Bug 117443
| Summary: | Consider using HTML parser to parse SVG | ||
|---|---|---|---|
| Product: | WebKit | Reporter: | Ryosuke Niwa <rniwa> |
| Component: | SVG | Assignee: | Nobody <webkit-unassigned> |
| Status: | RESOLVED WONTFIX | ||
| Severity: | Normal | CC: | ap, krit, mjs, ossy, pdr, rhodovan.u-szeged, sam, zherczeg, zimmermann |
| Priority: | P2 | Keywords: | BlinkMergeCandidate |
| Version: | 528+ (Nightly build) | ||
| Hardware: | Unspecified | ||
| OS: | Unspecified | ||
Ryosuke Niwa
https://chromium.googlesource.com/chromium/blink/+/b69618018614278cda72077611adc093f460dc57
Parse SVG as HTML
This change switches the SVG parser to be the HTML parser all the time. This
work is being done behind a flag to gauge the potential compat risk.
This is an interesting experiment.
| Attachments | ||
|---|---|---|
| Add attachment proposed patch, testcase, etc. |
Dirk Schulze
The experiment kind of failed.
HTML parser does not allow entities. There are thousands of SVGs out there which use custom XML entities. (Most exported by Illustrator :(.)
Either the HTML parser allows a legacy mode with entities or we can not proceed with this idea.
Philip Rogers
(In reply to comment #1)
> The experiment kind of failed.
>
> HTML parser does not allow entities. There are thousands of SVGs out there which use custom XML entities. (Most exported by Illustrator :(.)
>
> Either the HTML parser allows a legacy mode with entities or we can not proceed with this idea.
I discussed this with Kouhei and we don't think a basic level of xml entity support would be too hard to implement in the html parser.
Dirk Schulze
(In reply to comment #2)
> (In reply to comment #1)
> > The experiment kind of failed.
> >
> > HTML parser does not allow entities. There are thousands of SVGs out there which use custom XML entities. (Most exported by Illustrator :(.)
> >
> > Either the HTML parser allows a legacy mode with entities or we can not proceed with this idea.
>
> I discussed this with Kouhei and we don't think a basic level of xml entity support would be too hard to implement in the html parser.
Does the HTML parser handle renaming of ns shorthands?
Instead of
xmlns:xlink="http://www.w3.org/1999/xlink"
this
xmlns:l="http://www.w3.org/1999/xlink" and l:href
This was another pattern we saw.
Ryosuke Niwa
As far as I know, this experiment failed due to XSLT use.