Andreas Rumpf
Andreas Rumpf
> it would be helpful to know what requires a copy or maybe have some metrics to track that. Well the compiler follows the spec, https://nim-lang.org/docs/destructors.html#rewrite-rules but there is also...
**Update**: The "Nimony" compiler offers 4 out of 5 of these: - [x] Incremental recompilations. - [x] No forward declarations for procs and types required. - [ ] Allow for...
Yes, prepare for: - Explicit `nil ref/ptr T` annotations, benefit: no more null pointer crashes at runtime. - `proc mygeneric[T: ](...) = ...`; benefit: increased compile-time checking - `case` inside...
> but will the resulting functionality for the Nim language be similar? Yes, very! But already the plugin system is even more powerful and allows for some things that are...
Is this also reproducible with `nim check`?
Ok, so it really is nimsuggest specific then.
Just use `strscans`, it's soo much better than regexes.
It's a valid bug regardless. And there is nim-regex as a third-party solution.
I'm tired of this `uninit` bullshit. Make the compiler use the DFA and stop exposing these nukes to Nim users.
DFA is one thing and the other is better API design like `newSeqWith(size, value)` or maybe `newSeqIt(4, it + 3)` (which would produce `@[3, 4, 5, 6]`).