Fix a few CSS selector issues #178
Conversation
- fix standalone #ID (was working only as ELEM#ID) - fix non-lowercase element name (elements are internally lowercased, so we should lowercase there too to expect any match) - fix E+F selector, that should ignore immediate preceding text nodes, and consider the first met element node - adds E~F selector (like E+F, but *any* instead of "immediate" precedessor is considered for a match)
| @@ -94,6 +94,7 @@ enum LVCssSelectorRuleType | |||
| cssrt_parent, // E > F | |||
| cssrt_ancessor, // E F | |||
| cssrt_predecessor, // E + F | |||
| cssrt_predsibling, // E ~ F | |||
| else if ( css_is_alpha( *str ) ) | ||
| { | ||
| // ident | ||
| char ident[64]; | ||
| if (!parse_ident( str, ident )) | ||
| return false; | ||
| _id = doc->getElementNameIndex( lString16(ident).c_str() ); | ||
| _id = doc->getElementNameIndex( lString16(ident).lowercase().c_str() ); |
There was a problem hiding this comment.
What's the effect on the XML vs HTML split?
There was a problem hiding this comment.
HTML (including those found in EPUB) are parsed by HTMLParser (a subclass of XMLParser that has a setting m_citags=true - unlike XMLParser that has it false) for case-insensitive tags, which make it lowercasing the element names. So, they are all lowercased internally when added to the DOM tree.
There was a problem hiding this comment.
doc->getElementNameIndex( "DIV" ) would work (if I remember correctly) like: oh, you're asking me for DIV, I haven't seen it yet, let's make a new ID for DIV for you, here it is - but there would be no element with that new ID in the DOM.
There was a problem hiding this comment.
And this CSS rule only applies to HTML?
There was a problem hiding this comment.
to each DOM node: Selectors gets that node's elementnameID (which will always be lowercase) and compares it to its _id (which is now lowercase).
I don't think we build DOM from anything but HTML.
There was a problem hiding this comment.
So what happens to the XHTML? I thought you just said it's handled case sensitive? :-P
There was a problem hiding this comment.
Both standalone html and html/xhtml found in EPUB are parsed by the lowercasing HTMLParser (only other opf/ncx xml stuff in the epub are parsed by the XMLParser).
Then, this HTMLParser feeds "documentWriter" with tag/attribute/text.
standalone HTML uses ldomDocumentWriterFilter, which handle autoclose and LIB.RU.
EPUB uses ldomDocumentFragmentWriter, that doesn't do autoclose, but deals with appending multiple body from multiple HTML as docFragment into a single DOM.
But both are fed with lowercased element tag names.
There was a problem hiding this comment.
Weird, but okay in practice I suppose. :-p
Fix some of the issues reported in #176.