03:35 rakkable left 03:36 rakkable joined 04:18 melezhik joined 06:27 melezhik left 08:40 melezhik joined
ab5tract it feels like the nominalize code needs to be reverted or fixed, because as of now, it is actually and quite clearly broken 09:01
for example, calling SubsetHOW.nominalize doesn't return the subset type at all, it returns either the refinee or the nominalized refinee 09:02
That's a hell of a behavior change. I think the better approach would to ensure that nominalization happens prior to target attachment 09:03
Geth rakudo/ab5tract/r5912: 926c269893 | ab5tract++ | 2 files
RakuAST/v6.e - Fix subset type defaults in containers

The proper place for maybe-nominalization (which in turn also involves maybe-definite-ization, for versions < v6.e.
Only RakuAST needs to be fixed here, thankfully.
10:37
rakudo/ab5tract/r5912: f06806ff33 | ab5tract++ | 2 files
RakuAST/v6.e - Fix subset type defaults in containers

The proper place for maybe-nominalization (which in turn also involves maybe-definite-ization, for versions < v6.e.
Only RakuAST needs to be fixed here, thankfully, as I have had way too high a does of metamodel internals for the moment. ... (14 more lines)
10:45
rakudo: ab5tract++ created pull request #5996:
RakuAST/v6.e - Fix subset type defaults in containers
10:50 melezhik left 10:54 lizmat left 10:55 lizmat joined, lizmat left 11:05 librasteve_ left 11:09 melezhik joined
Geth rakudo/ab5tract/r5912: 77b98449fb | ab5tract++ | src/Raku/ast/type.rakumod
Why were coercions excluded?
11:48
12:38 Pixi joined 12:41 Pixi` left
ab5tract ok so a subtest of subtests fail with `make t/spec/S02-types/subset-6e.t` that do *not* fail when I extract said subtest, first class all its child subtests, and run 12:59
the result is every test passes. how useful for debugging? :O 13:01
can no one save me from this insanity? 👾 13:20
13:21 lizmat joined 16:28 finanalyst joined 19:07 melezhik_ joined 19:32 melezhik_ left, melezhik_ joined 19:43 timo left 19:48 melezhik_ left, melezhik_ joined 20:01 timo joined 20:07 melezhik_ left 20:20 melezhik_ joined 20:27 melezhik_ left, melezhik_ joined
melezhik_ . 20:34
tellable6 2025-10-25T21:31:12Z #raku-dev <ab5tract> melezhik: I kicked the container.
2025-10-25T21:39:32Z #raku-dev <ab5tract> melezhik: also wanted to suggest that you declare `config` as a term. that way you can use the associative access directly, no need for parentheses
melezhik_ put results of this weekend brownie demo to github.com/coke/blin2/issues/1#iss...448905787, thanks ab5tract: a lot for help and providing agents for testing
who ever wants to participate ask [Coke] for access to blin2 20:35
20:44 melezhik_ left 20:55 MasterDuke joined 20:56 melezhik_ joined 21:00 melezhik_ left 21:35 kjp left 21:42 MasterDuke left 21:50 melezhik left