From 822cc78d438e8df6a068d811d3cc825b883aba36 Mon Sep 17 00:00:00 2001 From: Martin Fischer Date: Fri, 1 Sep 2023 19:58:33 +0200 Subject: fix: too small char ref error spans --- src/machine.rs | 1 + src/tokenizer.rs | 9 +++++++++ 2 files changed, 10 insertions(+) (limited to 'src') diff --git a/src/machine.rs b/src/machine.rs index 149b64d..84fb525 100644 --- a/src/machine.rs +++ b/src/machine.rs @@ -1659,6 +1659,7 @@ where } }, State::CharacterReference => { + slf.some_offset = slf.reader.position() - "&".len(); slf.temporary_buffer.clear(); slf.temporary_buffer.push('&'); match slf.read_char()? { diff --git a/src/tokenizer.rs b/src/tokenizer.rs index f41f5ae..cecc0f6 100644 --- a/src/tokenizer.rs +++ b/src/tokenizer.rs @@ -50,6 +50,7 @@ pub struct Tokenizer> { last_start_tag_name: String, is_start_tag: bool, /// * Set to the start offset of `, O: Offset, E: Emitter> Tokenizer { | Error::MissingSemicolonAfterCharacterReference => { self.reader.position()..self.reader.position() } + Error::AbsenceOfDigitsInNumericCharacterReference + | Error::NullCharacterReference + | Error::CharacterReferenceOutsideUnicodeRange + | Error::SurrogateCharacterReference + | Error::NoncharacterCharacterReference + | Error::ControlCharacterReference + | Error::UnknownNamedCharacterReference => self.some_offset..self.reader.position(), + _ => self.reader.position() - 1..self.reader.position(), }; self.emitter.emit_error(error, span); -- cgit v1.2.3