451654 - [silme] New file-parser needed for bookmarks.html

Status: RESOLVED WONTFIX

(Mozilla Localizations :: Infrastructure, defect)

Description

[Adrian Kalla [:adriank]]

To have full compatibility of compare-locales for Silme we need to have a bookmarks file parser to parse the bookmarks.html file. 
Probably we can use much of the code of the parser which already is in compare-locales.

Comment 1

[Zibi Braniecki [:zbraniecki][:gandalf]]

So, here I was thinking about building a support for HTML files overall.

You can't cove all, but we can cover pretty much a lot. An ID of such entity would be it's XPath path, and w could cover all all texts and value="*"/label="" properties.

Then on top of it we can build a very simple bookmarks.html parser that would reuse html parser.

What do you think?

Comment 2

[Adrian Kalla [:adriank]]

If we want to parse all HTML files, the question will be, if we will need to use a Tidy tool first to make the HTML file XML-valid?
The last time I worked with XPaths on HTML files I had to Tidy them first - but that was about 3 years ago...

Comment 3

[Axel Hecht [:Pike]]

I'm not sure how much effort we should spend on doing bookmarks.html compat. The logic to check is somewhat peculiar, and I'm removing it from at least hg anyway.

Comment 4

[Adrian Kalla [:adriank]]

Pike: so we can leave this bug open as I nice-to-have feature, with a low priority and not blocking anymore bug 452311 ?
Maybe if someone implements a full HTML parser for Silme we could try to use it for bookmarks.html...

Comment 5

[Axel Hecht [:Pike]]

Yeah. I'm not sure if an html localization scheme would be a good fit for silme.

HTML is full of placeables, and I have no idea yet on how to handle those.

Comment 6

[Adrian Kalla [:adriank]]

Axel, Gandalf: as 1.9.1 is approaching, and we do not have bookmarks.html there anymore: I think we do not need this feature.
WONTFIX?

Comment 7

[Axel Hecht [:Pike]]

Yes, WONTIFXed. Dropping dependencies for this bug, too, to make those simpler.