| releasable6 | Next release in ≈1 day and ≈15 hours. There are no known blockers. Please log your changes in the ChangeLog: github.com/rakudo/rakudo/wiki/ChangeLog-Draft | 03:00 | |
| timo | oh, is this actually also a fix for the ndarray atpos spesh bug i diagnosed the other day? github.com/rakudo/rakudo/commit/a8745fe8 | 03:51 | |
| or i guess not since this commit is 2 months old? | |||
|
06:10
[Coke] left
07:06
sena_kun joined
07:17
sena_kun left
08:08
[Coke] joined
08:51
finanalyst joined
|
|||
| lizmat | m: class A { }; proto sub postcircumfix:<[; ]>(|) {*}; multi sub postcircumfix:<[; ]>(A:D, @a) { dd }; multi sub postcircumfix:<[; ]>(|c) { dd }; (A.new)[0;0] | 09:46 | |
| camelia | sub postcircumfix:<[; ]>(A:D, @a) | ||
| lizmat | aha! | ||
| ab5tract | There we go :) | 10:01 | |
|
10:04
finanalyst left
11:14
sena_kun joined
11:43
sena_kun left
|
|||
| lizmat | so I'm working on a utility module META::varauthapi that will set the ^ver, ^auth, ^api of all classes provided by a distribution, using the info in the META6.json (aka $?DISTRIBUTION) | 12:30 | |
| basically: add "use META::verauthapi $?DISTRIBUTION" and your classes will be in sync with the META info | 12:31 | ||
| problem is: I cannot seem to be able to map what is in META<provides>.keys to the actual classes | 12:33 | ||
| even passing UNIT:: as an additional argument in the use statement, doesn't work as UNIT:: does not appear to be populated at compile time | 12:34 | ||
| m: my class A { }; BEGIN dd UNIT::.keys; dd UNIT::.keys | |||
| camelia | ().Seq ("\$?PACKAGE", "::?PACKAGE", "\$=pod", "\$=finish", "\$?LANGUAGE-REVISION", "\$/", "GLOBALish", "!UNIT_MARKER", "EXPORT", "\$¢", "A", "\$!", "\$_").Seq |
||
| lizmat | nine ab5tract ideas ? | 12:35 | |
| vrurg_ maybe? | |||
| vrurg_ | I could only tell that I was trying to get this done on the compiler side not without a reason. | 13:29 | |
| jdv | blin finally finished. took longer than usual. | 13:30 | |
| lizmat | vrurg_ , yes, and I hope we will get there with RakuAST... but I want it now :-) | 13:31 | |
| jdv | looks like there is some fallout. i'll get it posted in a bit once i get it cleaned up | ||
| lizmat | jdv++ | ||
| vrurg_ | lizmat: Do you want it for private classes too? Because I'd stick to what can be found in the META. | 13:33 | |
| lizmat | yes, that was the plan | ||
| although for now you can specify the classes as additional arguments | 13:34 | ||
| raku.land/zef:lizmat/META::verauthapi fwiw :-) | |||
| jdv | github.com/rakudo/rakudo/issues/5648 is blin | 14:01 | |
| ab5tract: one might be yours in there ^ | |||
| lizmat: the rest might be yours ? | |||
| ab5tract | jdv: ah, that appears to be an error in the test corpus | 14:03 | |
| it's expecting 35 lines but I added another one to the preamble | |||
| lizmat | bisectable6: my %h = a => 1, b => 2; dd %h{all(<a b>)}:exists | 14:08 | |
| bisectable6 | lizmat, Will bisect the whole range automagically because no endpoints were provided, hang tight | ||
| lizmat, Output on all releases: gist.github.com/576db79d2743283014...0359315acc | 14:09 | ||
| lizmat, Bisecting by output (old=2024.08 new=0b6c7dc) because on both starting points the exit code is 0 | |||
| lizmat, bisect log: gist.github.com/6801879cda47ce6f91...3782ca8e34 | |||
| lizmat, (2024-09-08) github.com/rakudo/rakudo/commit/dd...d52b101319 | |||
| lizmat, Bisecting by exit code (old=2016.02 new=2016.03). Old exit code: 1 | |||
| lizmat, bisect log: gist.github.com/f9474615db3d828964...32d2b50f57 | |||
| lizmat, (2016-03-04) github.com/rakudo/rakudo/commit/25...d49404c588 | |||
| lizmat, ⚠ New output detected, please review the results manually | |||
| lizmat, Output on all releases and bisected commits: gist.github.com/305a8f3a1100b393c0...f055862bea | |||
|
14:13
sena_kun joined
|
|||
| Geth | rakudo/main: 3ca83dc138 | (Elizabeth Mattijsen)++ | src/core.c/hash_slice.rakumod Make sure we have a Bool to check against with :exists d66162a4095ea0afbd0 introduced a short-cut that would make {}:exists always return false if the key was a Junction. Fix this by making sure we have a Bool:D to compare against (which should be optimized away pretty quickly if we already had a Bool:D). The alternative would have been to either revert these optimizations, or to create almost identical candidates handling Junctions specifically. That felt like a not very DRY approach, so instead have a little larger bytecode footprint for the fast path. |
14:35 | |
| lizmat | jdv ^^ | 14:36 | |
| jdv ab5tract the other failure is because $?LANGUAGE-REVISION => 2 has been added to a scope, which alters the number of elements tested for | 14:42 | ||
| so it is a flaky test | |||
| jdv | cool, we're back on track for tomorrow's release then | ||
| lizmat | flaky well, fragile | ||
| jdv | yup | ||
| i added an issue to that repo | 14:43 | ||
| thanks all | |||
| timo | m: class Wibble { method Bool() { Backtrace.new.Str.say; True } }; my %foo; say %foo<hello>:delete:exists(Wibble.new); | 16:21 | |
| camelia | in method Bool at <tmp> line 1 in block <unit> at <tmp> line 1 False |
||
| timo | is this funny? i'm not sure if this is funny | ||
| m: class Wibble { method Bool() { Backtrace.new.full.say; True } }; my %foo; say %foo<hello>:delete:exists(Wibble.new); # this is what i wanted to look at | 16:28 | ||
| camelia | in method new at SETTING::src/core.c/Backtrace.rakumod line 99 in method Bool at <tmp> line 1 in any at src/vm/moar/dispatchers.nqp line 4390 in sub postcircumfix:<{ }> at SETTING::src/core.c/hash_slice.rakumod line 49 in block <uni… |
||
| timo | jdv: the Data::Dump::Tree output has "please contact the author [...] deprecated code [...]" but what exactly was deprecated isn't in the part that's posted. anything important? or just PERL6LIB? | 16:32 | |
| jdv | that's not why it failed its test suite | 16:35 | |
| timo | ok that's fair | ||
| jdv | a lot of stuff still emits that:( | 16:36 | |
| timo | we are generating links to modules.raku.org in the blin output, which unfortunately currently just redirects directly to raku.land/ without trying to get you to the right dist, or even the search function | ||
| i'll ticket that in blin | |||
| jdv | yeah. i meant to fix that but there didn't seem to be a 1:1 solution that was quick and easy | ||
| timo | fair | 16:37 | |
| jdv | iirc raku.land requires more than just the dist name in a query | ||
| timo | probably also the auth | 16:38 | |
| jdv | any only a few people read that report but it definitely should be fixed:) | ||
| [Coke] looks forward to working on blin tickets. :) | 17:49 | ||
| jdv | nice. i hope to rebuild the docker image i use by eoy. i just don't have the time:( | 18:00 | |
| i hope next month's run doesnt take 11h and oom my box again | 18:01 | ||
| [Coke] | jdv++ | 18:30 | |
|
19:46
finanalyst joined
21:26
sena_kun left
|
|||
| ab5tract | can we please move beyond the idea that `dd` is an implementation detail? | 21:43 | |
| I will outright refuse to use an Raku implementation that decides its too fancy to support `dd` | |||
| of all the bonkers, used once in a million years by a mostly mad person things that a Raku implementation is required to support, the idea that `dd` is a theoretically useful afterthought ... frankly it comes off as disrespectful | 21:45 | ||
| jdv | isn't .raku supposed to do "that" | 21:47 | |
| iirc its because dd shows impl details but i don't remember enough to be useful | 21:49 | ||
| ab5tract | m: dd Int; say Int.raku | 21:50 | |
| camelia | Int Int |
||
| ab5tract | even if all it is is a wrapper function that looks like: `dd ($arg) { say $arg.raku }` (which it isn't), it would still be a required element of a Raku implementation imo | 21:52 | |
| but I'll take a deep breath and do some searches through #raku history to see what kinds of justifications have been wielded against it | 21:53 | ||
| timo | i'm thinking it was just never proposed to become a part of the language | 21:54 | |
| ab5tract | what would such a proposal look like? spec tests? | ||
| because I'e | |||
| because I've added a few features to the language that still have spec tests pending :) | 21:55 | ||
| (not proud of that, but it is what it is) | |||
| timo | we do what we can | 21:57 | |
| ab5tract | in terms of routines, methods, etc marked as implementation details, there is nothing that comes close to the usage count of `dd` for me | ||
| anyway, I'm going to add an explicit exception to the Comma inspection for implementation details, which is the only reason I've had this somewhat contained explosive reaction | 21:58 | ||
| timo | i'm pushing the debian packages stuff a bit further to actually getting up-to-date packages into debian sid (one day before the next release, but oh well) | 22:01 | |
| ab5tract | timo: that's seriously cool | 22:05 | |
| so many distros hang off the debian vetting process | 22:06 | ||
| or maybe a better way to say, so many distro installations hang off that process | |||
| jdv | ab5tract: what's rrally buggin you? | 22:21 | |
| *really | |||
| :) | 22:22 | ||
| ab5tract | 0 < $bugging < Inf | 22:23 | |
| :) | |||
| luckily we aren't pushing <= on the last part of the clause ;) | |||
| jdv | ha | 22:24 | |
| timo++ | |||
| timo | well, you can \+\+timo for now instead of timo\+\+ | 22:27 | |
| ab5tract | timo: I've been meaning to ask what happened to the other half of timotimo :) | 22:28 | |
| timo | i'm timo on github and libera, but i'm timotimo on salsa | ||
| difficult to say, really! | |||
| ab5tract | I'm very pleased to have as much timo as we can get around here :) | 22:29 | |
| timo | hm, what's the limit on nickname length on here then? :P | ||
| ab5tract | I don't know if I ever properly expressed my level of stoked-ness upon seeing your `$nick.subst(0, $nick.comb / 2)` resurface here on the regular :) | 22:31 | |
| jdv | i did albeit privately. i concur. | ||
| timo | <3 | 22:33 | |
|
22:39
finanalyst left
|
|||
| timo | salsa is keeping me hanging :| | 22:54 | |
| leaving me hanging* | 22:55 | ||
| jdv | i like my salsa scorching, nothing less;) | 22:57 | |
| releasable6 | Next release in ≈19 hours. There are no known blockers. Please log your changes in the ChangeLog: github.com/rakudo/rakudo/wiki/ChangeLog-Draft | 23:00 | |
| jdv | pipe down there | ||