pawarser(init): start extracting the parser lib

This commit is contained in:
Schrottkatze 2024-10-13 15:32:26 +02:00
parent a693b57447
commit a3ab844ba7
No known key found for this signature in database
4 changed files with 61 additions and 1 deletions

9
Cargo.lock generated
View file

@ -1159,6 +1159,15 @@ version = "1.0.15"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "57c0d7b74b563b49d38dae00a0c37d4d6de9b432382b2892f0574ddcae73fd0a"
[[package]]
name = "pawarser"
version = "0.1.0"
dependencies = [
"drop_bomb",
"enumset",
"rowan",
]
[[package]]
name = "petgraph"
version = "0.6.5"

View file

@ -6,7 +6,7 @@ members = [
"crates/lang",
"crates/svg-filters",
"crates/prowocessing",
"crates/executor-poc",
"crates/executor-poc", "crates/pawarser",
]
resolver = "2"

View file

@ -0,0 +1,12 @@
[package]
name = "pawarser"
version = "0.1.0"
edition = "2021"
[dependencies]
rowan = "0.15.15"
drop_bomb = "0.1.5"
enumset = "1.1.3"
[lints]
workspace = true

View file

@ -0,0 +1,39 @@
#![feature(iter_collect_into)]
pub mod parser {
pub mod input {
use enumset::{EnumSet, EnumSetType};
struct Input<'src, 'toks, SyntaxKind: EnumSetType + Into<rowan::SyntaxKind>> {
raw: &'toks Vec<(SyntaxKind, &'src str)>,
// enumset of meaningless tokens
semantically_meaningless: EnumSet<SyntaxKind>,
// indices of non-meaningless tokens
meaningful_toks: Vec<usize>,
}
impl<'src, 'toks, SyntaxKind: EnumSetType + Into<rowan::SyntaxKind>>
Input<'src, 'toks, SyntaxKind>
{
pub fn new(
raw_toks: &'toks Vec<(SyntaxKind, &'src str)>,
meaningless: Option<EnumSet<SyntaxKind>>,
) -> Self {
let mut meaningful_toks = Vec::new();
if let Some(meaningless) = meaningless {
let meaningful_toks = raw_toks
.iter()
.enumerate()
.filter_map(|(i, tok)| (!meaningless.contains(tok.0)).then_some(i))
.collect_into(&mut meaningful_toks);
}
Self {
raw: raw_toks,
semantically_meaningless: meaningless.unwrap_or_default(),
meaningful_toks,
}
}
}
}
}