Summary: | Stylesheet rule cssText is all lowercase | ||||||
---|---|---|---|---|---|---|---|
Product: | WebKit | Reporter: | Moxiecode Systems <spam> | ||||
Component: | CSS | Assignee: | Nobody <webkit-unassigned> | ||||
Status: | RESOLVED WONTFIX | ||||||
Severity: | Major | CC: | ap, eric, ganesh.bksv, hyatt, mustaf.here, rniwa, sarap.karthik, simon.fraser | ||||
Priority: | P4 | ||||||
Version: | 528+ (Nightly build) | ||||||
Hardware: | All | ||||||
OS: | All | ||||||
URL: | http://tinymce.moxiecode.com/safari/stylesheet_case.html | ||||||
Attachments: |
|
Description
Moxiecode Systems
2011-03-02 14:43:47 PST
I suspect that fixing this bug requires considerable amount of work. I don't think any spec requires that case is round-tripped. This issue is because of the rule we have for class: in CSSGrammar.y in the following lines if (!p->m_strict) $2.lower(); While loading the page, the call flow is CSSStyleSheet::parseStringAtLine() => CSSParser::parseSheet()=> cssyyparse(this); And while parsing as per the above logic, we convert the class string to lower case if the parser is running in NON-strict mode We have verified that if we add DOCTYPE on top of my html (which makes the CSSParser in strict mode), the style class info is shown maintaining proper casing given in the html page. Therefore, Does any one know why do we have that logic(as mentioned above) of converting string to lowercase for non-strict mode in CSSGrammar.y? Based on this, we can decide whether or not to fix the issue It's a documented quirk of Mozilla's quirks mode: https://developer.mozilla.org/en/mozilla_quirks_mode_behavior Or rather the case-insensitivity is: In quirks mode css class names are case insensitive. In standards mode they are case sensitive. class name is nothing to do with this bug. This bug is about cssText being lower case. (In reply to comment #5) > Or rather the case-insensitivity is: > > In quirks mode css class names are case insensitive. In standards mode they are case sensitive. My interpretation is if a html page does not mention any DOCTYPE, then it is by default quirks mode, correct? Then does it mean Mozilla has an issue here because in quirks mode also FF is respecting the CSS class names, which it should not do as per link given above? Created attachment 107614 [details]
modifed test case
Attached the modifed test case . Just Added <!DOCTYPE html> to the actual test case.
http://www.w3.org/TR/html5-diff/#doctype In my opinion, the above link strengthens comments 5 & 7. As per the spec, browser need not render the page in standars mode unless <!DOCTYPE html> is mentioned. If I am correct, can some one move this to INVALID/WONTFIX state. |