🦋 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.
ugexe so doing a diff with `RAKUDO_MODULE_DEBUG=1 raku -Ilib t/spec/S10-packages/basic.t 2> new.out` with and without the commit does show its precompiling more with my change 00:15
that uses 5 `use lib ...`
gist.github.com/ugexe/63267181f024...-change-L9 00:18
in the rakudo code before my changes that line says 1 RMD: Settings CORE.c loaded 00:20
ugexe i suppose its quite possible it was cheating before and its doing the right thing now but im not doing a very good job of confirming or denying that 01:22
Geth rakudo: usev6++ created pull request #4576:
[JVM] Provide clean fix for broken build
06:35
Geth rakudo: 8fc643ec95 | (Christian Bartolomäus)++ | 2 files
[JVM] Provide clean fix for broken build

There was no need to add the special cases to src/Perl6/World.nqp.
The adjustment to p6captureouters2 was inspired by 02c13fad8d. Again thanks to jnthn++ for the pointer.
09:35
Geth rakudo/revert-curfs-fixes: 810b953e36 | (Elizabeth Mattijsen)++ | src/core.c/CompUnit/Repository/FileSystem.pm6
Revert CurFS fixes for now

Until we can figure out why this is causing the spectest to take about 2x as long (from 357 -> 616 wallclock for yours truly)
10:06
rakudo: lizmat++ created pull request #4577:
Revert CurFS fixes for now
ugexe github.com/Raku/roast/blob/40756ae...sic.t#L278 this test is surprisingly not agonizingly slow 14:17
Geth rakudo/new-disp-nativecall: 11 commits pushed by (Stefan Seifert)++
review: github.com/rakudo/rakudo/compare/7...54b6a7c207
17:36
Geth rakudo/ugexe-patch-3: 7f3aaf8da1 | (Nick Logan)++ (committed using GitHub Web editor) | src/core.c/CompUnit/Repository/FileSystem.pm6
Ensure all methods use the updated comp-unit-id

In a recent change the distribution id was added to the module name when generating a comp-unit-id for `need` to allow multiple versions of the same name to be loaded. However,`need` and `resolve` need to generate identical comp-unit-ids for a given module name since `resolve` will be called by CUR::PrecompilationRepository when `need`ing a module.
18:34
rakudo: ugexe++ created pull request #4578:
Ensure all methods use the updated comp-unit-id
MasterDuke ugexe: is ^^^ "just" a correctness fix, or does it also address the slowdown? 18:59
releasable6 Next release in ≈4 days and ≈23 hours. 1 blocker. Please log your changes in the ChangeLog: github.com/rakudo/rakudo/wiki/ChangeLog-Draft 19:00
ugexe both. i dunno how much faster the spectest is yet but it was faster for the one `use lib` heavy test i was using to benchmark. fwiw i would expect my changes to be slightly slower than before, but not more than slightly 19:04
MasterDuke i just tested it and a spectest is now back around what it was before the curfs fix (i.e., after the new-disp merge) at ~175s 19:20
ugexe cool, i'll merge it then 19:27
MasterDuke (and the spectest also passed) 19:28
Geth rakudo: bc2cac1366 | (Nick Logan)++ (committed using GitHub Web editor) | src/core.c/CompUnit/Repository/FileSystem.pm6
Ensure all methods use the updated comp-unit-id (#4578)

In a recent change the distribution id was added to the module name when generating a comp-unit-id for `need` to allow multiple versions of the same name to be loaded. However,`need` and `resolve` need to generate identical comp-unit-ids for a given module name since `resolve` will be called by CUR::PrecompilationRepository when `need`ing a module.
19:32