🦋 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.
00:00 reportable6 left 00:01 reportable6 joined 00:41 [Coke] left 01:38 linkable6 joined 02:57 quotable6 joined 04:45 reportable6 left, linkable6 left, greppable6 left, quotable6 left, nativecallable6 left, notable6 left, bisectable6 left, coverable6 left, evalable6 left, squashable6 left, releasable6 left, benchable6 left, sourceable6 left, bloatable6 left, tellable6 left, shareable6 left, unicodable6 left, statisfiable6 left, committable6 left, reportable6 joined, greppable6 joined 04:46 coverable6 joined, releasable6 joined 04:47 notable6 joined, bisectable6 joined 04:48 tellable6 joined, squashable6 joined, sourceable6 joined 05:48 quotable6 joined 06:00 reportable6 left 06:01 committable6 joined
lizmat ugexe re Impossible coercion from BOOTInt into Int: value is of unacceptable type BOOTInt any idea where that happened? 06:31
06:50 shareable6 joined 06:51 nativecallable6 joined 06:53 benchable6 joined 06:54 statisfiable6 joined 06:55 unicodable6 joined 06:57 evalable6 joined 07:08 reportable6 joined
Geth rakudo/main: e36ac593fe | (Elizabeth Mattijsen)++ | src/Raku/Grammar.nqp
RakuAST: now that =finish is a directive

We no longer need to check for it explicitely
07:21 ab5tract left 07:58 linkable6 joined 08:22 sena_kun joined 09:04 bloatable6 joined
Geth roast: da2685d8d5 | (Elizabeth Mattijsen)++ | S26-documentation/04-code.t
Fix incorrect test for implicit code block

lizmat 885!
Geth rakudo/main: a05c4061ad | (Elizabeth Mattijsen)++ | src/core.c/RakuAST/Fixups.pm6
RakuAST: =cell will also honor implicit code blocks

09:24 sena_kun left 09:30 sena_kun joined 09:34 sena_kun left
Geth rakudo/main: cf29e292a6 | (Elizabeth Mattijsen)++ | src/core.c/Version.pm6
Streamline Version a bit

  - newdisp makes verbatim dispatch faster, so use that for the common
   occurring version values
  - remove nqp::stmts at the outer level, they're not needed
  - expand ))) for better readability
No functional changes, just some old maintenance and preparation for some new features.
12:00 reportable6 left 12:02 reportable6 joined
ugexe lizmat: gist.github.com/ugexe/c92c193064a3...f540b59265 12:06
lizmat multi method indent(Cool:D: $steps) { self.Str.indent($steps) } ?? 12:08
could it be it's going awry while trying to produce an error ?
Geth rakudo/main: 2184c5a83e | (Elizabeth Mattijsen)++ | 2 files
RakuAST: C<> markup implies verbatimizing

Basically move the flattening of V<> and C<> to "from-string" by first having the string parsed, and then verbatimizing any V<> and C<> markup codes. Adapt podifying accordingly.
ugexe I'm not sure. But if its producing an error before encountering this one then there is at least two errors 12:21
lizmat indeed :-(
ugexe i'm kind of assuming it would be related to rakuast in some way
lizmat well, typically some implicit conversions that are handled by MoarVM, aren't handled by the JVM backend 12:22
there's some conditionals in the core handling that
I would assume something new got added to the core that isn't properly "escaped" for the JVM backend 12:23
16:38 notna joined
gfldex If I would write `use Foo::Module; #= zef install Foo;`, could Rakudo display that hint instead of just complaining about unfindable modules? 17:31
17:32 sena_kun joined
ugexe im not sure that is a great idea, it could very well be the module is provided in a location that isn't being provided in the current e.g. RAKULIB 17:43
some CURs aren't "installs (CURFS), and some CURs don't need to install (CompUnit::Repository::Github) 17:47
it could make sense to add some note to the general message though, like "maybe you need to install Foo::Module" 17:50
i guess what i'm saying is I don't trust users to provide technically accurate messages, although we could come up with something generic 17:52
like if i saw "you may need to `zef install Foo`" i think thats fine, but "you need to `zef install Foo`" not so much 17:54
of course keeping those custom messages in sync with the version/api/auth would quite redundant, i.e. `use Foo:auth<xxx>:ver<1>:api<1>; #= zef install "Foo:auth<xxx>:ver<1>:api<1>"` 17:55
even then there might be quoting differences between e.g. linux and windows cmd.exe
perl -e 'use Dfasfd' 17:56
Can't locate Dfasfd.pm in @INC (you may need to install the Dfasfd module) (@INC contains:
18:00 reportable6 left 18:01 reportable6 joined
gfldex I was thinking of stand-alone scripts that might not be shipped via zef or other such means. It would be nice to allow them to display a message when even compiling the script fails, not just when a module is not found. I always found it rather cumbersome to hunt down deps for perl-scripts. 18:22
ugexe it honestly would probably be just as easy to provide a META6.json that just lists the dependencies of the stand alone script, treating it like a lock file 18:37
so the mechanism is already there in a way. do we really want multiple ways of doing this, that could also potentially conflict with each other? 18:40
maybe conflict isn't the best way to put it 18:41
19:47 notna left
Geth rakudo/main: 4a66191130 | (Nick Logan)++ | src/vm/moar/ModuleLoaderVMConfig.nqp
Remove nqp/lib from vm search paths

The removed comment suggests that this code shouldn't be needed anymore, since `:from<NQP>` was implemented at some point. The passing `make test` and `make spectest` also suggest this code is not needed anymore. However, to be clear, I'm not sure why the
  `:from<NQP>` adverb might make this library inclusion required.
rakudo/main: 64fef8049f | (Nick Logan)++ | src/vm/moar/ModuleLoaderVMConfig.nqp
Remove rakudo-home/lib from vm_search_path

rakudo-home/lib is provided elsewhere, and thus this search path isn't needed. This makes the vm_search_path return an empty array, similar to the JS backend. Note that we don't get rid of the vm_search_path method altogether because the JVM can return a non empty array of paths.
rakudo/main: 04bf4f719a | (Nick Logan)++ (committed using GitHub Web editor) | src/vm/moar/ModuleLoaderVMConfig.nqp
Merge pull request #5304 from rakudo/ugexe/remove-unneeded-nqp-lib

Remove nqp-home/lib and rakudo-home/lib from vm search paths
21:47 sena_kun left 21:59 [Coke] joined
ugexe lizmat: it looks like jvm stopped building with github.com/rakudo/rakudo/commit/e9...3f8101db6b 22:28
that was awhile ago though, so we can't infer if thats what the current error is from 22:29
i was able to rerun the ci test on that commit though so we should get to see what the first jvm error was (the original build log on that commit is gone on azure pipelines) 22:57
doh, its the java heap error that i only recently fixed 23:42
which i guess is good in that there might not be very much to fix to get the jvm to work, but bad in that we have no idea when the non-heap space error started