Niko Matsakis
Niko Matsakis
Some investigation: The AST that we construct for match items is here: https://github.com/lalrpop/lalrpop/blob/896ab62e039a3a7f5f986bed102864c8b813f740/lalrpop/src/grammar/parse_tree.rs#L85-L92 This gets processed here into one of these data structures, where the catch all becomes just a...
@k0pernicus great! ping me if you run into any questions, and/or reach out on gitter
@k0pernicus I hope the info you want is there :) let me know otherwise
@k0pernicus > ndeed, I don't understand how to get the current number of precedence (match_group_precedence if I am correct ?) in order to compute the new one in add_literal_from_grammar... Is...
> As I am new in the project, I can't say if this two-level test is not covering enough this feature, or if I can use already written unit tests...
Note that you will want to run rustfmt too, thanks to https://github.com/lalrpop/lalrpop/pull/327
@k0pernicus d'oh, missed your updates (please do ping on gitter, I get overwhelmed easily =) Regarding those tests -- it seems ok to support the catch all in any place,...
Another use case for this is contextual keywords, as described in #112.
OK, so, I gave some more thought to this. I think it'd be fairly easy to use (e.g.) a combination of LR0 states and follow-sets to partition the input tokens...
OK, re-reading your comment, I think I understand better. So there are two distinct things going on here: - Converting `#![foo]` to `#[foo]` -- this may indeed be a good...