Perl 6 language and compiler development | Logs at colabti.org/irclogger/irclogger_logs/perl6-dev | For toolchain/installation stuff see #perl6-toolchain | For MoarVM see #moarvm
Set by AlexDaniel on 12 June 2018.
00:06 ufobat_ left
kjp timotimo: Certainly just removing "is required" makes simple applications work again, and indeed that change has already been merged into DBIish. 00:22
You're probably right about this stuff not being used directly by end-users.
I suspect that it's just $!pc_conn on the assumption that the previous "is required" will catch errors. 00:23
You're pushing my Perl 6 knowledge :-) Will just adding exclamations marks provide all that's required?
timotimo that will give a generic error message 00:24
let me write up an example
... actually my keyboard i having some trouble right now
i can't write closing curly brace
m: class Test { has $!foo; submethod BUILD(:$!foo!) { } }; Test.new
camelia Required named parameter 'foo' not passed
in submethod BUILD at <tmp> line 1
in block <unit> at <tmp> line 1
timotimo but with a die it'll be more helpful 00:25
m: class Test { has $!foo; submethod BUILD(:$!foo = die "you have to construct a Test with a foo attribute") { } }; Test.new
camelia you have to construct a Test with a foo attribute
in submethod BUILD at <tmp> line 1
in block <unit> at <tmp> line 1
kjp Okay; I'll have a go at that. 00:26
04:49 brrt joined 05:26 ufobat_ joined
geekosaur so fwiw I'm seeing occasional random segfaults and bus errors from perl6 master 05:40
rakudo master, sorry
am trying to see if turning off various things makes them go away, but they're too sporadic
06:37 brrt left 06:49 lizmat left 06:59 [TuxCM] left
[Tux] m: use Inline::Perl5; Inline::Perl5.new.use("Text::CSV_PP") 07:02
camelia ===SORRY!===
Could not find Inline::Perl5 at line 1 in:
/home/camelia/.perl6
/home/camelia/rakudo-m-inst-1/share/perl6/site
/home/camelia/rakudo-m-inst-1/share/perl6/vendor
/home/camelia/rakudo-m-inst-1/share/perl6
Comp…
07:03 lizmat joined
[Tux] For those that like to investigate that: Internal error: inconsistent bind result 07:04
Rakudo version 2018.06-193-gaff96baef - MoarVM version 2018.06-223-ga50a0b15b
csv-ip5xs1.093 - 1.277
csv-ip5xs-208.610 - 8.839
csv-parser25.569 - 25.614
csv-test-xs-200.438 - 0.447
test9.523 - 9.682
test-t2.360 - 2.398
test-t --race0.989 - 1.010
test-t-2040.351 - 42.948
test-t-20 --race14.754 - 14.816
 07:05
I have issued the I::P5 case as github.com/niner/Inline-Perl5/issues/116 07:11
07:11 robertle joined, stmuk_ joined 07:13 stmuk left
lizmat Files=1243, Tests=76378, 385 wallclock secs (16.06 usr 5.38 sys + 2467.24 cusr 233.55 csys = 2722.23 CPU) 08:01
08:04 pmurias joined 08:08 Ven` joined 08:12 brrt joined
brrt geekosaur: bus errors may be JIT related 08:15
try running with MVM_SPESH_BLOCKING=1 08:16
Geth nqp/truffle: 87c2e222d3 | (Paweł Murias)++ | 2 files
[truffle] Return nqp::null when using a VOID op as an OBJ
 09:10
nqp/truffle: 14fd9f0c3d | (Paweł Murias)++ | 5 files
[truffle] Implement nqp::shift and nqp::iterator
nqp/truffle: 8caec7c4f9 | (Paweł Murias)++ | src/vm/jvm/bin/run_tests.pl
[truffle] Run passing test 15
nqp/truffle: b85595b5cb | (Paweł Murias)++ | 3 files
[truffle] Implement nqp::atpos, nqp::bindpos.

Pass test 17		 09:19
nqp/truffle: 78be0c072e | (Paweł Murias)++ | 2 files
[truffle] Add missing files
 09:29
09:34 dogbert2 joined
Geth nqp/truffle: 4 commits pushed by (Paweł Murias)++ 09:34
jnthn So the failure in github.com/rakudo/rakudo/issues/2056 was, in the immediate, due to an overflow while processing the string heap during bytecode production 09:59
Which I fixed, but still, that file compiling into something with a string heap of 275,000+ string seems a bit...odd
When I look at what those strings are, though, they're all like 10:00
Int|999408
Int|999409
Or at least, many of them are. Wonder where on earth that comes from...
Ohhh 10:01
Unicode::PRECIS::Tables builds up some big sets
OK, that explains it somewhat, though still I don't understand why it once worked and now not 10:02
Let alone why it'd bissect to github.com/rakudo/rakudo/commit/d3...c5d8165116
AlexDaniel c: d3c5381b74^,d3c5381b74 gist.github.com/AlexDaniel/f5eb6ef...415fc7a639 10:08
yoleaux 15 Jul 2018 22:13Z <tbrowder_> AlexDaniel: please see my doc PR #2177 which implements Phase 2 of my proposed Language doc handling
committable6 AlexDaniel, gist.github.com/6290c4ae9bebe88fb2...49d2f6f606
AlexDaniel just in case
yeah, bisected correctly :)
jnthn ohh...I wonder 10:10
tbrowder_ AlexDaniel: i fixed typo for failing tests and replied to yr msg at PR 10:11
jnthn Yeah, OK, I know why 10:14
dogbert2 a bug or a feature
jnthn Previously, when we assigned to Scalar, we did it in cheating C code 10:15
Now we do it with nqp::bindattr
That includes the serialization context write barrier
Which is the thing that tracks when we have changed an object serialized in a different SC
In this case, that thing is the our-scoped Scalar 10:16
Which is exported
Since one could probably write a test that would consider the previous behavior a bug, this probably counts as improvement 10:18
10:20 robertle_ joined
dogbert2 interesting 10:22
so which regression is next on your list? 10:23
jnthn I probably need to look at the others in that issue some more 10:30
10:30 Zoffix joined
Zoffix .tell AlexDaniel don't know if this even have any meaning, but I was trying to install a module with zscript on HEAD-ish commit and zef apparently just bailed out after searching for a module. Re-trying installed it fine: gist.github.com/zoffixznet/8d72d3e...10d3fafd4b 10:32
yoleaux Zoffix: I'll pass your message to AlexDaniel.
pmurias MasterDuke: the unused variables in while code are there so that they can be used in the future 10:33
yoleaux 14 Jul 2018 20:53Z <MasterDuke> pmurias: *@ doesn't seem to work in signatures with Truffle. e.g. `'sub foo(*@f) { say(nqp::elems(@f)) }; foo(1, 2, 3)'` dies with `java.lang.RuntimeException: Can only call elems on arrays or hashes` (with a yet-to-be-pushed commit to make elems support NQPList). where would i look to fix that?
10:33 Zoffix left
AlexDaniel . 10:33
yoleaux 10:32Z <Zoffix> AlexDaniel: don't know if this even have any meaning, but I was trying to install a module with zscript on HEAD-ish commit and zef apparently just bailed out after searching for a module. Re-trying installed it fine: gist.github.com/zoffixznet/8d72d3e...10d3fafd4b
lizmat Zoffix: modules.perl6.org/dist/Readline:cpan:JGOFF still shows version 0.1.1 while at least 0.1.3 has been uploaded to CPAN ? 10:34
AlexDaniel interesting, but maybe zef itself failed to pull a list of modules or something like that?
pmurias MasterDuke: re *@ in signatures in Truffle.nqp you need to hook the support in the compile_params method 10:35
MasterDuke: on the java side we need a node that gets slurpy arguments (look into NQPGetPositionalNode for the node that gets a single positional node) 10:36
Geth nqp: 896f009d2a | (Aleks-Daniel Jakimenko-Aleksejev)++ | tools/build/MOAR_REVISION
[MoarVM Bump] Brings 4 commits

MoarVM bump brought: github.com/MoarVM/MoarVM/compare/2...gb7c6afd92 b7c6afd92 Use larger integer time to iterate strings heap a50a0b15b Make compare-oplist-interp-order.sh more POSIX compliant 0dd6ce547 Merge pull request #904 from samcv/remove-backref-hash2 e94d3dfdb Remove previous hash handle pointers in hash implementation		 10:49
¦ nqp: version bump brought these changes: github.com/MoarVM/MoarVM/compare/2...gb7c6afd92
rakudo: ee9314d02f | (Aleks-Daniel Jakimenko-Aleksejev)++ | tools/build/NQP_REVISION
[NQP Bump] Brings 2 commits

NQP bump brought: github.com/perl6/nqp/compare/2018....g896f009d2 896f009d2 [MoarVM Bump] Brings 4 commits 09d75a951 Add some return types to doc signatures
¦ rakudo: version bump brought these changes: github.com/perl6/nqp/compare/2018....g896f009d2
10:50 Ven` left
Geth roast/handle-6.d: a3194da14b | (Zoffix Znet)++ | S32-io/io-handle.t
Finish spec of IO::Handle.WRITE/.READ/.EOF
 10:53
rakudo/handle-6.d: 47 commits pushed by (Elizabeth Mattijsen)++, (Daniel Green)++, (Jonathan Worthington)++, (Nick Logan)++, (Zoffix Znet)++, (Aleks-Daniel Jakimenko-Aleksejev)++
review: github.com/rakudo/rakudo/compare/4...78685710d9		 10:54
10:56 Zoffix joined
Zoffix lizmat: unsure why but the metafile included in the dist itself still has version set at 0.1.0 10:57
lizmat aha.. ok, that could be the issue.. :)
Zoffix lizmat: oh, nm, it's updated actually: github.com/drforr/perl6-readline/b...META6.json 10:58
I was looking at modules.perl6.org
In build log I see it's being processed without errors: modules.perl6.org/update.log
Don't have time to look at it right now 10:59
|10h invesitage mp6 update breakatge: colabti.org/irclogger/irclogger_log...07-16#l110
ZofBot Zoffix, Will remind you on 2018-07-16T16:59:09.106161-04:00 about invesitage mp6 update breakatge: colabti.org/irclogger/irclogger_log...07-16#l110
10:59 Zoffix left
lizmat keeps a note of that link 10:59
11:02 brrt left
AlexDaniel weekly: reportable: gist.github.com/04593645cef335360a...eb32f21708 11:06
notable6 AlexDaniel, Noted!
Geth nqp/truffle: 2652a4a122 | (Paweł Murias)++ | src/vm/jvm/Truffle.nqp
[truffle] Fix ops definitions
 11:08
nqp/truffle: 31696e05a6 | (Paweł Murias)++ | src/vm/jvm/bin/run_tests.pl
[truffle] Stop running a test that doesn't pass
 11:09
jnthn m: class C { has Int $!x; method m(Mu $!x) { } }; C.new.m('nope') 11:14
camelia Type check failed in assignment to $!x; expected Int but got Str ("nope")
in method m at <tmp> line 1
in block <unit> at <tmp> line 1
jnthn committable6: 2018.06 class C { has Int $!x; method m(Mu $!x) { } }; C.new.m('nope')
committable6 jnthn, ¦2018.06: «Type check failed in assignment to $!x; expected Int but got Str ("nope")␤ in method m at /tmp/3ew1uL9lvw line 1␤ in block <unit> at /tmp/3ew1uL9lvw line 1␤␤ «exit code = 1»»
jnthn committable6: class C { has Int $!x; method m(Mu $!x) { } }; C.new.m('nope') 11:16
committable6 jnthn, ¦class: «Cannot find this revision (did you mean “Plano”?)»
jnthn committable6: ALL class C { has Int $!x; method m(Mu $!x) { } }; C.new.m('nope')
committable6 jnthn, gist.github.com/57f23f38fb0fd05a9b...92bd91ee64 11:17
11:25 Ven` joined
timotimo hum. we don't serialize an object's objectid, do we? 11:33
jnthn Apparently if we serialize a Set we do... 11:34
timotimo i mean, the current implementation doesn't really make it simple to do
we put the objectid we got into the string heap, but the object itself won't retain its object id, right?
jnthn Correct
timotimo that's probably trouble :) 11:35
jnthn *sigh* The spit regression in github.com/rakudo/rakudo/issues/2056 is bizzare 11:37
lunch &
11:43 klapperl joined 11:44 klapperl_ left
lizmat jnthn: Set uses .WHICH for the keys, which for non-value types uses nqp::objectid . Probably stating the bleedingly obvious, but I thought I'd mention it 11:50
12:01 brrt joined
samcv weekly: use siphash on MVM github.com/MoarVM/MoarVM/commit/d9...76dc8c8e47 and reduce the size of each hash entry by 8 bytes github.com/MoarVM/MoarVM/commit/e9...7d05c6917f 12:08
notable6 samcv, Noted!
jnthn back 12:31
brrt \o jnthn 12:39
Geth rakudo: ae25c4b061 | (Jonathan Worthington)++ | src/Perl6/Optimizer.nqp
Avoid a bunch of (caught) exceptions in Optimizer

This didn't affect correctness, but was a bit noisy when trying to debug and look for another more interesting exception being thrown. Exceptions are also somewhat costly, and a quick concreteness check before we try the `getattr` can save us that.		 12:40
rakudo: 8a76c00c2b | (Jonathan Worthington)++ | src/Perl6/Grammar.nqp
Never repossess the quote lang cache

We don't want to go serializing different copies of it into various precompilations.
rakudo: abfa1b1591 | (Jonathan Worthington)++ | src/Perl6/Grammar.nqp
Ensure compilations get a fresh %*SIG_INFO
jnthn Hmm, so the spit module that breaks has: 12:50
has Spit::Type $!param-of;
And then:
method set-param-of(Mu $,Mu $!param-of,Int:D $!param-pos) { } 12:51
The assignment to the attribute started failing after the assignment changes
But, if I go back to a version where it works, and do:
method set-param-of(Mu $,Spit::Type $!param-of,Int:D $!param-pos) { }
Then we get a parameter bind error
So it's almost as if the type check was being ignored 12:52
But I can't get it to ignore it in an isolated example 12:58
dogbert2 Spit eh :) 13:00
Ven` what's a Spit::Type? An enum? 13:02
jnthn A role 13:03
This thing is doing a bunch of meta-programming
Ven` Yeah, it seems. This is my first time seeing a .sp file :P
jnthn oh..hm
And more interestingly, the type being checked against isn't composed yet 13:04
Urgh. The old Scalar assign implementation has its own implementation of what istype does 13:06
And I'm guessing there might be a discrepancy between them 13:07
m: class Unfinished { ... }; class C { has Unfinished $!x; method m(Mu $!x) { } }; BEGIN C.new.m('nope') 13:16
camelia 5===SORRY!5=== Error while compiling <tmp>
An exception occurred while evaluating a BEGIN
at <tmp>:1
Exception details:
Died with X::TypeCheck::Assignment
in method m at <tmp> line 1
in code at <tmp> line 1
jnthn committable6: 2016.06 class Unfinished { ... }; class C { has Unfinished $!x; method m(Mu $!x) { } }; BEGIN C.new.m('nope')
committable6 jnthn, gist.github.com/723d263a757b2f1ff2...4ee06b44e3
jnthn committable6: 2016.06 constant \Unfinished = Metamodel::ClassHOW.new_type(:name<Unfinished>); class C { has Unfinished $!x; method m(Mu $!x) { } }; BEGIN C.new.m('nope') 13:17
committable6 jnthn, gist.github.com/e8a99efe943b713004...6d768af6a2
lizmat jnthn: after the latest nqp bump t/spec/S06-currying/positional.t segfaults at random places for me 13:18
timotimo i get segfaults there, too 13:20
lizmat same for t/spec/integration/eval-and-threads.t btw
jnthn lizmat: Yeah, I saw that too; see #moarvm, looks like it may be hashing changes 13:21
timotimo that as well
lizmat jnthn: I don't think my versio already has the siphash changes
jnthn There was a previous hashing change
Yeah, that was merged in yesterday 13:22
And somebody did a bump this morning 13:23
So that'll be in there
lizmat ok, yeah, that one is in there, yup
jnthn It seems there's a discrepancy between nqp::istype and the way that scalar assignment used to do type checking, which had its own similar-but-not-quite-the-same implementation of it 13:24
Still need to get a reproduction but it looks rather like the assignment one, if it didn't find a way to type check because the type wasn't composed yet, would just let the assignment proceed 13:25
The nqp::istype used pretty much everywhere else is stricter
timotimo demonic
lizmat what is wisdom: make nqp::istype as lenient ?
jnthn That feels...dubious. 13:26
Make everything dodgy because one thing used to be?
lizmat well... then nqp::istype should be able to support uncomposed types ? 13:27
in a slow path ?
jnthn Or if you write code to deal with uncomposed types you just have to deal with them as Mu
Note the code does this:
method set-param-of(Mu $,Mu $!param-of,Int:D $!param-pos) { } 13:28
But then this:
has Spit::Type $!param-of;
If it had used anything except Mu in the parameter list there, then it would have failed.
lizmat well, if that's going to be documented like that, I think that's a situation we can live with ? 13:29
jnthn Ouch, changing the default nqp::istype behavior to leninet = instant segv 13:30
I'm pretty sure we shouldn't do that anyway 13:33
aha, I might have found a way to work around this in the MOP 13:41
Ven` seems to me like we'd want the type checker to refuse assigning here ;o) 13:43
jnthn OK, I think I have a sensible solution that doesn't feel like a total hack :) 13:45
We can't *really* figure out all of the available roles until composition time, but we can at least use the set of those we've been directly given 13:46
Which in this case is enough, and using type information that's actually been provided up to that point 13:47
Hurrah, that gets another module's tests happy
If somebody has said $obj.^add_role(Foo) and then does $obj ~~ Foo, even without composing, it's not reasonable enough for us to say yes 13:50
So long as they don't expect us to deal in transitive roles until composition time, it's all good.
m: my $type = Metamodel::ClassHOW.new_type(); $type.^add_role(Positional); say $type ~~ Positional 13:53
camelia False
jnthn That says True with my local changes
lizmat whee! :-) 13:55
13:57 brrt left
Geth rakudo: 38d046fafb | (Jonathan Worthington)++ | src/Perl6/Metamodel/ClassHOW.nqp
Before composition, typecheck against added roles

Previously:
   my $type = Metamodel::ClassHOW.new_type();
   $type.^add_role(Positional);
   say $type ~~ Positional;
... (14 more lines)		 14:00
Ven` can you remove a role? 14:01
timotimo before composition it could be done
after composition, not so much
Ven` how do you do that?
jnthn I don't think there's any way to do it, though timotimo is right in terms of what would be in theory possible to implement. 14:02
timotimo well, you could use getattr and bindattr to get at the array it uses to store the roles 14:03
Ven` Just to make sure `MyType.^add_role(R); R t = MyType.new; MyType.^remove_role_somehow(R); MyType.^compose;`
14:03 MasterDuke left
timotimo yeah 14:03
Ven` doesn't actually do something quite insane
timotimo it does do something quite insane. but it'll be your own fault
Ven` (well it probably works after the commit :P)
okay, fair enough :D
timotimo we will not support something like .^remove_role_somehow, i don't think 14:04
we also don't support using nativecall to find a routine's bytecode and modifying that
Ven` wait, I can't get the vm bytecode of a function easily? how unpythonistic :P
timotimo right, python makes that easy 14:05
jnthn It could be something one can request via the debug protocol :)
timotimo yeah, i'd like to have that
jnthn Especially if you can also ask for the specializations :)
timotimo yes yes yes
Ven`
.oO( Dear Santa, ... )
 14:06
14:06 skids joined 14:40 ufobat_ left
lizmat notable6: weekly 15:04
notable6 lizmat, 4 notes: gist.github.com/a8d3f7b181517f9448...eeec740d8c
Geth nqp: 1ca664ac66 | (Jonathan Worthington)++ | src/vm/moar/QAST/QASTCompilerMAST.nqp
Disable the SC write barrier during load

The load time of a module is dynamically within the scope of the compilation of the using module. However, from a user perspective the module mainline is probably more fittingly seen as a bit of runtime during compile time.
... (18 more lines)		 15:06
15:10 fake_space_whale joined 15:51 squashable6 left, squashable6 joined 16:14 Ven` left 16:19 lizmat left 16:23 |Tux| left 16:24 lizmat joined 16:27 fake_space_whale left 16:28 robertle_ left
Geth 6.d-prep: c63404fb09 | (Zoffix Znet)++ (committed using GitHub Web editor) | d-docs/New-Features-Policy.md
Mod trial impl. in ecosystem requirement

For some features this would be impossible, so require features under `experimental` pragma for such cases instead.		 16:38
6.d-prep: fdf468e183 | (Zoffix Znet)++ (committed using GitHub Web editor) | d-docs/New-Features-Policy.md
Tweak wording for clarity
 16:40
16:48 |Tux| joined 16:57 Zoffix joined
Zoffix Geth ded? I see new commits in rakudo and nqp in the last 15m that weren't announced, were they? 16:57
16:59 Geth left
jnthn Hm, I saw it announce one by you about 10 minutes ago 16:59
16:59 Geth joined, ChanServ sets mode: +v Geth
jnthn Oh, but not from the Rakudo/NQP repo 16:59
Geth rakudo: 980f692ccd | (Aleks-Daniel Jakimenko-Aleksejev)++ | tools/build/NQP_REVISION
[NQP Bump] Brings 2 commits

NQP bump brought: github.com/perl6/nqp/compare/2018....g33e22b739 33e22b739 [MoarVM Bump] Brings 4 commits 1ca664ac6 Disable the SC write barrier during load
¦ rakudo: version bump brought these changes: github.com/perl6/nqp/compare/2018....g33e22b739
nqp/truffle: a0454b35c9 | (Paweł Murias)++ | 7 files
[truffle] Implement 'local' scoped variables
 17:00
nqp/truffle: 58d2d82400 | (Paweł Murias)++ | nqp-truffle.nqp
[truffle] Enable the optimizer
nqp/truffle: a167db98ae | (Paweł Murias)++ | 2 files
[truffle] Compile with the JVM setting

We need to swith to using a one dual compiled that we can load at some point Before we can deserialize we use a placeholder for QAST::WVal.
nqp: 33e22b739b | (Aleks-Daniel Jakimenko-Aleksejev)++ | tools/build/MOAR_REVISION
[MoarVM Bump] Brings 4 commits

MoarVM bump brought: github.com/MoarVM/MoarVM/compare/2...g2f807cd59 2f807cd59 learn facts from a couple of sp_ ops 1a4adb0e3 Merge pull request #889 from samcv/siphash 3049b0420 Mix in the instance pointer into the first hash secret d9a3270aa Implement SipHash, use as our hashing function w/ 64bit hashvals
¦ nqp: version bump brought these changes: github.com/MoarVM/MoarVM/compare/2...g2f807cd59
Zoffix k 17:01
17:02 Zoffix left 17:03 Zoffix joined
Zoffix lizmat: modules.perl6.org/dist/Readline:cpan:JGOFF points to right release now (after I told it to gen new db from scratch). Pretty sure it's dupe-dist bug that's causing the individual URL to a dist to pick among the available dist (based on insert order, I'm guessing): github.com/perl6/modules.perl6.org/issues/106 17:04
And readline shows up 3 times right now: modules.perl6.org/search/?q=readline
.oO( rewrite modules.perl.org using Cro )
 17:07
Should have the time in 2020 :)
17:07 Zoffix left 17:37 ZofBot left 17:39 ufobat_ joined
DrForr waves. 18:06
18:14 ExtraCrispy left 18:33 ufobat_ left
releasable6 Next release in ≈4 days and ≈23 hours. 10 blockers. Please log your changes in the ChangeLog 19:00
19:52 ZofBot joined, ChanServ sets mode: +v ZofBot 19:56 lizmat left
Geth rakudo: 9abc10212c | (Elizabeth Mattijsen)++ | src/core/Hash.pm6
Revert "Copy Hash.STORE logic to Object Hash for now"

This reverts commit 7848cdb484c3ab096d043b015b5465228db9b3e5.
The Hash.STORE refactor is probably not going to happen before the 2018.07 release. So remove this scaffolding for now.		 20:04
20:04 lizmat joined
Geth 6.d-prep: 8d4046e46c | (Zoffix Znet)++ (committed using GitHub Web editor) | ChangeLog.md
Log some constants changes
 20:05
geekosaur brrt, I disabled spesh entirely and still got them. there were segfaults coming from gc after a recent patch though, which seem unrelated. 20:07
gleh
.tell brrt I disabled spesh entirely and still got them. there were segfaults coming from gc after a recent patch though, which seem unrelated.
yoleaux geekosaur: I'll pass your message to brrt.
geekosaur .tell brrt (unrelated to spesh, that is)
yoleaux geekosaur: I'll pass your message to brrt.
20:42 pmurias left, pmurias joined 21:05 MasterDuke joined
lizmat and another Perl 6 Weekly hits the Net: p6weekly.wordpress.com/2018/07/16/...llination/ 21:06
hmmm... hmmm... Geth seems to have missed at least one commit 21:11
MasterDuke Zoffix: the poster say "of programming of the future". it looked like there was some chat about changing to "of programming in the future"? 21:15
lizmat yeah, I suggested that, but I think both are ok ( but then again, English is not my native language )
21:19 skids left
MasterDuke "of ... of" works, but as a native speaker i still had to think a little about it 21:24
but that might just be me, a wider poll would be good
geekosaur "of the future of programming"? 21:26
(which is still of ... of but reads a bit more naturally to me, and might be slightly less ambiguous) 21:27
21:35 lizmat left 21:36 lizmat joined
Geth rakudo: 8a8ce95efa | (Elizabeth Mattijsen)++ | src/core/PseudoStash.pm6
Make PseudoStash pseudoers hash a native hash

We won't be calling PseudoStashes so often that the optimizer would take care of this. So do it ourselves.		 21:51
22:00 pmurias left
Geth rakudo: jnthn self-assigned `Cannot call method 'ACCEPTS' on a null object` in Slang::SQL github.com/rakudo/rakudo/issues/2080
4766536b41 | (Jonathan Worthington)++ | src/vm/moar/spesh-plugins.nqp

While we can fix bugs in Rakudo where Perl 6 constructs wrongly produce it, we have some slangs that use `nqp::ops`, and our hands are rather tied if we're not wanting to break them. Since we're doing this with a spesh plugin, the cost is really low anyway (in the best case, it'll be free), so we might as well put the mapping back and not break a module that was depending on it.		 22:08
22:12 robertle left
roast/handle-6.d: 61 commits pushed by (Jonathan Worthington)++, (Zoffix Znet)++, (Elizabeth Mattijsen)++, MasterDuke17++
review: github.com/perl6/roast/compare/a31...2b1741ae1a		 23:09
roast/handle-6.d: 3e734149f1 | (Zoffix Znet)++ | 2 files
Create A02-some-day-maybe spec APPENDIX

Some things should have reasonably-defined behaviour, but ATM that behaviour is to just throw X::NYI (e.g. write methods on an IO::CatHandle). Some day we might come up with good things to do instead of X::NYI and until that day this APPENDIX will basically just list such things.		 23:18
rakudo/handle-6.d: ae9a3628eb | (Zoffix Znet)++ | 2 files
Move IO::CatHandle's NYI method to A02 spec APPENDIX

Since we document these, they should really live in a spec, not in an implementation-specific test suite.
Moved to: github.com/perl6/roast/commit/3e734149f1		 23:20
rakudo/handle-6.d: 9246b03520 | (Zoffix Znet)++ | src/core/IO/CatHandle.pm6
Block out .READ/.WRITE/.EOF in IO::CatHandle

Just making them all throw X::NYI ATM and sticking that into SAP (Spec APpendices) until someone comes up with something good for these to do.		 23:24
roast/handle-6.d: e0f6f389d5 | (Zoffix Znet)++ | APPENDICES/A02-some-day-maybe/io-cathandle.t
Add .WRITE/.EOF/.READ to SAP (Spec APendices)

Rakudo impl: github.com/rakudo/rakudo/commit/9246b03520		 23:25
23:34 Zoffix joined
Zoffix stresstest failures multiplied: gist.github.com/zoffixznet/8cddb31...354902ffe0 23:35
3 more files fail now
t/spec/S32-num/stress.t looks to be segfaulting now 23:36
Yeah, I get either Non-zero wait status: 11 or Non-zero wait status: 7 for these new three 23:37
.oO( do I have latest master merged in? )
 23:38
Geth rakudo: zoffixznet++ created pull request #2081:
Rename IO::Handle's generic methods to .READ/.WRITE/.EOF		 23:39
rakudo/master: 7 commits pushed by (Zoffix Znet)++
roast: zoffixznet++ created pull request #452:
Spec for IO::[Cat]Handle.WRITE/.READ/.EOF methods		 23:40
Zoffix man I see the hook announcement on IRC before the page even managed to load :P
Geth roast/master: 6 commits pushed by (Zoffix Znet)++
Zoffix Doc build died: travis-ci.org/perl6/doc/jobs/404390772 util/travis-build.sh: line 6: 6775 Segmentation fault (core dumped) zef --/tap-harness --depsonly install . 23:42
23:43 nebuchadnezzar left
Geth 6.d-prep: ba720e6462 | (Zoffix Znet)++ | 2 files
Mark IO::Handle.READ/.WRITE/.EOF as completed

POVI (Proof-Of-Viability Implementation): github.com/rakudo/rakudo/pull/2081 Spec and SAP: github.com/perl6/roast/pull/452 Docs: github.com/perl6/doc/pull/2181		 23:44
Zoffix w00t
One fewer items on the list \o/ 23:45
buggable: d
buggable Zoffix, Diwali is in 113 days, which is 16 weeks, 14 minutes, and 53 seconds. Need to review 31 commits per day (219/week)to complete. Need 8 teaser flyers.
23:48 Ven` joined, buggable left, buggable joined, ChanServ sets mode: +v buggable
Zoffix buggable: d 23:48
buggable Zoffix, Diwali is in 113 days, which is 16 weeks, 11 minutes, and 24 seconds. Need to review 34 commits per day (240/week)to complete. Need 8 teaser flyers. Still have 9 TODO features costing 81 hours. Still have 2.3 policies to write. Still have ~95% of ChangeLog to do.
Zoffix 🤘🤘🤘 www.youtube.com/watch?v=9jK-NcRmVcw
23:52 Ven` left
Zoffix heh, 6 failing files on master 23:54
ZOFVM: Files=1299, Tests=153365, 165 wallclock secs (21.28 usr 3.31 sys + 3578.31 cusr 166.99 csys = 3769.89 CPU) 23:55
updated R#2071 with list of new files
synopsebot R#2071 [open]: github.com/rakudo/rakudo/issues/2071 [⚠ blocker ⚠] Stresstest failures on v2018.06.179.g.9.dce.63318
23:56 Zoffix left

Powered by Cro and the Raku® Programming Language.

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

Thank you!