🦋 Welcome to the IRC channel of the core developers of the Raku Programming Language (raku.org #rakulang). This channel is logged for the purpose of history keeping about its development | evalbot usage: 'm: say 3;' or /msg camelia m: ... | Logs available at irclogs.raku.org/raku-dev/live.html | For MoarVM see #moarvm
Set by lizmat on 8 June 2022.
08:07 sena_kun joined 08:35 finanalyst joined
finanalyst .tell [Coke] please take another look at doc-website PR#363. Makes indexed=true default, and defaults are not written unless changed 08:38
tellable6 finanalyst, I'll pass your message to [Coke]
09:18 finanalyst left
lizmat github.com/Raku/problem-solving/issues/424 10:50
nine lizmat: LOL did I really just guess an actual module and that's even written by yourself? What are the odds? :) 12:16
lizmat hehe
nine Btw. I'm not convinced that small reductions in bytecode size would be worth anything but minor changes. The most likely source of any performance improvement would be getting some routines below the inlining limit. But if that really helps so much we could as well just increase that limit by 5 %. 12:58
lizmat github.com/Raku/problem-solving/issues/425
I think the current limit is there for a reason
but perhaps we need to play with that as well to see what gives better performance 12:59
nine There's already a different max_inline_size for Raku (384) than the default of 192 (which still gets used for NQP). Both were determined empirically. 13:01
lizmat % bceval c '.frames.grep({ 385 < .opcodes.elems < 400}).elems' 13:19
268
the C setting has 268 frames that are just above the threshold
currently c setting has 18720 frames below the threshold 13:20
% bceval c '.frames.grep({ .opcodes.elems <= 384}).elems'
18720
all of this with a grain of salt, of course, as the bytecode from disk is really quickly altered by spesh in the runtime 13:21
nine Speshed bytecode can be a lot smaller than the original 13:22
That's the norm actually
14:08 Altai-man joined 14:10 sena_kun left 16:25 finanalyst joined
finanalyst [Coke] I've implemented a reset button as requested for the search options. Its working on new-raku. Tests still in progress on doc-website 17:51
[Coke] tests passing 18:11
Geth rakudo/main: 4 commits pushed by (Stefan Seifert)++ 18:43
lizmat 1058! 19:01
ab5tract heck yeah :) 19:12
Geth rakudo/main: d6dc3610ea | ab5tract++ (committed using GitHub Web editor) | src/core.c/Any.rakumod
Supply an Any.match method candidate for Any:D (#5559)

This addresses the use of non-`Str` candidates arriving in `where` constraints that attempt to match against a regex. (R#5464 / #5464).
  ```
... (32 more lines)		 19:25
linkable6 R#5464 [open]: github.com/rakudo/rakudo/issues/5464 [dispatching] Invocant of method 'match' must be a type object of type 'Any'
rakudo/main: 51b0ef2561 | (Elizabeth Mattijsen)++ | 2 files
Final tweaks in Any.match fixes

  - The JVM backend isn't smart enough to make the an "empty" proto
   disappear, so it needs an extre level of nqp::getlexcaller.
  - Int.match(foo) would still fail: fixed by making the Cool.match
   method a multi
  ab5tract++ for perseverance!		 19:47
ab5tract lizmat: shouldn't Int.match still fail though? 19:55
lizmat it warns, that should be enough? 19:56
m: dd Int ~~ Cool
camelia Bool::True
19:57 nativecallable6_ joined 19:58 reportable6 joined, sourceable6_ joined, notable6_ joined, unicodable6_ joined, coverable6_ joined, linkable6_ joined, bloatable6_ joined
ab5tract locally without your commit I get: `Invocant of method 'match' must be an object instance of type 'Cool', t a type object of type 'Int'. Did you forget a '.new'?` 19:58
19:59 tellable6_ joined
ab5tract that seems to match the expected behavior, no?\ 19:59
19:59 evalable6_ joined, greppable6_ joined
lizmat no, that was actually incorrect 19:59
19:59 squashable6 joined, releasable6_ joined
lizmat type objects stringify to the empty string with a warning 19:59
ab5tract ie, the `Any:U:` candidate that calls `self.Str`
19:59 statisfiable6 joined
lizmat m: say Int.Str 19:59
19:59 bisectable6_ joined
camelia Use of uninitialized value of type Int in string context.
Methods .^name, .raku, .gist, or .say can be used to stringify it to something meaningful.

in block <unit> at <tmp> line 1		 19:59
20:00 shareable6_ joined
lizmat m: my $a = Int.Str; dd $a 20:00
camelia Use of uninitialized value of type Int in string context.
Methods .^name, .raku, .gist, or .say can be used to stringify it to something meaningful.
in block <unit> at <tmp> line 1
Mu $a = ""
20:00 benchable6_ joined, committable6_ joined, quotable6_ joined
ab5tract ah, gotcha 20:00
thanks for the fix :) 20:01
lizmat hmmm on that thought....
m: say "" ~~ /^ $/ 20:02
camelia ｢｣
lizmat so it looks like it should set the empty match if that is what was requested
so we can actually simplify maters 20:10
matters
20:15 Altai-man left, sena_kun joined
Geth rakudo/main: 881c1e87a7 | (Elizabeth Mattijsen)++ | 2 files
Simplify the Any.match issue

There shouldn't really be a difference between instantiated or uninstantiated Any matching. Both should first stringify and then call .match on the result. If a class doesn't handle strinngification of its type object, this will be handled by Any.Str, which will warn and produce the empty string (which *can* potentially match).
Because this is now done by Any.match, Cool.match doesn't need to exist anymore as it will fall back to Any.match.		 20:18
20:26 committable6_ left 20:29 committable6_ joined 20:41 sena_kun left, sena_kun joined 21:02 sena_kun left 21:33 sena_kun joined 22:30 finanalyst left 23:06 sena_kun 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!