🦋 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.
00:00 reportable6 left 00:11 TempIRCLogger left, TempIRCLogger joined 00:18 shareable6 joined, quotable6 joined, greppable6 joined 00:19 notable6 joined, linkable6 joined 00:20 benchable6 joined 01:20 releasable6 joined 01:22 frost left 01:23 qorg11 left, qorg11 joined 01:40 Xliff left 02:08 ggoebel left 02:20 committable6 joined 02:35 drakonis joined 02:50 ggoebel joined
releasable6 Next release in ≈1 day and ≈15 hours. GitHub is down. There are no known blockers. Please log your changes in the ChangeLog: github.com/rakudo/rakudo/wiki/ChangeLog-Draft 03:00
04:01 reportable6 joined 04:33 frost joined 06:00 reportable6 left 08:48 squashable6 left, squashable6 joined 09:05 ggoebel left 10:01 reportable6 joined 11:01 reportable6 left, coverable6 left, bloatable6 left, nativecallable6 left, bisectable6 left, statisfiable6 left, releasable6 left, tellable6 left, linkable6 left, unicodable6 left, squashable6 left, evalable6 left, quotable6 left, notable6 left, benchable6 left, shareable6 left, sourceable6 left, greppable6 left, committable6 left 11:02 reportable6 joined, bloatable6 joined 11:03 greppable6 joined 11:04 quotable6 joined, benchable6 joined, bisectable6 joined
lizmat I think there's something weird going on with module loading: gist.github.com/lizmat/8b55f762472...7086f5ee63 11:42
in the first gist (after a re-compile of the setting) it strikes me that it tries to load the .repo-id from places it did *NOT* find the sha 11:43
gist.github.com/lizmat/8b55f762472...xt-L21-L23 11:44
in the second case, it claims that the repo has changed and that it needs to re-check dependencies
gist.github.com/lizmat/8b55f762472...xt-L22-L29 11:45
it does this *every* time
which implies to me something's amiss and fixing this would help the time spectests take 11:46
(and testing in general)
finally (probably related), gist.github.com/lizmat/8b55f762472...xt-L30-L31 shows that it actually loads a *different* sha 11:47
nine ugexe ^^
vrurg ^^
nine lizmat: trying to load .repo-id files from everywhere is perfectly normal. Maybe you're wondering what these files are in the first place? 11:52
lizmat perhaps, indeed :-)
I mean, aren't they supposed to come in pairs, residing in the same dir ? 11:53
the sha file, and the sha.repo-id file ? 11:54
nine It's like this: you've got module A installed in site and have a precomp file for it in site/precomp. How do you know that the precomp file is up to date? On the one side we have to check whether the dependencies' precomp files have changed since. But even if not, what if there is a newer version of one of the dependencies in ~/.raku?
lizmat but a precomp file can only change if the sha changes ?
isn't the precomp file supposed to be immutable for a given sha ?
nine So when we load a precomp file we have to look for any newer dependency in every repository of the repo chain. That's epensive, so what we do is we generate a SHA of all repositories' state and record that in the .repo-id files. If the SHA is still the same, we don't have to even check. 11:55
So whenever you change the repository chain, by e.g. running with -Ilib, we have to check those dependencies again and if the precomp file is still up to date, we can put a .repo-id file into lib/.precomp/... and know through this that the precomp file from site is still ok 11:56
lizmat ok, so .repo-sha would have been a better extension? :-) 11:57
nine Maybe. That would bake the implementation into the name. repo-id is about the identity of the repository chain (as in "is it identical to when we did the dependency check?") 11:58
A bit of a mathy wording
lizmat ok, gotcha yeah
so what about the changed dependencies ? 11:59
nine That will need a closer look. Would be nice to find that there is a bug that we can fix to improve module loading performance
12:00 reportable6 left
lizmat ok, will focus on that then for now :-) 12:00
12:00 reportable6 joined 12:02 squashable6 joined 12:03 releasable6 joined, linkable6 joined 12:04 nativecallable6 joined
Geth rakudo: 9425d0faaf | (Elizabeth Mattijsen)++ | lib/Test.rakumod
Test should work with any default version of the language
12:15 Geth left, Geth joined 12:19 Geth left 12:20 Geth joined 12:37 ggoebel joined
Geth rakudo: patrickbkr++ created pull request #4817:
Corect usage instructions of precompiled release archives
13:02 notable6 joined 13:03 statisfiable6 joined 13:04 evalable6 joined
Geth rakudo: 29fc65074a | (Patrick Böker)++ | 2 files
Corect usage instructions of precompiled release archives

The Linux and MacOS instructions contained Windows-y bits.
rakudo: 599e33efd0 | (Patrick Böker)++ (committed using GitHub Web editor) | 2 files
Merge pull request #4817 from patrickbkr/correct-bin-rel-instructions

Corect usage instructions of precompiled release archives
lizmat nine: OOC, have you ever used the tools/install-dist.p6 script to install modules that have a "build" section? 13:15
what I'm reading in the source atm, is that if it *has* a build, it will exit on successful bulding, and never actually run the install logic ?
*building 13:16
13:58 ggoebel left 13:59 ggoebel joined 14:55 frost left 15:03 Xliff joined 15:04 unicodable6 joined, shareable6 joined
ugexe Inline::Perl5 uses it 15:22
yeah looks like you would have to run it once with --build and once without 15:25
i think the exit on success may not have been intentional though 15:26
yeah that must be so, since build-only behavior would have its own path github.com/rakudo/rakudo/commit/4c...fa1e014a03 15:28
re changed dependencies, it wouldn't surprise me to find some module getting loaded that has `use lib ...` 15:30
although from the output you posted that isnt the case 15:32
16:01 committable6 joined 16:03 sourceable6 joined
ugexe looks like the behavior started between 2020.02.1 and 2020.07 (don't have 2020.05 build locally to test) 16:11
the 2020.05 release has like 20 commits labeled as "CompUnit::* optimizations"
if i were to guess it would be related to those 16:12
16:26 MasterDuke left 16:33 MasterDuke joined 17:03 coverable6 joined
Geth rakudo: edb8a200fd | (Stefan Seifert)++ | tools/install-dist.raku
Remove erroneous early exit from install-dist.raku

  --only build has its own MAIN sub, so it makes no sense to exit unconditionally
after building
nine lizmat++ ugexe++
Geth rakudo: 6e383c2044 | (Elizabeth Mattijsen)++ | tools/install-dist.p6
Same as edb8a200fd for the old script
18:00 reportable6 left
Geth Linenoise/main: 44 commits pushed by (Juan Julián Merelo Guervós)++, (JJ Merelo)++, (Elizabeth Mattijsen)++
review: github.com/raku-community-modules/...4b59a38424
19:03 linkable6 left
Geth Linenoise/main: 43725a7b10 | (Elizabeth Mattijsen)++ | 14 files
Changes so far
19:05 linkable6 joined
Geth Linenoise/main: 086a3e1b81 | (Elizabeth Mattijsen)++ | 2 files
CI tweaks
Linenoise/main: ff5d057eb9 | (Elizabeth Mattijsen)++ | 2 files
More CI tweaks
Linenoise/main: e0483ea522 | (Elizabeth Mattijsen)++ | .github/workflows/test.yml
Drop windows from test.yml
Linenoise/main: 98094e926b | (Elizabeth Mattijsen)++ | README.md
Hopefully fix badges
Linenoise/main: 779df5fa18 | (Elizabeth Mattijsen)++ | README.md
Fix URLencoding on badge link
Linenoise/main: 44eccfd534 | (Elizabeth Mattijsen)++ | 3 files
Various CI fixes (hopefully)
Linenoise/main: 13d2607ce4 | (Elizabeth Mattijsen)++ | Changes
20:01 reportable6 joined 20:21 sena_kun left 20:23 sena_kun joined 20:45 Xliff left 21:01 Xliff joined
Geth nqp: 3f1438e6e4 | (Daniel Green)++ | tools/templates/MOAR_REVISION
Bump MoarVM for a NativeCall+mimalloc fix
rakudo: 200579f702 | Coke++ | docs/windows.md
Stop telling users to avoid mimalloc

  github.com/MoarVM/MoarVM/pull/1686 resolved the issue
rakudo: 8d284d8c1e | (Daniel Green)++ | tools/templates/NQP_REVISION
Bump NQP for a NativeCall+mimalloc fix
21:26 Xliff left
releasable6 Next release in ≈19 hours. GitHub is down. There are no known blockers. Please log your changes in the ChangeLog: github.com/rakudo/rakudo/wiki/ChangeLog-Draft 23:00
23:20 linkable6 left, linkable6 joined 23:52 ggoebel left