🦋 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.
00:07 reportable6 left 00:10 reportable6 joined
lizmat nine: OOC, could you explain why reverting 6b50a940af7721 fixes a problem? 00:44
00:44 linkable6 left
lizmat cause I'd like to understand what I did wrong! 00:45
00:45 linkable6 joined 01:42 vrurg left, vrurg_ joined 02:42 reportable6 left, bloatable6 left, evalable6 left, benchable6 left, releasable6 left, bisectable6 left, nativecallable6 left, shareable6 left, tellable6 left, quotable6 left, linkable6 left, notable6 left, sourceable6 left, greppable6 left, committable6 left, coverable6 left, unicodable6 left, statisfiable6 left, squashable6 left, statisfiable6 joined 02:43 sourceable6 joined, notable6 joined, benchable6 joined, bloatable6 joined, linkable6 joined, greppable6 joined, shareable6 joined, tellable6 joined 02:44 coverable6 joined, releasable6 joined, quotable6 joined, reportable6 joined, bisectable6 joined, nativecallable6 joined 02:45 committable6 joined, squashable6 joined, unicodable6 joined, evalable6 joined 03:12 vrurg joined, vrurg_ left 04:36 vrurg left 04:37 vrurg joined 04:46 vrurg_ joined, vrurg left 05:27 vrurg joined, vrurg_ left 06:07 reportable6 left 06:08 reportable6 joined 07:08 unicodable6 left, quotable6 left, bloatable6 left, committable6 left, reportable6 left, releasable6 left, linkable6 left, squashable6 left, tellable6 left, bisectable6 left, sourceable6 left, coverable6 left, shareable6 left, nativecallable6 left, benchable6 left, evalable6 left, greppable6 left, notable6 left, statisfiable6 left 07:09 linkable6 joined, sourceable6 joined, bisectable6 joined, squashable6 joined, tellable6 joined, quotable6 joined, releasable6 joined, nativecallable6 joined 07:10 coverable6 joined, shareable6 joined, statisfiable6 joined, reportable6 joined, committable6 joined 07:11 unicodable6 joined, evalable6 joined, benchable6 joined, notable6 joined, greppable6 joined, bloatable6 joined 07:56 vrurg_ joined, vrurg left 08:07 vrurg joined, vrurg_ left 08:17 vrurg_ joined, vrurg left 08:26 vrurg joined, vrurg_ left
nine lizmat: I'd have hoped the commit message explains it? Or did you mean reverting "Add CUR::Staging.candidates method"? 08:32
08:37 vrurg_ joined, vrurg left 08:46 vrurg joined, vrurg_ left 08:57 vrurg_ joined, vrurg left 09:09 sena_kun joined 09:13 frost joined
Geth rakudo/master: 5 commits pushed by (Ben Davies)++, (Christian Bartolomäus)++ 09:31
nine I still see errors like: Could not open /home/abuild/rpmbuild/BUILDROOT/perl6-Cro-TLS-0.8.7-2.206.x86_64/usr/share/perl6/vendor/sources/B2B08F2B66FB431BA3673FD5666577402EEDE723. Failed to stat file: no such file or directory 10:28
10:50 frost left
lizmat nine: ok, I think I see it now... but I think there is still an issue if there is a $*REPO in the call stack that is not PROCESS::<$REPO> ? 10:52
so shouldn't that be: my $precomp = self.precomp-repository; ?? 10:53
nine That's exactly what it is? my $precomp = self.precomp-repository; 11:00
To clarify, I get the "Failed to stat file: no such file directory" errors with "Add CUR::Staging.candidates method" restored, but one of the bugs in it fixed. 11:10
So there's yet another one.
And given the error message and what that commit does, I think it's at the heart of the implementation, the new candidate method. This method can now return dists that are contained in the shadowed repository. But candidates is also used by !matching-dist which is in turn used by need. 11:14
And need will take a candidate, assume it's one of its own (because why wouldn't it be) and construct paths like my $source-file-name := %meta<source>; my $loader := $.prefix.add('sources').add($source-file-name); 11:15
Of course this path does not point to a file because its using the wrong prefix. lizmat: how is this supposed to work?
lizmat hmm... that was not in the revert: I was looking at that, so that at least is correct 11:18
nine In which revert? 11:19
lizmat your commit message said it was a revert of 6b50a940af77211709ef87256 11:20
so I was looking at that, not at the current state of the CURIU code
nine No, that's the one commit that I didn't actually revert. I said that my commit was fixing a regression caused by 6b50a940af77211709ef87256
The rest of the packaging problems is caused by "Add CUR::Staging.candidates method" 11:21
11:21 linkable6 left
nine Which I did revert on master. I'm now trying to find out what's actually wrong with it and so far have found two issues and one of them makes me question the whole commit. 11:21
11:24 linkable6 joined
lizmat looking at 50025c1a48dd19 and not having all of that in my quick access memory, I'd say I reacted to a potential issue that ugexe spotted 11:26
nine But what was the plan for loading one of these additional candidates? 11:27
lizmat as it said in the commit message: The candidates should *also* take into account any candidates from the CURI repository it is shadowing. Spotted by ugexe++ 11:28
and as candidates is not supposed to just return the first, it would return all it could find ? 11:29
nine But how would the winning candidate be loaded if its from a different repo? You can't just have one repo load a candidate from a different repo
lizmat well, then the API of candidates is abused I guess... because .candidates is just supposed to return all candidates from all repos, is it not ? 11:30
nine I wouldn't know. candidates is not part of the original design. But it has been used for !matching-dist since it's introduction in commit 2d007bf431c173c70bc6f964c5cb8d2da0e5bdae 11:44
11:45 linkable6 left
lizmat not part of the design? I remember distinctly writing that design.perl6.org/S22.html#candidates 11:47
nine It isn't mentioned in github.com/rakudo/rakudo/blob/mast...agement.md 11:48
11:48 linkable6 joined
lizmat well, it is mentioned in S22, which is from 2014 11:51
11:52 sena_kun left
nine Anyway as implemented, this just cannot work. 11:54
11:54 sena_kun joined
nine Btw. the commit actually introduces two totally unrelated features and the commit message explains for neither why they would be needed. 12:04
12:07 reportable6 left 12:08 reportable6 joined
lizmat clearly I do not understand the problem, so I'm recusing myself from the CUR* namespace 12:08
if you feel it is better to revert all of the work I've done there, please do! 12:09
nine Do we leave "Add CUR::Staging.candidates method" reverted then?
lizmat by all means, If you think it is the better way
as I said, I'm recusing myself from the CUR* namespace
nine Well I don't have a good idea to make it work as intended. But maybe ugexe does. 12:10
lizmat Well, I thought I had an idea, but clearly I don't understand the problem space properly
meanwhile, Inline::Perl5 tests are now skipped when spectesting on master while Inline::Perl5 is installed 12:19
can't reinstall either :-( gist.github.com/lizmat/28bc69c2a31...179baded45 12:22
nine ^^
nine But what's the failure? 12:38
zef --debug? 12:39
lizmat gist updated
nine That output is not helpful :/ Can you try to build it manually? 12:42
lizmat nine: gist.github.com/lizmat/01892c51416...26ba8bd023 12:49
nine Doesn't like nice but it's just warnings 12:50
lizmat just added make install output 12:51
nine Can you please try rakudo /path/to/your/rakudo/tools/install-dist.p6 --to=home 12:55
lizmat only output: make: Nothing to be done for `all'. 12:56
seems to have installed
nine No output is actually good.
Then I have no idea what zef's problem is 12:57
lizmat pretty sure it used to work until a few rakudo commits ago... want me to check that ? 13:01
nine Please, yes. Would be nice to get to a stable state again.
lizmat % zef uninstall Inline::Perl5 13:04
===> Uninstalled from inst#/Users/liz/.raku
that's correct, right?
nine yep 13:06
13:08 linkable6 left, evalable6 left
lizmat on 90709d5334d59f2e292fd "make install" fails 13:08
13:10 evalable6 joined 13:11 linkable6 joined
lizmat looks like installing Inline::Perl5 stopped working at 7342d9dc28f7f3056f574b9f841064e2f396eb7d 13:18
at least
13:19 linkable6 left
lizmat I don't want to go further back, as that will completely screw up my local installation :-( 13:19
13:21 linkable6 joined, sena_kun left
nine at or before? 13:53
For such tests I just use a separate installation. 13:54
lizmat at
nine That commit brings just one nqp commit which is a MoarVM bump which brings the fallback resolver and Add missing :deoptallpoint flags to invokish ops 13:56
lizmat right 13:57
nine No hypothesis how any of that can break Inline::Perl5. Much less only on your system.
lizmat: can you try going back on those individual MoarVM commits? 14:00
lizmat well, if it is only me, I can live with that
no, not right now: trying to wrap up App::Rak 0.1 at the moment... and I don't want to deal with further breakage at 14:01
maybe tomorrow after the Weekly 14:03
14:16 dogbert17 joined 15:36 vrurg_ left 15:37 vrurg joined 16:47 vrurg_ joined, vrurg left 17:39 hythm joined
hythm Speaking of "Add CUR::Staging.candidates method" commit, Staging Inline::Perl5 started failing for me and complaining about not able to find `dist` file. the error was coming from `Inline::Perl5::Interpreter` module. 17:42
so I created a minimal example of `Inline::Perl5::Interpreter` to minimize the code. now after reverting "Add CUR::Staging.candidates method" staging Inline::Perl5 started working again, but the minimal example , now showing different error:
`Cannot locate native library '/tmp/workdir/stage/resources/lib705D783E180D440A8BE0A0832525197FF9B9FF6C..so'`
looks like it is calling `platform-library-name` on the already compiled resource. Its confusing so I'm not sure if it's a genuine bug or I messed something when creating the minimal example of `Inline::Perl5::Interpreter`. more information is on this comment github.com/rakudo/rakudo/issues/49...1261577747 17:43
18:06 reportable6 left 18:07 reportable6 joined 18:40 sena_kun joined 19:11 sena_kun left, sena_kun joined 19:14 Altai-man joined 19:16 sena_kun left 19:28 vrurg joined, vrurg_ left 19:43 epony left 19:52 vrurg_ joined, vrurg left
Geth nqp: ebff3c7aba | (Christian Bartolomäus)++ | 2 files
[JVM] Handle unicode numbers in nqp::radix(_I)
nqp: bddcbca4b7 | (Christian Bartolomäus)++ | src/vm/jvm/runtime/org/raku/nqp/runtime/Ops.java
[JVM] Support unicode numbers in Rakudo better

This is everything but a proper implementation of getuniprop_str and unipropcode. But it makes the evaluations mentioned in
  github.com/rakudo/rakudo/issues/4760 work on the JVM backend.
20:37 epony joined
nine hythm: it should work fine on rakudo master 20:50
hythm nine: rephrasing my previous statements, this gist gist.github.com/hythm7/d41cdd318cb...66786c1e20 is an attempt to reproduce an issue I saw during staging Inline::Perl5. 21:33
at this commit 50025c1a "Add CUR::Staging.candidates method" it gives this error: "Failed to open file /home/hythm/.raku/dist/0A9FF6017CA75303FC6271D4E739E7D0C7AC83FB: No such file or directory". on master (just built now) gives a different error "Cannot locate native library
'/tmp/workdir/stage/resources/lib705D783E180D440A8BE0A0832525197FF9B9FF6C..so': /tmp/workdir/stage/resources/lib705D783E180D440A8BE0A0832525197FF9B9FF6C..so: cannot open shared object file: No such file or directory".
so staging Inline::Perl5 at "Add CUR::Staging.candidates method" commit fails, on master works. while, the gist mentioned above at "Add CUR::Staging.candidates method" fails, and on master, gives a different error. so I'm trying to know if its a genuine bug, or I just messed something when making the minimal example (of Inline::Perl5::Interpreter)
in the gist.
21:34 linkable6 left 21:36 linkable6 joined 22:36 linkable6 left, evalable6 left 22:37 evalable6 left 22:38 evalable6 joined 22:39 linkable6 joined 22:48 Altai-man left 23:04 epony left