rust-lang/rust
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

Make char::is_lowercase and char::is_uppercase const

Browse Source 
Implements #101400.
This commit is contained in:
Sage Mitchell
2022-09-02 20:52:28 -07:00
parent b11bf65e4a
commit 4a3e169da7
5 changed files with 55 additions and 27 deletions
Download Patch File Download Diff File Expand all files Collapse all files

11
src/tools/unicode-table-generator/src/range_search.rs
View File

@@ -1,5 +1,6 @@
#[rustc_const_unstable(feature = "const_unicode_case_lookup", issue = "101400")]
#[inline(always)]
fn bitset_search<
const fn bitset_search<
const N: usize,
const CHUNK_SIZE: usize,
const N1: usize,
@@ -15,14 +16,14 @@ fn bitset_search<
let bucket_idx = (needle / 64) as usize;
let chunk_map_idx = bucket_idx / CHUNK_SIZE;
let chunk_piece = bucket_idx % CHUNK_SIZE;
let chunk_idx = if let Some(&v) = chunk_idx_map.get(chunk_map_idx) {
v
let chunk_idx = if chunk_map_idx < chunk_idx_map.len() {
chunk_idx_map[chunk_map_idx]
} else {
return false;
};
let idx = bitset_chunk_idx[chunk_idx as usize][chunk_piece] as usize;
let word = if let Some(word) = bitset_canonical.get(idx) {
*word
let word = if idx < bitset_canonical.len() {
bitset_canonical[idx]
} else {
let (real_idx, mapping) = bitset_canonicalized[idx - bitset_canonical.len()];
let mut word = bitset_canonical[real_idx as usize];

15
src/tools/unicode-table-generator/src/raw_emitter.rs
View File

@@ -76,7 +76,7 @@ impl RawEmitter {
writeln!(
&mut self.file,
"static BITSET_CANONICAL: [u64; {}] = [{}];",
"const BITSET_CANONICAL: [u64; {}] = [{}];",
canonicalized.canonical_words.len(),
fmt_list(canonicalized.canonical_words.iter().map(|v| Bits(*v))),
)
@@ -84,7 +84,7 @@ impl RawEmitter {
self.bytes_used += 8 * canonicalized.canonical_words.len();
writeln!(
&mut self.file,
"static BITSET_MAPPING: [(u8, u8); {}] = [{}];",
"const BITSET_MAPPING: [(u8, u8); {}] = [{}];",
canonicalized.canonicalized_words.len(),
fmt_list(&canonicalized.canonicalized_words),
)
@@ -96,7 +96,12 @@ impl RawEmitter {
self.blank_line();
writeln!(&mut self.file, "pub fn lookup(c: char) -> bool {{").unwrap();
writeln!(
&mut self.file,
r#"#[rustc_const_unstable(feature = "const_unicode_case_lookup", issue = "101400")]"#
)
.unwrap();
writeln!(&mut self.file, "pub const fn lookup(c: char) -> bool {{").unwrap();
writeln!(&mut self.file, " super::bitset_search(",).unwrap();
writeln!(&mut self.file, " c as u32,").unwrap();
writeln!(&mut self.file, " &BITSET_CHUNKS_MAP,").unwrap();
@@ -130,7 +135,7 @@ impl RawEmitter {
writeln!(
&mut self.file,
"static BITSET_CHUNKS_MAP: [u8; {}] = [{}];",
"const BITSET_CHUNKS_MAP: [u8; {}] = [{}];",
chunk_indices.len(),
fmt_list(&chunk_indices),
)
@@ -138,7 +143,7 @@ impl RawEmitter {
self.bytes_used += chunk_indices.len();
writeln!(
&mut self.file,
"static BITSET_INDEX_CHUNKS: [[u8; {}]; {}] = [{}];",
"const BITSET_INDEX_CHUNKS: [[u8; {}]; {}] = [{}];",
chunk_length,
chunks.len(),
fmt_list(chunks.iter()),