  .data being the string or buffer of data and
  .hostname and .port being the sender's address.		 14:11
timotimo (just a rebase)
Geth rakudo: timo++ created pull request #1473:
[RFC] first draft of "datagram"-oriented API for udp sockets		 14:14
roast: 3dab7c7745 | (Aleks-Daniel Jakimenko-Aleksejev)++ | S32-exceptions/misc.t
Test malformed sigs with return types

Adds tests for RT#127100.		 15:02
synopsebot RT#127100 [new]: rt.perl.org/Ticket/Display.html?id=127100 [LTA] error message when specifying return type before parameters ( --> Bool, Int $x, Int $y )
TimToady looks forward to the day he can wrap a datagram interface around stdin 17:47
(which is how xinetd passes the socket in) 17:48
otoh, not sure we really need a new type for the connection, if recv can just figure out what it has with an ordinary socket 17:53
I can see us auto-figuring-out that stdin is a socket, but I dunno that we can figure out that it's a datagram socket
I guess we can figure that out 17:56
yeah, getsockopt(SO_TYPE), at least on linux 18:01
we really should do socket autodetection on every passed in fd 18:02
*passed-in 18:05
TimToady now pictures a passed-out socket...
Geth roast: 3ca315596c | (Aleks-Daniel Jakimenko-Aleksejev)++ | S29-os/system.t
Test Proc.encoding

Resolves RT#129296		 18:07
synopsebot RT#129296 [open]: rt.perl.org/Ticket/Display.html?id=129296 Splitting non-binary handles no longer works (run(:out, "ls").out.split: 0.chr)
TimToady in my telephone logging program, if I force gc after every line, it stabilizes at about 500M, but if I let the gc have its own way, it keeps growing, currently at 1.4G, with the resident set at 1G, which is twice what the program actually needs 18:22
so it seems as though either something is fragmenting long-term memory, or we have some kind of leak that only happens outside the nursery 18:23
or the gc is just not very well tuned to my particular mix 18:24
200% overhead seems a bit much though 18:25
in any case, the fact that it stabilizes at .5G when I force gc would seem to indicate I don't have a normal memory leak 18:26
in the "stable" state, it periodically gives a meg back, then takes it back almost immediately 18:28
I dunno if it ever gives anything back when I let the gc decide when to run
which is what makes me think that perhaps the meg it would have given back under force ends up getting something in it that prevents reclamation but also discourages reuse of the other memory in that meg somehow 18:30
Geth rakudo: df1712d81c | (Nick Logan)++ (committed using GitHub Web editor) | src/core/CompUnit/RepositoryRegistry.pm
Fix home repository to point at correct location

Fixes the following incorrect blead-only behavior:
  ```
perl6 -e 'say CompUnit::RepositoryRegistry.repository-for-name("home")' inst#/Users/ugexe ... (11 more lines)		 22:19
roast: ceabda6337 | (Aleks-Daniel Jakimenko-Aleksejev)++ | S32-exceptions/misc.t
Test line number of the corresponding starter

Adds tests for RT #130261. I don't find this test elegant, but it should be enough to close the ticket.		 22:22
synopsebot RT#130261 [new]: rt.perl.org/Ticket/Display.html?id=130261 [LTA] Error message for unclosed curly quote points to the last line of the file
Geth roast: ff0472adfc | (Moritz Lenz)++ | S32-temporal/DateTime.t
RT#125555: DateTime comparison ops
 22:31
synopsebot RT#125555 [open]: rt.perl.org/Ticket/Display.html?id=125555 [NYI] Comparison ops for DateTimes in Rakudo
Geth roast: 7557eef426 | (Aleks-Daniel Jakimenko-Aleksejev)++ | S02-lexical-conventions/unspace.t
Tests unspace + map + constants

Adds tests for RT #125985.		 22:51
synopsebot RT#125985 [open]: rt.perl.org/Ticket/Display.html?id=125985 [BUG] Internal error when a constant, unspace, and a map are involved in Rakudo
stmuk rsync --delete is a good flag after a few Belgium beers 22:56
geekosaur --no-ossifer-Im-not-drunk-really 22:58
Zoffix .tell nine *I* didn't get the error. It only happens on Travis. And to get the error to print, I had to make temporary modifications to X::Comp::BeginTime to printe the message from a Failure being stuffed into a `str` attribute: github.com/rakudo/rakudo/commit/2339d8dd00 23:33
yoleaux Zoffix: I'll pass your message to nine.

Powered by Cro and the RakuĀ® Programming Language.

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

Thank you!