Bug 5996 – SVG <view> is unimplemented

• Component: SVG
• Status: RESOLVED FIXED

[Eric Seidel (no email)]

SVG <view> is unimplemented

It actually doesn't look that hard.  It's the initial zoom/pan to apply to the document when activating a 
link, that's all.  Shouldn't be very difficult to build up the DOM objects, and then it's just a bit of wiring into 
the linking system.

[Alexander Kellett]

possibly useful test

LayoutTests/svg/W3C-SVG-1.1/resources/linking-uri-01-b.svg

[Eric Seidel (no email)]

I'm sorta surprised no one has jumped on this yet.  This would be (IMO) a fun, easy fix to make.

[Eric Seidel (no email)]

Ok, maybe this isn't so super easy to fix.

[Eric Seidel (no email)]

Created attachment 10216 [details]
rough outline patch (doens't actually work)

[Eric Seidel (no email)]

Here is another test of view fragment urls:
http://www.w3.org/Graphics/SVG/Test/20061213/htmlEmbedHarness/full-linking-a-03-b.html

[Eric Seidel (no email)]

and another:
http://www.w3.org/Graphics/SVG/Test/20061213/htmlEmbedHarness/full-linking-uri-01-b.html

[Eric Seidel (no email)]

Rob had expressed interest in fixing this.

[Rob Buis]

(In reply to comment #7)
> Rob had expressed interest in fixing this.
> 

and still does.
Cheers,

Rob.

[Rob Buis]

Created attachment 15361 [details]
Work in progress

This patch is almost complete, it is based on Eric's patch. There are some bugs to fix and ofcourse some tests to create, but from my simple testing it seems to work fine. I hope to have something reviewable soon :)
Cheers,

Rob.

[Nikolas Zimmermann]

(In reply to comment #9)
> Created an attachment (id=15361) [edit]
> Work in progress
> 
> This patch is almost complete, it is based on Eric's patch. There are some bugs
> to fix and ofcourse some tests to create, but from my simple testing it seems
> to work fine. I hope to have something reviewable soon :)
> Cheers,
> 
> Rob.

Excellent work! Just had a quick look, can't do real reviewing atm - as I'm blocked by
exams (Saturday!). I love the RenderSVGRoot code removals though...

What bugs are still existant? Did you file follow up bugs already for these?

Greetings,
Niko

[Rob Buis]

Created attachment 15381 [details]
Complete code patch, no tests yet

As the description says, code is complete, I am not sure how to tests yet though. Obviously the manual tests work. If it can be automated I need to find some examples of how to do tests where a link is followed. The code part can be reviewed now though I think.
Cheers,

Rob.

[Oliver Hunt]

Comment on attachment 15381 [details]
Complete code patch, no tests yet

the "static const UChar svgViewSpec[] = {'s','v','g','V', 'i', 'e', 'w'};" declarations may trigger our initialiser logic, so you may ned to turn them into functions wrapping scoped globals...

Other thasn that it looks fine, but this wasn't as thorough review as really necessary

[Rob Buis]

Created attachment 15504 [details]
Now with testcases

A first shot at some testcases.
Cheers,

Rob.

[Nikolas Zimmermann]

(In reply to comment #13)
> Created an attachment (id=15504) [edit]
> Now with testcases
> 
> A first shot at some testcases.
> Cheers,
> 
> Rob.

Hey Rob,

I'm very tempted to r+ it now. Did you think about Oliver's issue?
I think that's the last thing holding it back from r+.

One last thing:
// FIXME: All this setup should be done after attributesChanged, not here.
This comment is outdated in SVGSVGElement::createRenderer.
Maybe just move it to the approriate place?

Greetings,
Niko

[Nikolas Zimmermann]

Comment on attachment 15504 [details]
Now with testcases

Good job.

[Rob Buis]

Landed in r24349.