lizmat Files=1336, Tests=113586, 221 wallclock secs (29.69 usr 8.23 sys + 3109.45 cusr 284.38 csys = 3431.75 CPU) 08:35
Geth rakudo/IO-Notification-Change-ValueObjAt: ad92787103 | (Elizabeth Mattijsen)++ | src/core.c/IO/Notification.pm6
Simplify .WHICH creation

It's all about uniqueness more than readability.
09:22
rakudo: 17f352f510 | (Elizabeth Mattijsen)++ | src/core.c/IO/Notification.pm6
Make IO::Notification::Change a value type

Which should make it better to filter for it, e.g. with Supply.unique(expires => X).
09:57
rakudo: ad92787103 | (Elizabeth Mattijsen)++ | src/core.c/IO/Notification.pm6
Simplify .WHICH creation

It's all about uniqueness more than readability.
rakudo: 0e7ec139b1 | (Elizabeth Mattijsen)++ (committed using GitHub Web editor) | src/core.c/IO/Notification.pm6
Merge pull request #3905 from rakudo/IO-Notification-Change-ValueObjAt

Make IO::Notification::Change a value type
Geth rakudo: 3ba803d214 | (Christian Bartolomäus)++ | 2 files
Start to use nqp::execname on the JVM backend

The op is now available in Rakudo. (It's only NQP on the JVM backend that still needs a bootstrap.)
09:58
rakudo: 49f3d433c7 | (Elizabeth Mattijsen)++ (committed using GitHub Web editor) | 2 files
Merge pull request #3904 from usev6/jvm_nqp_execname

Start to use nqp::execname on the JVM backend
Geth rakudo/precomp_store_parallelism: ea7aaf9497 | (Stefan Seifert)++ | src/core.c/CompUnit/PrecompilationStore/File.pm6
Fix race conditions when writing .repo-id files

CompUnit::PrecompilationRepository.precompile not only writes the precomp file but also an accompanying .repo-id file, but it only locked the precomp file itself. This left the repo-id file open to race conditions ending in an empty repo-id file which led to explosions later on.
Turns out the new locking scheme was a bit too granular after all. Using the same lock file for precomp file and repo-id file (while still not affecting other precomp units) takes care of the race condition without necessitating a change in the API.
12:08
rakudo/rakuast: ed13652efb | (Jonathan Worthington)++ | 2 files
RakuAST compilation of .WHAT, .HOW, et al
12:09
Geth rakudo/rakuast: 8b53825f9c | (Jonathan Worthington)++ | 3 files
Compile a few more cases of numeric literals
12:44
Geth rakudo/rakuast: 3fac0d0ab7 | (Jonathan Worthington)++ | 9 files
Preparations for BEGIN-time code evaluation

With RakuAST we'll:
  * Interpret simple things
  * Send the rest through the compilation pipeline
... (12 more lines)
13:37
AlexDaniel [Coke]: how will you publish the results? Will it be all results, or just the winners? 13:38
[Coke] list of emails who voted, final counts. invididual ballots will not be reported. 13:41
*individual 13:42
You can see the dumb WIP script I have here: github.com/Raku/Raku-Steering-Coun...tally.raku 13:43
basically the output of that script.
Xliff Who is in the Raku Sterring Council?
[Coke] That's what the vote is to find out 13:44
We're coming up on an initial vote to decide the first group 13:45
jdv79 will there be minutes? i'm not fond of the "behind closed doors" stuff. 13:51
[Coke] I'm just tallying the votes 13:55
other information is at the github repo (look under papers) - please open an issue on that if you don't think it's addressed.
timotimo that last rakuast commit almost looks like someone could go ahead and add implementations for a couple of ast nodes to be interpreted ... 13:56
Xliff Ah. OK. Good to know that I should vote. 14:26
jdv79 i imagine it'll be in an upcoming weekly. i almost forgot about it myself. 14:28
lizmat it was in the last weekly: rakudoweekly.blog/2020/08/31/2020-...ndidacies/ 14:29
jdv79 i know. except the voting period is in the future so easy to forget;) 14:31
lizmat The voting period will be mentioned prominently in at least RWN issues :-) 14:36
*two
[TuxCM] Rakudo version 2020.08.2-49-g49f3d433c - MoarVM version 2020.08-91-g590bac47e
csv-ip5xs0.896 - 0.959
csv-ip5xs-209.902 - 9.989
csv-parser25.601 - 26.151
csv-test-xs-200.391 - 0.396
test7.607 - 7.770
test-t1.847 - 1.957
test-t --race0.817 - 0.847
test-t-2031.890 - 32.266
test-t-20 --race9.044 - 9.826
14:41
Geth rakudo/rakuast: 54a9ffd875 | (Elizabeth Mattijsen)++ | 9 files
Make "make test" pass

By adding the new identifiers of RakuAST
15:17
timotimo make "make "make test" test" pass 15:21
nine So RakuAST passes make test? Ship it! 15:22
timotimo AFAIK it already only acts when you set an env var or so 15:24
Xliff `rakudobrew build moar rakuast` 15:28
Actually, I probably want moar-blead
timotimo i think nothing especially fresh from moar is needed 15:29
Xliff rut-roh 15:30
cp: will not overwrite just-created '/home/cbwood/Projects/rakudobrew/versions/moar-rakuast/install/share/perl6/lib/Perl6/Grammar.moarvm' with 'blib/Raku/Grammar.moarvm'
That looks like an error with the build system. 15:33
Wasn't there earlier this week,.
lizmat ah, indeed I forgot to set RAKUDO_RAKUAST=1 for make test
so it now passes the tests that were previously failing because of missing identifiers 15:34
Xliff: yeah, jnthn also ran into that
Xliff someone with some build chops is welcome to investigate :-) 15:35
sena_kun We still have github.com/rakudo/rakudo/issues/3824 15:42
Geth rakudo/rakuast: 46fec6c4e9 | (Jonathan Worthington)++ | 5 files
Basic RakuAST support for `use lib`
16:06
rakudo/rakuast: 5d5e6c51b3 | (Jonathan Worthington)++ | 3 files
RakuAST for $?FILE and $?LINE

Only compilation from source can really provide values for these, so they are to be constructed with the appropriate value.
rakudo/rakuast: 2d8a33b9a7 | (Jonathan Worthington)++ | 3 files
Implement a little more AST interpretation

This is aimed quite directly at the `use lib` statements that appear in numerous spectest files, although being able to interpret method calls is of course quite generally useful.
Geth rakudo/rakuast: eb2bfe640d | (Jonathan Worthington)++ | src/Raku/Actions.nqp
Fix mis-compile of `,` list with trailing `,`
16:30
AlexDaniel c: rakuast say 42 17:48
committable6 AlexDaniel, ¦rakuast: «Cannot test this commit (Commit exists, but an executable could not be built for it)»
AlexDaniel lizmat: I assume it is buildable if you use the same branch for nqp?
lizmat: if so, maybe bump it?
Altai-man AlexDaniel, I think it needs an env variable to actually use rakuast frontend? (can be simulated with run, yes) 17:50
tellable6 2020-09-01T19:58:30Z #raku <El_Che> Altai-man: building locally now
2020-09-01T19:59:20Z #raku <El_Che> Altai-man: profile test failes, but running raku --profile does not give an error
2020-09-01T20:00:18Z #raku <El_Che> Altai-man: funnily enough the other ubuntu 16.04 pkg builds fine. The only difference I can find is that I build each project separately
hey Altai-man, you have a message: gist.github.com/381590a180aefeb167...eb99c00621
AlexDaniel Altai-man: yeah, that'd be just one `run` away 17:51
Geth rakudo/precomp_store_parallelism: f2106afcd1 | (Stefan Seifert)++ | src/core.c/CompUnit/PrecompilationRepository.pm6
Fix long standing race condition in precompilation causing spurious test failures

When we load a precomp file we expect to find an accompanying .repo-id file with the recorded state of the repo chain at the time of precompilation. However we wrote that file only after writing the precomp file itself. Since loading of modules is lock free this has always left a window of opportunity for the dependency check to end up without a repo-id and a successive blow up in the argument check of store-repo-id.
Fix by writing the repo-id file before the precomp file.
19:10
nine That ^^^ may finally take care of some inexplicable make test failures that popped up occasionally on CI 19:11
lizmat nine++ 19:13
[Coke] ten 19:15
nine Though almost more important for me was that I finally dug into how to get at the real error messages instead of just a "Dubious, test returned 1 (wstat 256, 0x100)" when running parallel tests 19:16
Just needed to install Perl's TAP::Harness::Archive and add --archive=make-test.tar to the call in m-coretest5 in the Makefile 19:19
timotimo yes!! 19:28
make that available to everybody everywhere please
does it write results as files finish, or something? so in theory you could "tail" the tar file?
nine No, it writes the tar file once the tests are done 19:29
timotimo dang ... 19:30
nine timotimo: build.opensuse.org/package/rdiff/h...mp;rev=586 20:04
timotimo "please login" :( 20:15
Geth rakudo/rakuast: 8653a24212 | (Jonathan Worthington)++ | src/Raku/Actions.nqp
Fix copy pasta; japhb++
21:04
rakudo/rakuast: c7d8f60e74 | (Elizabeth Mattijsen)++ | 6 files
Annotate tests with the associated Raku code

  - as close as possible
  - intended to aid in learning RakuAST
  - for me, and hopefully for other people as well
  - first batch, more tomorrow
21:55