🦋 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.
[Tux] Rakudo v2022.03-73-g15abaf412 (v6.d) on MoarVM 2022.03-1-gefff4cd31
csv-ip5xs0.765 - 0.792
csv-ip5xs-204.926 - 6.589
csv-parser3.680 - 3.710
csv-test-xs-200.401 - 0.401
test6.535 - 6.568
test-t1.359 - 1.393
test-t --race0.807 - 0.839
test-t-2020.388 - 22.028
test-t-20 --race6.476 - 6.694
07:29
5 2022-03-07 17:53:18 test-t 1.402 07:30
4 2022-03-07 17:51:38 test-t 1.399
3 2022-03-24 08:25:36 test-t 1.393
2 2022-03-22 15:10:25 test-t 1.386
1 2022-03-24 08:27:53 test-t 1.359
[Tux] whistles
lizmat Files=1351, Tests=117057, 289 wallclock secs (35.04 usr 9.84 sys + 4046.55 cusr 326.50 csys = 4417.93 CPU) 09:25
lizmat MasterDuke: looks like the YAMLish failure happens after merging #4807 10:22
"Cannot assign to a readonly variable or a value" at Str.pm6 line 1815 10:23
MasterDuke ha, guess that `try` was needed. what's the YAMLish code doing that causes that? 10:24
lizmat argh, I just removed that... 10:28
MasterDuke ?
lizmat basically, you need to run the test with --ll-exception *and* disable the CATCH in load-parse 10:29
s/load-parse/load-yaml
MasterDuke well, jnthn did suggest testing with a substitution in an action method, guess my example that worked wasn't good enough 10:36
lizmat I don't think YAMLish uses separate actions? 10:39
MasterDuke github.com/Leont/yamlish/blob/mast...h.pm6#L500
adding back the `try`s fixes the YAMLish test, i'll PR that in a minute 10:40
lizmat stands corrected :-) 10:43
but isn't that just covering up the issue ?
MasterDuke well, maybe your suggestion of changing to `if nqp::iscont(cds)` is better 10:45
i'll give that a try
nope 10:49
same error in YAMLish test 10:50
ah, but looks like maybe `if nqp::isrwcont(cds)` does the trick 10:53
lizmat interesting... I always assumed that nqp::iscont implied assignabiluty... guess we need to look at more places that check like that 10:57
MasterDuke this will be the only use of isrwcont in the src/core.* 11:00
only a couple other uses elsewhere
leont YAML has a habit of exposing weird issues by virtue of being a far more complicated grammar than most 11:06
Geth rakudo: MasterDuke17++ created pull request #4830:
Only assign to caller's $/ during a regex subsititution if it's writable
11:46
lizmat MasterDuke: looks like e.g. TEMP-LET-LOCALIZE is also using iscont incorrectly (line 441 in Hash.p6) 11:56
Proc::Async!pipe and !wrap-decoder
Regex.ACCEPTS 11:57
Hash::Object.TEMP-LET-LOCALIZE 11:58
Map.ASSIGN-KEY 11:59
MasterDuke: maybe I should just do all of them myself ? :-)
MasterDuke i'm going to be afk for a while, won't stop you 12:00
lizmat ok, will do 12:02
Geth rakudo: aed27864f6 | MasterDuke17++ (committed using GitHub Web editor) | src/core.c/Str.pm6
Only assign to caller's $/ during a regex subsititution if it's writable (#4830)

  github.com/rakudo/rakudo/pull/4807 removed a `try` before the
assignment that was added in 14c1a6f58a73ed0ab92f3fb379e332f682afa916 because I couldn't find an example that needed it. However, that broke YAMLish (it was doing a subsititution in an actions method). However, a
  `try` is overkill here, and we can just make the cheaper check for
writability.
12:12
Geth rakudo: 90d876a3fd | (Elizabeth Mattijsen)++ | 7 files
Fix incorrect use of nqp::iscont

In these places, nqp::iscont was being used to see if the target was a writeable container. nqp::iscont only guarantees there's a container. nqp::isrwcont guarantees a writeable (assigneable) one.
12:26
problem-solving: 81cfdd3a23 | (Elizabeth Mattijsen)++ (committed using GitHub Web editor) | solutions/meta/sunsetting-p6c-cpan.md
Resolution: Preparing the Raku Ecosystem for the Future (#318)

Resolution document for problem solving issue #316
  github.com/Raku/problem-solving/issues/316
13:03
japhb Looks like the morning rakudo commits helped -- was able to do a full rebuild-all just now. 16:08
vrurg docs.raku.org is down... ping rba 19:36
maettu rba is tring to reach people
vrurg Oh, back up again
Thanks! 19:37
Not really. Cloudflare 'always live' feature. 19:38
Geth rakudo: f66b598132 | (Elizabeth Mattijsen)++ | src/core.c/Distribution/Path.pm6
Streamline creation of Distribution::Path

  - fewer IO -> Str -> IO conversions
  - no intermediate hashes to create
  - make it a more normal object using new Raku features
  - $!meta-file now has IO of default file, instead of being unset
  - expose .meta-file: it's also shown in .raku, so makes sense
  - add more gradual typing
20:13
Geth rakudo: fde422c602 | (Elizabeth Mattijsen)++ | src/core.c/Distribution/Path.pm6
Distribution::Path.new should only take IO::Path:D

Spotted by MasterDuke++
22:54