Summary: | Feed libxml2 with raw data, relying on it to do character set decoding | ||||||||
---|---|---|---|---|---|---|---|---|---|
Product: | WebKit | Reporter: | Patrick R. Gansterer <paroga> | ||||||
Component: | XML | Assignee: | Patrick R. Gansterer <paroga> | ||||||
Status: | NEW --- | ||||||||
Severity: | Normal | CC: | ap, darin | ||||||
Priority: | P2 | ||||||||
Version: | 528+ (Nightly build) | ||||||||
Hardware: | All | ||||||||
OS: | All | ||||||||
Bug Depends on: | 52547, 52085, 53398 | ||||||||
Bug Blocks: | 43085 | ||||||||
Attachments: |
|
Created attachment 79118 [details] Work in progress I did some small performance tests (see bug 52547) with this new patch: avg median stdev min max HTML 6517.25 6770.5 505 5242 7286 XML (original) 6254.5 6366 573 5462 7118 XML (with patch) 5735.45 5385.5 704 4159 6853 -8.3% -15.4% |
Created attachment 78193 [details] Work in progress I created a patch of the work I've done already. Maybe you can give me some early feedback. At the moment only about 5 test fail, because of some missing encoding problem. (I don't teach libxml2 about all known TextEncodings in the current state). XMLDocumentParser is a subclass of ScriptableDocumentParser which is a subclass of DecodedDataDocumentParser. Normally DecodedDataDocumentParser handles the appendBytes method, which I've implemented in the XMLDocumentParser to get the raw data. IMHO this is a kind of "layer violation". Can you give me a tip how to implement this in a correct way? Do I need to change the inheritance of all "parser classes"?