This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
Bug https://www.w3.org/Bugs/Public/show_bug.cgi?id=17558 describes that --> needs to be banned from the allowed syntax in cue data. However, when XML data is carried in metadata tracks, the XML comments end tag uses the sequence of characters -->. This should be acceptable to a parser, since metadata tracks don't use escapes.
We don't support XML comments or empty lines in cue payloads. Both cause the parser to assume the cue has ended.
I can understand why we wouldn't want to support empty lines. But I think we could tolerate --> if we would only look for the next "-->" after the next empty line. The less restrictions we pose on cue text, the less broken files we will get.
https://www.w3.org/Bugs/Public/show_bug.cgi?id=14550
Should we have different parsing rules for metadata tracks?
You don't know what's a metadata track when parsing. It would be crazy to take a WebVTT file linked to as a metadata cue file, switch it to be a caption file, and suddenly find you had different cues.
If we make it dependent on the presence of a "Kind: metadata" marker in the header of the file, we could do that.