|
6macros: discussing the finer points of Perl 6 macros, Qtrees, and how to stay sane | irclog: irclog.perlgeek.de/6macros/today Set by moderator on 28 July 2015. |
|||
|
02:25
FROGGS_ joined
07:54
FROGGS joined
09:01
FROGGS joined
09:40
Ven joined
|
|||
| Ven | o/ | 09:40 | |
| masak: I thought I had linked some here, but apparently not – whoops! It's a scheme / racket-style system, at least | 09:41 | ||
| masak | what is? that felt like a reply to something, but I don't know what | 14:57 | |
|
16:17
Ven joined
|
|||
| Ven | masak: yes, 'twas a reply to the rust macro link | 16:21 | |
| masak | aha; Rust's macro system is a scheme/racket-style one? | 16:24 | |
| I think I rejected that interpretation because I found it unlikely, for some reason :) | |||
| Ven | yes, "templates" | 16:27 | |
| at least macro_rules :) | 16:28 | ||
| well, it's more racket-ish than scheme-ish, because it allows more IIRC | 16:29 | ||
| > Rust's AST is untyped for two reasons, I think (though I'm not a Rust dev). First, type inference may depend on the expansion of macros, procedural or otherwise, and so the pre-expanded AST can't be fully typed properly. Second, the type system provides the fundamental soundness guarantees that Rust relies on, and typechecking must be run after all AST manipulations are finished. This is why the AST doesn't have access to Rust's seman | 16:30 | ||
| masak | cut off at "Rust's semant" | 16:57 | |
| Ven | > access to Rust's semantic type information. | 16:59 | |
| mmh, didn't cut off in the web client | |||
| masak | well, the quote makes sense | 17:00 | |
|
18:53
vendethiel- joined
19:19
FROGGS joined
19:34
vendethiel joined
|
|||
| masak is on a roll tonight with the 007 commits | 21:08 | ||
| vendethiel | haha. cute implementation of @ :) | 21:14 | |
| masak | it's coming along :) | 21:15 | |
| the code is a tad repetitive so far, but I'm hoping to improve on that once I have more Q types covered | |||
| it feels like there should be some factoring that just puts everything in a hash, like we talked about | |||
| vendethiel | yeah, that seems feasible | 21:17 | |
| "@" <.ws> <name=...> '{' <.ws> <::(%rules{$<name>} // die "No such Q type: $<name>"> <.ws> '}' | 21:18 | ||
| masak | sump'n like that | 21:19 | |
| wacky idea #27: convert the 007 source code into a form where it regularly sets state in a global hash. have the program dump the hash at the end. randomly produce/twiddle/evolve 007 programs with the goal of getting as much "coverage" in terms of new states set in the hash. | 23:03 | ||
| inspired by en.wikipedia.org/wiki/American_fuz...p_(fuzzer) | 23:04 | ||