Bug 75257
| Summary: | XMLDocumentParser needs a PreloadScanner | ||
|---|---|---|---|
| Product: | WebKit | Reporter: | Tony Gentilcore <tonyg> |
| Component: | XML | Assignee: | Nobody <webkit-unassigned> |
| Status: | NEW | ||
| Severity: | Normal | CC: | abarth, ap, eric, koivisto, simonjam, skyul, tonyg |
| Priority: | P2 | ||
| Version: | 528+ (Nightly build) | ||
| Hardware: | Unspecified | ||
| OS: | Unspecified | ||
Tony Gentilcore
The XMLDocumentParser doesn't use the PreloadScanner like the HTMLDocumentParser does. In cases like the one described in http://crbug.com/108592 , this leads to very poor performance.
This is also mentioned in http://gent.ilcore.com/2011/01/webkit-preloadscanner.html .
I thought we had a bug about this, but can't find it. Please dupe this one if you remember where that is.
| Attachments | ||
|---|---|---|
| Add attachment proposed patch, testcase, etc. |
Tony Gentilcore
I seem to also recall abarth asking why we need to pause the parser at all for scripts in XHTML documents (since there is no doc.write). We should probably investigate that further first.
Adam Barth
Note that preload scanning XML is much easier because we know exactly how the whole document will parse (e.g. there is no document.write).