🦋 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.
00:03 reportable6 left 00:04 reportable6 joined 00:23 Kaiepi joined 01:17 vrurg joined 01:20 vrurg_ left 02:11 vrurg_ joined 02:14 vrurg left 03:22 unicodable6 left, bloatable6 left, reportable6 left, shareable6 left, squashable6 left, quotable6 left, greppable6 left, coverable6 left, releasable6 left, committable6 left, statisfiable6 left, sourceable6 left, nativecallable6 left, notable6 left, evalable6 left, bisectable6 left, tellable6 left, benchable6 left, linkable6 left 03:24 tellable6 joined, statisfiable6 joined, linkable6 joined, squashable6 joined, coverable6 joined, greppable6 joined, vrurg_ left, sourceable6 joined 03:25 vrurg joined 04:25 reportable6 joined, evalable6 joined, quotable6 joined, notable6 joined, nativecallable6 joined, bisectable6 joined 05:03 frost joined 05:23 releasable6 joined 05:24 bloatable6 joined 05:25 unicodable6 joined 06:02 reportable6 left 06:04 reportable6 joined 06:23 benchable6 joined 06:25 shareable6 joined 06:26 committable6 joined 07:33 Xliff left
Geth roast: 8506e372d1 | (Christian Bartolomäus)++ | S06-advanced/dispatching.t
[JVM] Fudge some tests for dispatching (todo)
roast: usev6++ created pull request #764:
[JVM] Don't test encoding/decoding for windows-932
roast: 6c906f15b1 | (Christian Bartolomäus)++ | S32-encoding/registry.t
[JVM] Don't test encoding/decoding for windows-932
roast: 3049084f5d | (Christian Bartolomäus)++ | S32-io/io-handle.t
[JVM] Fudge test for TTY detection (todo)

The op nqp::isttyfh does only guess if a handle is a TTY, and is not even implemented for FileHandles. As noted in
  github.com/Raku/nqp/commit/765ef2c10a there is no clean way
to implement a fully correct version.
[Tux] Rakudo v2021.10-1-g6dc3a4700 (v6.d) on MoarVM 2021.10-1-g07060faf1
csv-ip5xs1.389 - 1.410
csv-ip5xs-2015.761 - 16.194
csv-parser4.402 - 4.447
csv-test-xs-200.368 - 0.370
test7.057 - 7.235
test-t1.636 - 1.641
test-t --race0.960 - 0.981
test-t-2023.961 - 24.322
test-t-20 --race7.770 - 8.073
09:57 linkable6 left, evalable6 left, linkable6 joined
Geth nqp: 78d087815b | (Christian Bartolomäus)++ | t/nqp/059-nqpop.t
[JVM] Run tests for with/without on JVM backend

These tests have been working since commit ad700a25f9, but I forgot to unskip them.
nqp: aa1160e40e | (Christian Bartolomäus)++ | t/nqp/019-file-ops.t
Remove check for non-existing backends
nqp: 0295c671cf | (Christian Bartolomäus)++ | t/nqp/116-streaming-decoder.t
[JVM] Run test for translation of newlines

The translate_newlines config has been implemented by
  MasterDuke++ with 0412c22303 and c482a31526 earlier this year.
The test passes now.
10:37 Xliff joined 10:58 evalable6 joined
Geth nqp: usev6++ created pull request #743:
[JVM] Run pod test on JVM backend
11:58 linkable6 left, evalable6 left 12:00 linkable6 joined 12:01 evalable6 joined 12:02 reportable6 left 12:04 reportable6 joined 12:23 Xliff left 13:23 linkable6 left, evalable6 left 14:25 linkable6 joined
Geth nqp/new-disp-nativecall: 1f6de10695 | (Stefan Seifert)++ | 4 files
API for asking whether the compiler supports a certain nqp op

This can be used to conditionally compile backend specific code in modules like NativeCall
nqp/fix-native-result-dispatch: 9ba4f1e51d | (Stefan Seifert)++ | src/vm/moar/QAST/QASTRegexCompilerMAST.nqp
Fix compilation of dispatch to routines with non-standard native result types

For non-64-bit (or unsigned) native result types we need to use a 64 bit temp register as required by the op and coerce it to the register kind prescribed by the called routine's return signature.
rakudo/fix-native-result-dispatch: 7b7d564bfe | (Stefan Seifert)++ | src/Perl6/Optimizer.nqp
Fix unnecessary boxing by routines with native return types

Commit 3cc9d765b2b350c9d15d0164ed53a9914b333afb caused a regression that made us box natively typed return values of called routines. Before this the optimizer copied the return type from the routine to the call QAST::Op node. After the type got added to the QAST::Want node instead. This made code-gen always emit invoke_o/dispatch_o instead of natively typed versions.
Fix by copying the type to both the QAST::Op and QAST::Want nodes. However we don't do so for lvalue routines as those must return containers that can be assigned into.
nqp: niner++ created pull request #744:
Fix compilation of dispatch to routines with non-standard native resu…
rakudo: niner++ created pull request #4589:
Fix unnecessary boxing by routines with native return types
14:53 MasterDuke left 15:03 MasterDuke joined
Geth nqp: ffe7007a59 | (Christian Bartolomäus)++ | t/nqp/114-pod-panic.t
[JVM] Unskip block for JVM: execname is available

  ... now (implemented with 755fa3f97c).
nqp: 34c67c0109 | (Christian Bartolomäus)++ | t/nqp/114-pod-panic.t
[JVM] Run test for known bad pod on JVM

The test passes locally for me. Let's see if the CI likes it, too.
16:25 evalable6 joined 16:27 dogbert17 left 16:42 dogbert17 joined 18:02 reportable6 left 18:04 reportable6 joined 19:04 linkable6 left, evalable6 left 19:05 linkable6 joined 19:08 vrurg_ joined 19:11 vrurg left 20:14 vrurg joined, vrurg_ left 21:14 linkable6 left, linkable6 joined
japhb lizmat: logs.liz.nl/raku-dev/2021-10-24.html is returning 500, even though I'm having no trouble with other channels. 21:26
22:13 vrurg_ joined, vrurg left, vrurg_ left 22:14 vrurg joined
lizmat japhb: please make an issue for it... :-) 23:02
23:05 evalable6 joined