RE: Live WebVTT Validator update

Hi Anne, and other WebVTT experts,

I'm experimenting with a TTML to WebVTT converter and just want to check a few details with you all.

The validator claims that:

WEBVTT { generated-from : TTML; }

/* ============ cue style here =========
-- styles here ignored by webvtt
*/
*::cue(c.style25)
{
    font-family: "Segoe UI Light", Sans-Serif;
    color: #303030;
    font-size: 36px;
    line-height: 44px;
    margin: 0;
}
/* cues here ignored by css

00:00:00.101 --> 00:00:03.000
<c.style22><c.style25><c.style26><c.style27> This is an <c.style28>SDP-US</c> TTML file. </c></c></c></c>

NOTE
end of file   */

Is incorrect, specifically:

1.Line 4, column 1: Timestamp must start with a character in the range 0-9.
2.Line 20, column 1: Timestamp must start with a character in the range 0-9.

My understanding of the parser is that content is ignored between the header up to the first timestamp line, is that not correct? I want to use that area for the generated style-sheet,  rather than have to put it in a separate file (i.e. the generated file serves both as cues and its own style).

Also it seems that *::cue(c.style25) is not a legal CSS 3 selector, since the syntax for functions in pseudo classes only admits an expression as argument, not a selector. Is that correct?


Thanks,

Sean.

From: Anne van Kesteren [mailto: annevk@annevk.nl]
To: public-texttracks@w3.org
Subject: Live WebVTT Validator update


Hey,



I updated http://quuz.org/webvtt/ (source code:

https://github.com/annevk/webvtt ), feedback welcome. In particular I

added support for the new escapes, align values, <lang> (though no BCP

47 validation, patches welcome), ability to select kind

(metadata/chapters), and cue identifiers that start with NOTE are now

treated as comments by the parser and won't trigger an error.



https://github.com/annevk/webvtt/commits has a more detailed overview

of the changes.



The parser not flagging conformance errors is kind of annoying, but so

far it's manageable to merge the conformance requirements with the

requirements of the parser, although there's probably a bunch of bugs.



Cheers,



PS: I'm not subscribed to this list, so you might want to cc me on replies.





--

http://annevankesteren.nl/