Skip to content

Commit 584406e

Browse files
MingunMingun
committed
Inline ReaderState::emit_bang with obvious dead code elimination (removed unused arms in match)
1 parent 40c1cfc commit 584406e

File tree

1 file changed

+82
-82
lines changed

1 file changed

+82
-82
lines changed

‎src/reader/state.rs‎

Lines changed: 82 additions & 82 deletions
Original file line numberDiff line numberDiff line change
@@ -83,86 +83,6 @@ pub(super) struct ReaderState {
8383
}
8484

8585
impl ReaderState {
86-
/// reads `BytesElement` starting with a `!`,
87-
/// return `Comment`, `CData` or `DocType` event
88-
pub fn emit_bang<'b>(&mut self, bang_type: BangType, buf: &'b [u8]) -> Result<Event<'b>> {
89-
let uncased_starts_with = |string: &[u8], prefix: &[u8]| {
90-
string.len() >= prefix.len() && string[..prefix.len()].eq_ignore_ascii_case(prefix)
91-
};
92-
93-
let len = buf.len();
94-
match bang_type {
95-
BangType::Comment if buf.starts_with(b"!--") => {
96-
debug_assert!(buf.ends_with(b"--"));
97-
if self.config.check_comments {
98-
// search if '--' not in comments
99-
let mut haystack = &buf[3..len - 2];
100-
let mut off = 0;
101-
while let Some(p) = memchr::memchr(b'-', haystack) {
102-
off += p + 1;
103-
// if next byte after `-` is also `-`, return an error
104-
if buf[3 + off] == b'-' {
105-
// Explanation of the magic:
106-
//
107-
// - `self.offset`` just after `>`,
108-
// - `buf` contains `!-- con--tent --`
109-
// - `p` is counted from byte after `<!--`
110-
//
111-
// <!-- con--tent -->:
112-
// ~~~~~~~~~~~~~~~~ : - buf
113-
// : =========== : - zone of search (possible values of `p`)
114-
// : |---p : - p is counted from | (| is 0)
115-
// : : : ^ - self.offset
116-
// ^ : : - self.offset - len
117-
// ^ : - self.offset - len + 2
118-
// ^ - self.offset - len + 2 + p
119-
self.last_error_offset = self.offset - len + 2 + p;
120-
return Err(Error::IllFormed(IllFormedError::DoubleHyphenInComment));
121-
}
122-
// Continue search after single `-` (+1 to skip it)
123-
haystack = &haystack[p + 1..];
124-
}
125-
}
126-
Ok(Event::Comment(BytesText::wrap(
127-
// Cut of `!--` and `--` from start and end
128-
&buf[3..len - 2],
129-
self.decoder(),
130-
)))
131-
}
132-
BangType::CData if uncased_starts_with(buf, b"![CDATA[") => {
133-
debug_assert!(buf.ends_with(b"]]"));
134-
Ok(Event::CData(BytesCData::wrap(
135-
// Cut of `![CDATA[` and `]]` from start and end
136-
&buf[8..len - 2],
137-
self.decoder(),
138-
)))
139-
}
140-
BangType::DocType if uncased_starts_with(buf, b"!DOCTYPE") => {
141-
match buf[8..].iter().position(|&b| !is_whitespace(b)) {
142-
Some(start) => Ok(Event::DocType(BytesText::wrap(
143-
// Cut of `!DOCTYPE` and any number of spaces from start
144-
&buf[8 + start..],
145-
self.decoder(),
146-
))),
147-
None => {
148-
// Because we here, we at least read `<!DOCTYPE>` and offset after `>`.
149-
// We want report error at place where name is expected - this is just
150-
// before `>`
151-
self.last_error_offset = self.offset - 1;
152-
return Err(Error::IllFormed(IllFormedError::MissingDoctypeName));
153-
}
154-
}
155-
}
156-
_ => {
157-
// <!....>
158-
// ^^^^^ - `buf` does not contain `<` and `>`, but `self.offset` is after `>`.
159-
// ^------- We report error at that position, so we need to subtract 2 and buf len
160-
self.last_error_offset = self.offset - len - 2;
161-
Err(bang_type.to_err())
162-
}
163-
}
164-
}
165-
16686
/// Wraps content of `buf` into the [`Event::End`] event. Does the check that
16787
/// end name matches the last opened start name if `self.config.check_end_names` is set.
16888
pub fn emit_end<'b>(&mut self, buf: &'b [u8]) -> Result<Event<'b>> {
@@ -417,7 +337,57 @@ impl ReaderState {
417337
debug_assert!(content.starts_with(b"<!--"), "{:?}", Bytes(content));
418338
debug_assert!(content.ends_with(b"-->"), "{:?}", Bytes(content));
419339

420-
self.emit_bang(BangType::Comment, &content[1..content.len() - 1])
340+
let bang_type = BangType::Comment;
341+
let buf = &content[1..content.len() - 1];
342+
let uncased_starts_with = |string: &[u8], prefix: &[u8]| {
343+
string.len() >= prefix.len() && string[..prefix.len()].eq_ignore_ascii_case(prefix)
344+
};
345+
346+
let len = buf.len();
347+
match bang_type {
348+
BangType::Comment if buf.starts_with(b"!--") => {
349+
debug_assert!(buf.ends_with(b"--"));
350+
if self.config.check_comments {
351+
// search if '--' not in comments
352+
let mut haystack = &buf[3..len - 2];
353+
let mut off = 0;
354+
while let Some(p) = memchr::memchr(b'-', haystack) {
355+
off += p + 1;
356+
// if next byte after `-` is also `-`, return an error
357+
if buf[3 + off] == b'-' {
358+
// Explanation of the magic:
359+
//
360+
// - `self.offset`` just after `>`,
361+
// - `buf` contains `!-- con--tent --`
362+
// - `p` is counted from byte after `<!--`
363+
//
364+
// <!-- con--tent -->:
365+
// ~~~~~~~~~~~~~~~~ : - buf
366+
// : =========== : - zone of search (possible values of `p`)
367+
// : |---p : - p is counted from | (| is 0)
368+
// : : : ^ - self.offset
369+
// ^ : : - self.offset - len
370+
// ^ : - self.offset - len + 2
371+
// ^ - self.offset - len + 2 + p
372+
self.last_error_offset = self.offset - len + 2 + p;
373+
return Err(Error::IllFormed(IllFormedError::DoubleHyphenInComment));
374+
}
375+
haystack = &haystack[p + 1..];
376+
}
377+
}
378+
Ok(Event::Comment(BytesText::wrap(
379+
&buf[3..len - 2],
380+
self.decoder(),
381+
)))
382+
}
383+
_ => {
384+
// <!....>
385+
// ^^^^^ - `buf` does not contain `<` and `>`, but `self.offset` is after `>`.
386+
// ^------- We report error at that position, so we need to subtract 2 and buf len
387+
self.last_error_offset = self.offset - len - 2;
388+
Err(bang_type.to_err())
389+
}
390+
}
421391
}
422392
FeedResult::EmitDoctype(_) => {
423393
debug_assert!(content.len() > 9, "{:?}", Bytes(content));
@@ -428,7 +398,37 @@ impl ReaderState {
428398
);
429399
debug_assert!(content.ends_with(b">"), "{:?}", Bytes(content));
430400

431-
self.emit_bang(BangType::DocType, &content[1..content.len() - 1])
401+
let bang_type = BangType::DocType;
402+
let buf = &content[1..content.len() - 1];
403+
let uncased_starts_with = |string: &[u8], prefix: &[u8]| {
404+
string.len() >= prefix.len() && string[..prefix.len()].eq_ignore_ascii_case(prefix)
405+
};
406+
407+
let len = buf.len();
408+
match bang_type {
409+
BangType::DocType if uncased_starts_with(buf, b"!DOCTYPE") => {
410+
match buf[8..].iter().position(|&b| !is_whitespace(b)) {
411+
Some(start) => Ok(Event::DocType(BytesText::wrap(
412+
&buf[8 + start..],
413+
self.decoder(),
414+
))),
415+
None => {
416+
// Because we here, we at least read `<!DOCTYPE>` and offset after `>`.
417+
// We want report error at place where name is expected - this is just
418+
// before `>`
419+
self.last_error_offset = self.offset - 1;
420+
return Err(Error::IllFormed(IllFormedError::MissingDoctypeName));
421+
}
422+
}
423+
}
424+
_ => {
425+
// <!....>
426+
// ^^^^^ - `buf` does not contain `<` and `>`, but `self.offset` is after `>`.
427+
// ^------- We report error at that position, so we need to subtract 2 and buf len
428+
self.last_error_offset = self.offset - len - 2;
429+
Err(bang_type.to_err())
430+
}
431+
}
432432
}
433433
FeedResult::EmitPI(_) => {
434434
debug_assert!(content.starts_with(b"<?"), "{:?}", Bytes(content));

0 commit comments

Comments
 (0)