supporting ascii literals

Signed-off-by: Jakub Doka <jakub.doka2@gmail.com>
This commit is contained in:
Jakub Doka 2024-12-14 21:02:29 +01:00
parent dc96c8b10a
commit 00f6729d31
No known key found for this signature in database
GPG key ID: C6E9A89936B8C143
5 changed files with 56 additions and 25 deletions

View file

@ -121,10 +121,12 @@ main := fn(): uint {
decimal := 255
octal := 0o377
binary := 0b11111111
ascii := '\n'
if hex == decimal & octal == decimal & binary == decimal {
return 0
}
return 1
}
```

View file

@ -314,6 +314,7 @@ impl<'a> Formatter<'a> {
}
Expr::Slf { .. } => f.write_str("Self"),
Expr::String { literal, .. } => f.write_str(literal),
Expr::Char { literal, .. } => f.write_str(literal),
Expr::Comment { literal, .. } => f.write_str(literal),
Expr::Mod { path, .. } => write!(f, "@use(\"{path}\")"),
Expr::Embed { path, .. } => write!(f, "@embed(\"{path}\")"),

View file

@ -306,7 +306,7 @@ fn endoce_string(
literal: &str,
str: &mut Vec<u8>,
report: impl Fn(&core::str::Bytes, &str),
) -> Option<()> {
) -> Option<usize> {
let report = |bytes: &core::str::Bytes, msg: &_| {
report(bytes, msg);
None::<u8>
@ -332,7 +332,9 @@ fn endoce_string(
};
let mut bytes = literal.bytes();
let mut char_len = 0;
while let Some(b) = bytes.next() {
char_len += 1;
if b != b'\\' {
str.push(b);
continue;
@ -354,11 +356,7 @@ fn endoce_string(
str.push(b);
}
if str.last() != Some(&0) {
report(&bytes, "string literal must end with null byte (for now)");
}
Some(())
Some(char_len)
}
pub fn quad_sort<T>(mut slice: &mut [T], mut cmp: impl FnMut(&T, &T) -> core::cmp::Ordering) {

View file

@ -356,6 +356,7 @@ impl<'a, 'b> Parser<'a, 'b> {
T::Idk => E::Idk { pos },
T::Die => E::Die { pos },
T::DQuote => E::String { pos, literal: self.tok_str(token) },
T::Quote => E::Char { pos, literal: self.tok_str(token) },
T::Packed => {
self.packed = true;
let expr = self.unit_expr()?;
@ -896,6 +897,11 @@ generate_expr! {
pos: Pos,
literal: &'a str,
},
/// `'\'([^']|\\\')\''`
Char {
pos: Pos,
literal: &'a str,
},
/// `'//[^\n]' | '/*' { '([^/*]|*/)*' | Comment } '*/'
Comment {
pos: Pos,

View file

@ -2975,7 +2975,7 @@ impl<'a> Codegen<'a> {
}
Expr::Ident { id, pos, .. } => self.find_type_as_value(pos, self.ci.parent, id, ctx),
Expr::Comment { .. } => Some(Value::VOID),
Expr::String { pos, literal } => {
Expr::Char { pos, literal } | Expr::String { pos, literal } => {
let literal = &literal[1..literal.len() - 1];
let mut data = core::mem::take(&mut self.pool.lit_buf);
debug_assert!(data.is_empty());
@ -2984,7 +2984,30 @@ impl<'a> Codegen<'a> {
self.error(pos + (literal.len() - bytes.len()) as u32 - 1, message);
};
crate::endoce_string(literal, &mut data, report).unwrap();
let char_count = crate::endoce_string(literal, &mut data, report).unwrap();
if matches!(expr, Expr::Char { .. }) {
if char_count != 1 {
return self.error(
pos,
fa!("character literal can only contain one character, \
but you supplied {char_count}"),
);
}
let value = match data.as_slice() {
&[v] => v as i64,
_ => return self.error(pos, "TODO: support utf-8 characters"),
};
data.clear();
self.pool.lit_buf = data;
self.gen_inferred_const(ctx, ty::Id::U8, value, ty::Id::is_integer)
} else {
if data.last() != Some(&0) {
self.error(pos, "string literal must end with null byte (for now)");
}
let ty = self.tys.make_ptr(ty::Id::U8);
let global = self
@ -3006,6 +3029,7 @@ impl<'a> Codegen<'a> {
Some(Value::new(global).ty(ty))
}
}
Expr::Defer { pos, value } => {
self.ci.defers.push((pos, ExprRef::new(value)));
Some(Value::VOID)