Geth rakudo: ugexe++ created pull request #6417:
RakuAST: optimize for loops
00:42
rakudo: ugexe++ created pull request #6418:
RakuAST: interpret a curried argument of a BEGIN-time call
01:19
rakudo: ugexe++ created pull request #6419:
RakuAST: look the GLOBAL package up at run time
01:53
02:41 apogee_ntv left 02:43 apogee_ntv joined
Geth rakudo: ugexe++ created pull request #6420:
RakuAST: report compiler worries on the node instead of the resolver
03:14
rakudo: ugexe++ created pull request #6421:
RakuAST: intern Rat and Complex literals
03:47
04:26 guifa joined
Geth rakudo: ugexe++ created pull request #6422:
RakuAST: apply val to quoted atoms in quote words literals
04:37
rakudo: ugexe++ created pull request #6423:
RakuAST: remove TODO comments whose work has been done
04:53
04:55 guifa left
Geth rakudo: ugexe++ created pull request #6424:
RakuAST: use Scalar as the container type of $ sigil variables
05:08
rakudo/main: 008acbbba5 | (Nick Logan)++ | 2 files
RakuAST: honor a version/auth constraint on `require`

  `require Foo:ver(...)` (and :auth/:api/:from) built the dependency
specification with the whole `Foo:ver<...>` string as the short-name and no matchers, so it looked up a module literally named "Foo:ver<...>" and never found it. Only a plain `require Foo` worked.
... (5 more lines)
05:10
rakudo/main: 4adfb24b0c | (Nick Logan)++ (committed using GitHub Web editor) | 2 files
Merge pull request #6412 from ugexe/ugexe/rakuast-require-adverbs

RakuAST: honor a version/auth constraint on `require`
rakudo/main: e030558a01 | (Nick Logan)++ (committed using GitHub Web editor) | 2 files
Give the experimental macros flag a load-independent value (#6413)

A module that does `use v6.e.PREVIEW; use experimental`, once precompiled and loaded with `use` alongside a fresh load of experimental, died at compile time:
   Merging GLOBAL symbols failed: duplicate definition of symbol EXPERIMENTAL-MACROS
... (14 more lines)
05:11
rakudo/main: 6e67320659 | (Nick Logan)++ (committed using GitHub Web editor) | 2 files
RakuAST: iterate a sunk `xx` result so its thunk runs (#6414)

Previously `gather do { take 1 } xx *` collected nothing under RakuAST. A sunk `A xx *` returns a lazy Seq, and nothing iterated it, so the repeated thunk never ran and its takes never reached the gather.
ApplyInfix.needs-sink-call is `infix.is-pure || infix.IMPL-RESULT-NEEDS-ITERATION`. Both were False for `xx`, so no sink call was emitted. A finite `A xx N` reifies eagerly and ran its thunk regardless, which hid the gap. Have Infix report IMPL-RESULT-NEEDS-ITERATION True for `xx` so a sunk result is iterated.
rakudo: ugexe++ created pull request #6425:
RakuAST: resolve the 'maybe also check?' questions on synthesized nodes
05:22
06:02 finanalyst joined 08:03 vrurg_ joined, vrurg left 09:09 finanalyst left 13:04 guifa joined 14:02 jdv joined
Geth rakudo: ugexe++ created pull request #6426:
RakuAST: form the meta-op of a setting operator reduce at compile time
14:50
ugexe currently going through rakuast TODO comments and addressing them 14:51
Geth rakudo: ugexe++ created pull request #6427:
RakuAST: die on undefined nodes in SET-NODE-ORIGIN
15:08
15:28 Pixi` is now known as Pixi
Geth rakudo: ugexe++ created pull request #6428:
RakuAST: report attributes declared without a package as a sorry
15:52
16:18 guifa left
Geth rakudo: ugexe++ created pull request #6429:
RakuAST: drop the dead second-chance resolution for qualified calls
16:53
rakudo: ugexe++ created pull request #6430:
RakuAST: set attr_package on attributive parameters
17:07
rakudo: ugexe++ created pull request #6431:
RakuAST: assert the condition modifier thunk stays innermost
17:19
rakudo: ugexe++ created pull request #6432:
RakuAST: Drop the dead resolve-only flag from IMPL-CHECK
17:52
rakudo: ugexe++ created pull request #6433:
RakuAST: detect a DEPARSE slang name with a type check
17:59
[Coke] ok, I put in some debug statements in my copy of zef on the blin box... and now it's just *completely* hanging. 18:01
wonder if the recompile is problematic, or... 18:02
it ran for 20m with no output past the print statement just before running zef update before I killed it 18:13
/home/azureuser/raku/bin/rakudo data/zef/bin/zef update is fast. 18:55
/home/azureuser/raku/bin/rakudo -I data/zef data/zef/bin/zef update is sloooow
the zef version is v1.1.3 from the git repo 18:56
Installed zef on the box is 1.0.0 18:57
(presumably getting those versions without the -I
with the -I, it's now not even emitting the "updating"... with the 1.0 version of libs and the config file, it's slow after. 19:00
ok. if I remove the store folder, it's fast. 19:09
ls -R data/zef-data/store-bkp/ | wc -l -- 180284 19:10
I suppose I could remove store before every blin run, but I feel like that would slow down the individual pulls later, no? 19:13
Yup, seems to be the presence of the store directory. 19:16
Geth Blin: 7186032137 | (Will Coleda)++ | util/run-blin.sh
Start with a fresh zef store each run

  ... or spend 10 minutes waiting for the initial zef update to complete.
19:20
19:33 apogee_ntv left
Geth Blin/coke/cleanup: 14 commits pushed by (Will Coleda)++
review: github.com/Raku/Blin/compare/ee89a...562dc7f1a4
19:35
Blin/coke/cleanup: 4353257654 | (Will Coleda)++ | META6.json
add self as author
19:41
20:23 finanalyst joined
Geth rakudo: ugexe++ created pull request #6434:
RakuAST: support the :dba regex modifier and pass it to FAILGOAL
20:35
20:52 apogee_ntv joined
Geth rakudo: ugexe++ created pull request #6435:
RakuAST: give the S meta-operator its base operator's properties
21:00
rakudo: ugexe++ created pull request #6436:
RakuAST: report errors on a failed cursor from its highwater mark
21:14
21:40 finanalyst left
Geth rakudo: ugexe++ created pull request #6437:
RakuAST: let the do statement prefix use its translatable token
22:23
rakudo: ugexe++ created pull request #6438:
RakuAST: name the bracket in the arrow postfix dereference error
22:41
rakudo: ugexe++ created pull request #6439:
RakuAST: support DOC-prefixed use statements
22:52
rakudo: ugexe++ created pull request #6440:
RakuAST: don't eat the newline after a trailing paren declarator doc
23:15
rakudo/main: 7ba38717ae | (Nick Logan)++ (committed using GitHub Web editor) | 10 files
RakuAST: resolve `$?FILE` at parse time so it matches legacy (#6415)

Previously `$?FILE` was declared as an implicit lexical constant in RakuAST::CompUnit::PRODUCE-IMPLICIT-DECLARATIONS, holding the source path with the cwd prefix stripped, and each `$?FILE` reference compiled to a generic lexical lookup. This had two problems. When the CORE setting itself was the compunit being compiled, the constant leaked ... (26 more lines)
23:33
rakudo/main: b837303ce3 | (Nick Logan)++ (committed using GitHub Web editor) | 2 files
RakuAST: give an attributive parameter its declaring package for the full binder (#6416)

A parameter with a three-name alias (`:x(:y(:$z))`) forces the whole signature through the full binder. A private attributive parameter (`:$!x`) in such a signature was bound with no declaring package, so the binder fetched the attribute off Mu and died with "no such attribute". The inline bindings path was fine because it reaches the attribute through its own lookup. ... (5 more lines)
23:34
rakudo: ugexe++ created pull request #6441:
RakuAST: parse a custom postcircumfix subscript as a semilist
23:52