|
01:29
iwuvmika left
|
|||
| apogee_ntv | tadzik: I genuinely don't think it's unreasonable to push back on reviews you cannot safely approve. | 01:43 | |
| It's not obstructionist to ask someone to break the PR up and an LLM can do that just fine. | 01:44 | ||
| I'm as pro-AI as pretty much anyone and the purpose of LLMs is not and cannot be to offload our work onto our colleagues. Throwing huge unchecked PRs at people is that. | 01:48 | ||
| If you cant review it, they almost certainly didnt either. | |||
|
01:50
kylese left
01:51
kylese joined
|
|||
| disbot | <antononcube> At some point someone will come with a "new Agile" that can streamline these kind developer interaction that include LLMs. | 01:58 | |
| <antononcube> (Might be called "Viber" or "Vibrator".) | 01:59 | ||
|
02:15
kylese left,
kylese joined
02:26
kylese left
02:27
kylese joined
|
|||
| [Coke] | Pretty sure it wqon't be called the latter. | 02:30 | |
| apogee_ntv | Its on my very long todo list to build proper flows for this kind of thing. | 02:33 | |
| I kindof have them but they're enforced from local prompt files that have a lot of project-specific stuff in. Not in a reusable format. | |||
|
03:11
stanrifkin_ joined
03:14
stanrifkin left
|
|||
| Voldenet | meh, good code, even made by AI looks nice, but your instructions have to be precise and you have to work over a lot of pieces of code | 03:54 | |
|
03:59
apogee_ntv left
04:06
apogee_ntv joined
04:44
stanrifkin_ left
|
|||
| Voldenet | btw, gotta say that Selkie::UI is amazing, but handler errors are true pain in the neck ;\ | 06:06 | |
| Question: is ⚛== (atomic ==) a valid operator? | 07:52 | ||
| m: my atomicint $x = 0; say ⚛$x == 1; say $x ⚛== 1; say $x # it clearly assigns instead of comparing, but it's not documented anywhere | 07:53 | ||
| camelia | False 1 1 |
||
| Voldenet | I got it here github.com/m-doughty/Selkie/blob/m...kumod#L403 | ||
| lizmat | $x ⚛== 1 actually parses into | 08:42 | |
| RakuAST::ApplyInfix.new( | 08:43 | ||
| left => RakuAST::Var::Lexical.new("\$x"), | |||
| infix => RakuAST::MetaInfix::Assign.new( | |||
| RakuAST::Infix.new("⚛=") | |||
| ), | |||
| right => RakuAST::IntLiteral.new(1) | |||
| ) | |||
| so it looks like it's losing the extra "=" in the parse | |||
| so you're getting ⚛= instead of ⚛== | 08:44 | ||
| looks like the legacy grammar has the same issue | 08:45 | ||
| Voldenet: ^^ care to make an issue ? | 08:46 | ||
| Voldenet | sure | 08:48 | |
| done, i'm not 100% sure about the title, but description is informative enough I hope github.com/rakudo/rakudo/issues/6873 | 08:53 | ||
| lizmat | thanks! Voldenet++ | ||