Skip to content
This repository was archived by the owner on May 28, 2025. It is now read-only.

Commit 8ff3903

Browse files
fee1-deadfee1-dead
committed
initial step towards implementing C string literals
1 parent 7b99493 commit 8ff3903

File tree

17 files changed

+310
-80
lines changed

17 files changed

+310
-80
lines changed

‎compiler/rustc_ast/src/ast.rs‎

Lines changed: 3 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -1814,6 +1814,8 @@ pub enum LitKind {
18141814
/// A byte string (`b"foo"`). Not stored as a symbol because it might be
18151815
/// non-utf8, and symbols only allow utf8 strings.
18161816
ByteStr(Lrc<[u8]>, StrStyle),
1817+
/// A C String (`c"foo"`).
1818+
CStr(Lrc<[u8]>, StrStyle),
18171819
/// A byte char (`b'f'`).
18181820
Byte(u8),
18191821
/// A character literal (`'a'`).
@@ -1868,6 +1870,7 @@ impl LitKind {
18681870
// unsuffixed variants
18691871
LitKind::Str(..)
18701872
| LitKind::ByteStr(..)
1873+
| LitKind::CStr(..)
18711874
| LitKind::Byte(..)
18721875
| LitKind::Char(..)
18731876
| LitKind::Int(_, LitIntType::Unsuffixed)

‎compiler/rustc_ast/src/token.rs‎

Lines changed: 7 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -74,6 +74,8 @@ pub enum LitKind {
7474
StrRaw(u8), // raw string delimited by `n` hash symbols
7575
ByteStr,
7676
ByteStrRaw(u8), // raw byte string delimited by `n` hash symbols
77+
CStr,
78+
CStrRaw(u8),
7779
Err,
7880
}
7981

@@ -141,6 +143,10 @@ impl fmt::Display for Lit {
141143
delim = "#".repeat(n as usize),
142144
string = symbol
143145
)?,
146+
CStr => write!(f, "c\"{symbol}\"")?,
147+
CStrRaw(n) => {
148+
write!(f, "cr{delim}\"{symbol}\"{delim}", delim = "#".repeat(n as usize))?
149+
}
144150
Integer | Float | Bool | Err => write!(f, "{symbol}")?,
145151
}
146152

@@ -170,6 +176,7 @@ impl LitKind {
170176
Float => "float",
171177
Str | StrRaw(..) => "string",
172178
ByteStr | ByteStrRaw(..) => "byte string",
179+
CStr | CStrRaw(..) => "C string",
173180
Err => "error",
174181
}
175182
}

‎compiler/rustc_ast/src/util/literal.rs‎

Lines changed: 54 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -2,7 +2,10 @@
22
33
use crate::ast::{self, LitKind, MetaItemLit, StrStyle};
44
use crate::token::{self, Token};
5-
use rustc_lexer::unescape::{byte_from_char, unescape_byte, unescape_char, unescape_literal, Mode};
5+
use rustc_lexer::unescape::{
6+
byte_from_char, unescape_byte, unescape_c_string, unescape_char, unescape_literal, CStrUnit,
7+
Mode,
8+
};
69
use rustc_span::symbol::{kw, sym, Symbol};
710
use rustc_span::Span;
811
use std::{ascii, fmt, str};
@@ -158,6 +161,52 @@ impl LitKind {
158161

159162
LitKind::ByteStr(bytes.into(), StrStyle::Raw(n))
160163
}
164+
token::CStr => {
165+
let s = symbol.as_str();
166+
let mut buf = Vec::with_capacity(s.len());
167+
let mut error = Ok(());
168+
unescape_c_string(s, Mode::CStr, &mut |span, c| match c {
169+
Ok(CStrUnit::Byte(0) | CStrUnit::Char('\0')) => {
170+
error = Err(LitError::NulInCStr(span));
171+
}
172+
Ok(CStrUnit::Byte(b)) => buf.push(b),
173+
Ok(CStrUnit::Char(c)) if c.len_utf8() == 1 => buf.push(c as u8),
174+
Ok(CStrUnit::Char(c)) => {
175+
buf.extend_from_slice(c.encode_utf8(&mut [0; 4]).as_bytes())
176+
}
177+
Err(err) => {
178+
if err.is_fatal() {
179+
error = Err(LitError::LexerError);
180+
}
181+
}
182+
});
183+
error?;
184+
buf.push(b'\0');
185+
LitKind::CStr(buf.into(), StrStyle::Cooked)
186+
}
187+
token::CStrRaw(n) => {
188+
let s = symbol.as_str();
189+
let mut buf = Vec::with_capacity(s.len());
190+
let mut error = Ok(());
191+
unescape_c_string(s, Mode::RawCStr, &mut |span, c| match c {
192+
Ok(CStrUnit::Byte(0) | CStrUnit::Char('\0')) => {
193+
error = Err(LitError::NulInCStr(span));
194+
}
195+
Ok(CStrUnit::Byte(b)) => buf.push(b),
196+
Ok(CStrUnit::Char(c)) if c.len_utf8() == 1 => buf.push(c as u8),
197+
Ok(CStrUnit::Char(c)) => {
198+
buf.extend_from_slice(c.encode_utf8(&mut [0; 4]).as_bytes())
199+
}
200+
Err(err) => {
201+
if err.is_fatal() {
202+
error = Err(LitError::LexerError);
203+
}
204+
}
205+
});
206+
error?;
207+
buf.push(b'\0');
208+
LitKind::CStr(buf.into(), StrStyle::Raw(n))
209+
}
161210
token::Err => LitKind::Err,
162211
})
163212
}
@@ -191,6 +240,8 @@ impl fmt::Display for LitKind {
191240
string = symbol
192241
)?;
193242
}
243+
// TODO need to reescape
244+
LitKind::CStr(..) => todo!(),
194245
LitKind::Int(n, ty) => {
195246
write!(f, "{n}")?;
196247
match ty {
@@ -237,6 +288,8 @@ impl MetaItemLit {
237288
LitKind::Str(_, ast::StrStyle::Raw(n)) => token::StrRaw(n),
238289
LitKind::ByteStr(_, ast::StrStyle::Cooked) => token::ByteStr,
239290
LitKind::ByteStr(_, ast::StrStyle::Raw(n)) => token::ByteStrRaw(n),
291+
LitKind::CStr(_, ast::StrStyle::Cooked) => token::CStr,
292+
LitKind::CStr(_, ast::StrStyle::Raw(n)) => token::CStrRaw(n),
240293
LitKind::Byte(_) => token::Byte,
241294
LitKind::Char(_) => token::Char,
242295
LitKind::Int(..) => token::Integer,

‎compiler/rustc_ast_pretty/src/pprust/state.rs‎

Lines changed: 2 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -210,6 +210,8 @@ pub fn literal_to_string(lit: token::Lit) -> String {
210210
token::ByteStrRaw(n) => {
211211
format!("br{delim}\"{string}\"{delim}", delim = "#".repeat(n as usize), string = symbol)
212212
}
213+
// TODO
214+
token::CStr | token::CStrRaw(_) => todo!(),
213215
token::Integer | token::Float | token::Bool | token::Err => symbol.to_string(),
214216
};
215217

‎compiler/rustc_builtin_macros/src/concat.rs‎

Lines changed: 4 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -32,6 +32,10 @@ pub fn expand_concat(
3232
Ok(ast::LitKind::Bool(b)) => {
3333
accumulator.push_str(&b.to_string());
3434
}
35+
Ok(ast::LitKind::CStr(..)) => {
36+
cx.span_err(e.span, "cannot concatenate a C string literal");
37+
has_errors = true;
38+
}
3539
Ok(ast::LitKind::Byte(..) | ast::LitKind::ByteStr(..)) => {
3640
cx.emit_err(errors::ConcatBytestr { span: e.span });
3741
has_errors = true;

‎compiler/rustc_builtin_macros/src/concat_bytes.rs‎

Lines changed: 4 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -18,6 +18,10 @@ fn invalid_type_err(
1818
};
1919
let snippet = cx.sess.source_map().span_to_snippet(span).ok();
2020
match ast::LitKind::from_token_lit(token_lit) {
21+
Ok(ast::LitKind::CStr(_, _)) => {
22+
// TODO
23+
cx.span_err(span, "cannot concatenate C string litearls");
24+
}
2125
Ok(ast::LitKind::Char(_)) => {
2226
let sugg =
2327
snippet.map(|snippet| ConcatBytesInvalidSuggestion::CharLit { span, snippet });

‎compiler/rustc_expand/src/proc_macro_server.rs‎

Lines changed: 4 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -61,6 +61,8 @@ impl FromInternal<token::LitKind> for LitKind {
6161
token::StrRaw(n) => LitKind::StrRaw(n),
6262
token::ByteStr => LitKind::ByteStr,
6363
token::ByteStrRaw(n) => LitKind::ByteStrRaw(n),
64+
// TODO
65+
token::CStr | token::CStrRaw(_) => todo!(),
6466
token::Err => LitKind::Err,
6567
token::Bool => unreachable!(),
6668
}
@@ -436,6 +438,8 @@ impl server::FreeFunctions for Rustc<'_, '_> {
436438
| token::LitKind::StrRaw(_)
437439
| token::LitKind::ByteStr
438440
| token::LitKind::ByteStrRaw(_)
441+
| token::LitKind::CStr
442+
| token::LitKind::CStrRaw(_)
439443
| token::LitKind::Err => return Err(()),
440444
token::LitKind::Integer | token::LitKind::Float => {}
441445
}

‎compiler/rustc_hir/src/lang_items.rs‎

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -332,6 +332,7 @@ language_item_table! {
332332
RangeTo, sym::RangeTo, range_to_struct, Target::Struct, GenericRequirement::None;
333333

334334
String, sym::String, string, Target::Struct, GenericRequirement::None;
335+
CStr, sym::CStr, c_str, Target::Struct, GenericRequirement::None;
335336
}
336337

337338
pub enum GenericRequirement {

‎compiler/rustc_hir_typeck/src/fn_ctxt/checks.rs‎

Lines changed: 5 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -1300,6 +1300,11 @@ impl<'a, 'tcx> FnCtxt<'a, 'tcx> {
13001300
opt_ty.unwrap_or_else(|| self.next_float_var())
13011301
}
13021302
ast::LitKind::Bool(_) => tcx.types.bool,
1303+
ast::LitKind::CStr(_, _) => tcx.mk_imm_ref(
1304+
tcx.lifetimes.re_static,
1305+
tcx.type_of(tcx.require_lang_item(hir::LangItem::CStr, Some(lit.span)))
1306+
.skip_binder(),
1307+
),
13031308
ast::LitKind::Err => tcx.ty_error_misc(),
13041309
}
13051310
}

‎compiler/rustc_lexer/src/lib.rs‎

Lines changed: 30 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -186,12 +186,16 @@ pub enum LiteralKind {
186186
Str { terminated: bool },
187187
/// "b"abc"", "b"abc"
188188
ByteStr { terminated: bool },
189+
/// `c"abc"`, `c"abc`
190+
CStr { terminated: bool },
189191
/// "r"abc"", "r#"abc"#", "r####"ab"###"c"####", "r#"a". `None` indicates
190192
/// an invalid literal.
191193
RawStr { n_hashes: Option<u8> },
192194
/// "br"abc"", "br#"abc"#", "br####"ab"###"c"####", "br#"a". `None`
193195
/// indicates an invalid literal.
194196
RawByteStr { n_hashes: Option<u8> },
197+
/// `cr"abc"`, "cr#"abc"#", `cr#"a`. `None` is invalid.
198+
RawCStr { n_hashes: Option<u8> },
195199
}
196200

197201
#[derive(Clone, Copy, Debug, PartialEq, Eq, PartialOrd, Ord)]
@@ -391,6 +395,32 @@ impl Cursor<'_> {
391395
_ => self.ident_or_unknown_prefix(),
392396
},
393397

398+
// TODO deduplicate this code
399+
// c-string literal, raw c-string literal or identifier.
400+
'c' => match (self.first(), self.second()) {
401+
('"', _) => {
402+
self.bump();
403+
let terminated = self.double_quoted_string();
404+
let suffix_start = self.pos_within_token();
405+
if terminated {
406+
self.eat_literal_suffix();
407+
}
408+
let kind = CStr { terminated };
409+
Literal { kind, suffix_start }
410+
}
411+
('r', '"') | ('r', '#') => {
412+
self.bump();
413+
let res = self.raw_double_quoted_string(2);
414+
let suffix_start = self.pos_within_token();
415+
if res.is_ok() {
416+
self.eat_literal_suffix();
417+
}
418+
let kind = RawCStr { n_hashes: res.ok() };
419+
Literal { kind, suffix_start }
420+
}
421+
_ => self.ident_or_unknown_prefix(),
422+
},
423+
394424
// Identifier (this should be checked after other variant that can
395425
// start as identifier).
396426
c if is_id_start(c) => self.ident_or_unknown_prefix(),

0 commit comments

Comments
 (0)