|
🦋 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: ... | Logs available at irclogs.raku.org/raku-dev/live.html | For MoarVM see #moarvm Set by lizmat on 8 June 2022. |
|||
| MasterDuke | every once in a while i think about trying to at least rebase it up to HEAD of main, but it's so old that it would be quite an effort | 00:03 | |
| but if that branch ever replaced the existing jvm backend then we would have a good path toward using java (and other langs!) libs | 00:04 | ||
| japhb | MasterDuke: Maybe rebase on top of each release one at a time and see how far you can get before it gets extra difficult? That would probably reduce the "background bitrot" a bit | 00:36 | |
| Or at least constrain the range in which pain might reside | 00:38 | ||
|
02:10
MasterDuke left
|
|||
| patrickb plans to work on an Inline::Java some time in the future. That's a different approach to using JVM libs. | 07:05 | ||
|
09:28
sena_kun joined
|
|||
| Geth | rakudo/main: e294a3e042 | (Elizabeth Mattijsen)++ | 2 files Always use Routine.dispatch_order to get candidates Make sure that the serialization context is set correct when generating the sorted list. Without it, we break the reproducibility test in "make test" although I have yet to come up with a reason why that is. |
10:53 | |
|
12:10
MasterDuke joined
|
|||
| Geth | rakudo/main: fa853abc7f | (Daniel Green)++ | src/Perl6/bootstrap.c/BOOTSTRAP.nqp Revert 99f52bc for the JVM, unbreaking its build Without this the JVM build dies while compiling CORE.c.setting with: Cannot call infix:<+|>; no signatures match in throw_or_die (gen/jvm/Metamodel.nqp:51) in <anon> (gen/jvm/BOOTSTRAP/v6c.nqp:4179) in <anon> (gen/jvm/BOOTSTRAP/v6c.nqp:4149) ... (7 more lines) |
12:14 | |
| rakudo/main: 3eed834ccf | (Elizabeth Mattijsen)++ | src/Perl6/bootstrap.c/BOOTSTRAP.nqp Abstract the graph checking logic into a Node class This allowed for simplification and way fewer hash lookups and better readability. Execution speed improved but only very slightly better than noise levels for make (spec)test. |
13:22 | ||
|
15:32
MasterDuke left
|
|||
| Geth | rakudo/main: 43017b218f | (Elizabeth Mattijsen)++ | src/Perl6/World.nqp Use single constant It looks like nqp::const::foo +| nqp::const::bar is *not* statically optimized into a single constant value (at least judging from the QAST with --target==optimize. Since nqp::const::foo handling is really a MAST operation, perhaps that would be the place to optimize nqp::const::foo +| nqp::const::bar handling. |
18:31 | |
| rakudo/main: 4d91de6d1b | (Elizabeth Mattijsen)++ | 4 files Update internal documentation on attributes of classes created in the bootstrap: a few new parameters were forgotten |
19:03 | ||
| lizmat | .tell MasterDuke re JVM issues in NQP, maybe it needs a bootstrap ? | 19:23 | |
| tellable6 | lizmat, I'll pass your message to MasterDuke | ||
| Geth | nqp/main: 216a063547 | (Elizabeth Mattijsen)++ | 2 files Add nqp::const::BIND_(RESULT|VAL)_xxx constants So that they can be used instead of just numbers like 0, 1, 2, 3 or 10 for bind value types, for instance. |
19:47 | |
| nqp/main: 106c9c8503 | (Elizabeth Mattijsen)++ | 10 files Bootstrap MoarVM to get nqp::const::BIND_(RESULT|VAL)_ constants So that we can use them inside NQP as well |
|||
| lizmat | .tell MasterDuke I'll wait with using them inside NQP until the JVM has been bootstrapped (github.com/Raku/nqp/commit/106c9c8503) | 19:48 | |
| tellable6 | lizmat, I'll pass your message to MasterDuke | ||
| Geth | rakudo/main: b4b895d799 | (Elizabeth Mattijsen)++ | tools/templates/NQP_REVISION Bump NQP to get nqp::const::BIND_(RESULT|VAL)_xxx constants |
20:00 | |
|
20:00
MasterDuke joined
|
|||
| Geth | nqp/main: 306001e2da | (Daniel Green)++ | 10 files Bootstrap JVM backend |
20:14 | |