🦋 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: ... | log inspection situation still under development | For MoarVM see #moarvm
Set by lizmat on 22 May 2021.
00:02 reportable6 left 00:04 reportable6 joined
Xliff_ sena_kun: Are you talking about #4570? It's still open. 00:06
00:14 linkable6 joined 00:56 frost joined 02:22 ggoebel left 03:39 linkable6 left, quotable6 left, greppable6 left, statisfiable6 left, reportable6 left, evalable6 left, squashable6 left, bisectable6 left, notable6 left, sourceable6 left, committable6 left, nativecallable6 left, benchable6 left, coverable6 left, shareable6 left, tellable6 left, unicodable6 left, releasable6 left, bloatable6 left, nativecallable6 joined, tellable6 joined, benchable6 joined, notable6 joined 03:40 statisfiable6 joined, sourceable6 joined, committable6 joined 03:42 reportable6 joined, bisectable6 joined, shareable6 joined 04:39 releasable6 joined, evalable6 joined 04:42 linkable6 joined, coverable6 joined, unicodable6 joined 05:39 bloatable6 joined 05:40 MasterDuke left 05:41 quotable6 joined, squashable6 joined 06:02 reportable6 left 06:03 reportable6 joined 07:12 MasterDuke joined 07:41 greppable6 joined
Xliff_ sena_kun: segfault (#4570) is still occurring 08:47
(BTW...Good morning, #cro -- how rude of me!)
09:29 linkable6 left, evalable6 left 09:34 Kaiepi left, Kaiepi joined
Kaiepi o/ 09:54
pr for #4558's going up soon
while working on it i looked into what needs doing in order to get generic return values in signatures 09:56
spectest seems favourable to Metamodel::GenericHOW delegating method calls to Mu and accepting all typechecks on the RHS of a smartmatch, but the compiler's doesn't do method calls on generics (easy fix), and "my ::T $" doesn't install a symbol for T at the very least 09:58
and role typechecking becomes a problem again 09:59
s/compiler's/compiler/ 10:00
[Tux] Rakudo v2021.09-267-g8d1f24f9c (v6.d) on MoarVM 2021.09-641-gf7d6bc614
csv-ip5xs1.459 - 1.578
csv-ip5xs-2016.449 - 17.809
csv-parser4.901 - 5.368
csv-test-xs-200.385 - 0.390
test7.484 - 7.737
test-t1.729 - 1.856
test-t --race1.100 - 1.102
test-t-2025.852 - 26.226
test-t-20 --race8.651 - 9.504
Kaiepi forgot to mention, generic return values don't work because T !=:= the comparand's type in the last line of Signature.ACCEPTS
besides all that 10:03
Geth rakudo: Kaiepi++ created pull request #4573:
Fix handling of constrained Mu parameters in signature smartmatching
roast: Kaiepi++ created pull request #760:
Add more tests for type constraints in signature smartmatches
10:31 evalable6 joined 10:33 ggoebel joined
lizmat sometimes I miss something like: 10:50
sub foo($a //= 42)
aka, if $a is not specified, or it received a non-defined value, then assign the default 10:51
similarly for attributes:
has $foo //= 42
sena_kun releasable6, status 10:55
releasable6 sena_kun, Next release in ≈8 days and ≈8 hours. 5 blockers. Changelog for this release was not started yet
sena_kun, Details: gist.github.com/ece92e268cee239124...bbb84e507d
11:05 ggoebel left 11:27 squashable6 left
Kaiepi lizmat, with metaops for assignments in parameters, maybe `sub foldr(&f, \init, +result [**@xs is raw] [R[[&f]]]= |@xs, init) { result }` would be valid? 11:32
could make for some weird golfs
lizmat I was just thinking about //= and possibly //:= 11:33
Kaiepi fair 11:34
lizmat has $foo := 42 as shorthand for: $foo is built(:bind) = 42
as another syntactic sugar 11:35
Kaiepi that would be nice
lizmat I think we need to wait for RakuAST to become more fleshed out, but yes :-)
Kaiepi forgot slurpies would need defaults in order to write foldr in that way 11:41
m: sub foldr(&f, \init, +result [**@xs is raw] = [R[[&f]]] |@xs, init) { result }
camelia 5===SORRY!5=== Error while compiling <tmp>
Cannot put default on slurpy parameter result
at <tmp>:1
------> 3lt [**@xs is raw] = [R[[&f]]] |@xs, init7⏏5) { result }
expecting any of:
Kaiepi *$ is another feature i've seen in the design docs that would be cool to have when RakuAST comes around 11:42
lizmat what was that again ? 11:43
feels linenoisy though
Kaiepi yeah... 11:44
:(**@head, *$tail)
:(**@init, tail)
*$tail 11:45
12:02 reportable6 left 12:03 reportable6 joined
lizmat github complains about this for one of my modules: /Users/runner/hostedtoolcache/rakudo/2021.09-01/x86_64/bin/rakudo is loading libcrypto in an unsafe way 12:03
on MacOS... any idea what that is about? 12:04
hmmm... probably Cro related, should probably ask there
MasterDuke wonder if we should switch away from sha1 duo.com/decipher/sha-1-fully-and-p...-collision 12:45
lizmat you mean in core? 12:56
seems like a very unlikely attack vector ?
.oO( famous last words :-)
MasterDuke yeah. sure it's unlikely, but it probably wouldn't be too terribly difficult to switch to something like sha256 13:04
lizmat well, implement a nqp::sha256 op then, :-) 13:06
and then *that* could be potentially exposed at the language level ? 13:07
13:32 linkable6 joined 13:41 squashable6 joined 13:53 frost left
timo why stay at 256 when there is sha3 14:18
lizmat CPU? 14:22
timo hm. 14:34
15:13 [Tux] left 15:22 [Tux] joined, vrurg left
leont sha3 isn't necessarily better or worse than sha2 15:40
lizmat and definitely not for what we would use it for in the core, now? 15:42
leont has no idea what it's used for in core 15:43
ugexe used a lot for generating ids that are safe filenames 15:46
lizmat and for some .WHICH cases 15:50
m: dd (1,2,3).Set.WHICH 15:51
camelia ValueObjAt.new("Set|C31107F1B1119ABCDC812ED45D2368222884B304")
16:08 AlexDaniel left, CIAvash left 16:12 AlexDaniel joined 16:20 CIAvash joined
Xliff_ Has there been any movement on #4570? 16:36
16:48 melezhik joined 17:20 vrurg joined
Xliff_ Temporary workaround found, please see github.com/rakudo/rakudo/issues/4570 17:22
Maybe that will help with finding the root cause of the SEGV?
18:02 reportable6 left, reportable6 joined
nine Xliff_: it's a tough one. I'm reporting progress in #moarvm 18:50
Xliff_ nine: Thanks. 19:35
19:36 Geth left, Geth joined 21:12 linkable6 left, evalable6 left 21:14 evalable6 joined 21:15 linkable6 joined
Geth rakudo: 33296d4065 | (Nick Logan)++ (committed using GitHub Web editor) | src/core.c/CompUnit/Repository/FileSystem.pm6
Use distribution specific comp unit ids in CURFS (#4572)

Previously CU::R::FileSystem would only use the short name when generating a comp unit id. However when two different versions of the otherwise same module were loaded the same precomp file would get used. This changes to use the distribution id (which already accounts for version/api/auth) in combination with the short name of the requested module.
23:39 linkable6 left