824327 - Convert direct subclasses of SVGElement to WebIDL

Status: RESOLVED FIXED

(Core :: SVG, defect)

Comment 1

[David Zbarsky (:dzbarsky)]

Attachment: Part 1: Add a namespaced version of NS_NewxxxSVGElement

Comment 2

[David Zbarsky (:dzbarsky)]

Attachment: SVGDescElement

Comment 3

[David Zbarsky (:dzbarsky)]

Attachment: SVGTitleElement

Comment 4

[David Zbarsky (:dzbarsky)]

Attachment: SVGStopElement

Comment 5

[David Zbarsky (:dzbarsky)]

Attachment: SVGStyleElement

Comment 6

[David Zbarsky (:dzbarsky)]

Attachment: SVGMetadataElement

Comment 7

[David Zbarsky (:dzbarsky)]

Attachment: SVGDescElement

Comment 8

[David Zbarsky (:dzbarsky)]

Attachment: SVGTitleElement

Comment 9

[David Zbarsky (:dzbarsky)]

Attachment: SVGStopElement

Comment 10

[David Zbarsky (:dzbarsky)]

Attachment: SVGStyleElement

Comment 11

[David Zbarsky (:dzbarsky)]

Attachment: SVGMetadataElement

Comment 12

[David Zbarsky (:dzbarsky)]

Attachment: SVGMPathElement

Comment 13

[David Zbarsky (:dzbarsky)]

Attachment: SVGScriptElement

Comment 14

[David Zbarsky (:dzbarsky)]

Attachment: SVGScriptElement

Comment 15

[Boris Zbarsky [:bzbarsky]]

Comment on attachment 695429 [details] [diff] [review]
Part 1: Add a namespaced version of NS_NewxxxSVGElement

Why this instead of keeping the NS_NewSVGWhateverElement functions but either use a different macro name (as this patch) that creates the namespaced version or do what Peter did for HTML and have the existing macro auto-detect the element concrete class?

Comment 16

[David Zbarsky (:dzbarsky)]

Attachment: Part 1: Add a namespaced version of NS_NewxxxSVGElement

Comment 17

[Boris Zbarsky [:bzbarsky]]

Comment on attachment 697665 [details] [diff] [review]
Part 1: Add a namespaced version of NS_NewxxxSVGElement

r=me

Comment 18

[Boris Zbarsky [:bzbarsky]]

Comment on attachment 695671 [details] [diff] [review]
SVGDescElement

>+#ifndef SVGDescElement_h

I think Peter and I have been using mozilla_dom_SVGDescElement_h

r=me with that and the NS_New thing updated.

Comment 19

[Boris Zbarsky [:bzbarsky]]

Comment on attachment 695671 [details] [diff] [review]
SVGDescElement

Oh and:

>+#include "SVGDescElement.h"

I'd prefer mozilla/dom/SVGDescElement.h.  Yes, that can matter on Windows, kinda.

Comment 20

[Boris Zbarsky [:bzbarsky]]

Comment on attachment 695672 [details] [diff] [review]
SVGTitleElement

>+++ b/browser/base/content/browser.js

Probably better, for now, to set 'hasInstanceInterface': 'nsIDOMSVGTitleElement' in Bindings.conf for SVGTitleElement.  Especially because you missed a comm-central use equivalent to the one here...

>+++ b/content/svg/content/src/SVGTitleElement.cpp
>+#include "SVGTitleElement.h"

mozilla/dom/SVGTitleElement.h

>+++ b/content/svg/content/src/SVGTitleElement.h
>+#ifndef SVGTitleElement_h

And mozilla_dom_SVGTitleElement_h

>+  virtual JSObject* WrapNode(JSContext *aCx, JSObject *aScope, bool *aTriedToWrap);

MOZ_OVERRIDE.  I missed that on SVGDescElement, so add it there too?

r=me with those nits.

Comment 21

[Boris Zbarsky [:bzbarsky]]

Comment on attachment 695673 [details] [diff] [review]
SVGStopElement

>+++ b/content/svg/content/src/SVGStopElement.cpp
>+#include "SVGStopElement.h"

mozilla/dom/SVGStopElement.h

>+// sSVGElement methods

Drop that first 's' too?

>+++ b/content/svg/content/src/SVGStopElement.h
>+#ifndef SVGStopElement_h

mozilla_dom_SVGStopElement_h

r=me

Comment 22

[Boris Zbarsky [:bzbarsky]]

Comment on attachment 695674 [details] [diff] [review]
SVGStyleElement

>+++ b/content/svg/content/src/SVGStyleElement.cpp
>+#include "SVGStyleElement.h"

mozilla/dom/SVGStyleElement.h

>+++ b/content/svg/content/src/SVGStyleElement.h
>+#ifndef SVGStyleElement_h

mozilla_dom_SVGStyleElement_h

How about we assume those two comments for all these patches and I stop making them?

>+++ b/dom/webidl/SVGStyleElement.webidl
>+  attribute DOMString xmlspace; // Spec claims this should be on SVGElement
>+  attribute DOMString type;
>+  attribute DOMString media;
>+  attribute DOMString title;

These should all be [SetterThrows].  You can use the XPCOM getters fine, but the setters will need to look somewhat like SetHTMLAttr (the version that takes an ErrorResult).

Maybe we should hoist an equivalent of SetHTMLAttr on Element...

r=me with that fixed.

Comment 23

[Boris Zbarsky [:bzbarsky]]

Comment on attachment 695676 [details] [diff] [review]
SVGMetadataElement

With the usual bits about the include guard and the #include in the cpp, r=me.

Comment 24

[Boris Zbarsky [:bzbarsky]]

Comment on attachment 695678 [details] [diff] [review]
SVGMPathElement

With the usual bits about the include guard and the #include in the cpp, r=me.

Comment 25

[Boris Zbarsky [:bzbarsky]]

Comment on attachment 696244 [details] [diff] [review]
SVGScriptElement

>+++ b/content/html/content/test/reflect.js

Peter has convinced me that we should do the explicit list thing here... Would you mind doing that?

It looks like the SVGScriptElement.h was created via hg add, not hg cp.  I'd prefer hg cp.

Usual comments about the include guard and #include line.

Usual comment about MOZ_OVERRIDE on WrapNode (which I think I forgot on some earlier patches).

>+++ b/dom/webidl/SVGScriptElement.webidl
>+  attribute DOMString type;
>+  attribute DOMString crossOrigin;

Both of these need to be [SetterThrows]; see comments above for SVGStyleElement.

r=me with those fixed.

Comment 26

[:Ms2ger (he/him; ⌚ UTC+1/+2)]

Comment on attachment 695678 [details] [diff] [review]
SVGMPathElement

Review of attachment 695678 [details] [diff] [review]:
-----------------------------------------------------------------

::: dom/webidl/SVGMPathElement.webidl
@@ +10,5 @@
> + * liability, trademark and document use rules apply.
> + */
> +
> +interface SVGMPathElement : SVGElement {
> +};

Hrm, this used to be called SVGMpathElement (lower case p), no? If so, you'll probably need to add something to a test_interfaces.html

Comment 27

[David Zbarsky (:dzbarsky)]

(In reply to :Ms2ger from comment #26)
> Comment on attachment 695678 [details] [diff] [review]
> SVGMPathElement
> 
> Review of attachment 695678 [details] [diff] [review]:
> -----------------------------------------------------------------
> 
> ::: dom/webidl/SVGMPathElement.webidl
> @@ +10,5 @@
> > + * liability, trademark and document use rules apply.
> > + */
> > +
> > +interface SVGMPathElement : SVGElement {
> > +};
> 
> Hrm, this used to be called SVGMpathElement (lower case p), no? If so,
> you'll probably need to add something to a test_interfaces.html

I think XPConnect still exposes this interface as SVGMpathElement

Comment 28

[David Zbarsky (:dzbarsky)]

Landed everything except SVGScriptElement.
remote:   https://hg.mozilla.org/integration/mozilla-inbound/rev/68ded29fc68e
remote:   https://hg.mozilla.org/integration/mozilla-inbound/rev/60c936f77dce
remote:   https://hg.mozilla.org/integration/mozilla-inbound/rev/2a2fdd3161f8
remote:   https://hg.mozilla.org/integration/mozilla-inbound/rev/00e9c72fa410
remote:   https://hg.mozilla.org/integration/mozilla-inbound/rev/83dd11d6dc7f
remote:   https://hg.mozilla.org/integration/mozilla-inbound/rev/c4470a13f216
remote:   https://hg.mozilla.org/integration/mozilla-inbound/rev/13acd3b74386

Comment 29

[David Zbarsky (:dzbarsky)]

(In reply to Boris Zbarsky (:bz) from comment #25)
> Comment on attachment 696244 [details] [diff] [review]
> SVGScriptElement
> 
> >+++ b/content/html/content/test/reflect.js
> 
> Peter has convinced me that we should do the explicit list thing here...
> Would you mind doing that?
> 

reflect.js is only used for HTML elements and the SVG script element, so I think the easiest solution may be to convert HTMLScriptElement and land it at the same time as SVGScriptElement.

Comment 30

[Cameron McCormack (:heycam)]

(In reply to David Zbarsky (:dzbarsky) from comment #27)
> > Hrm, this used to be called SVGMpathElement (lower case p), no? If so,
> > you'll probably need to add something to a test_interfaces.html
> 
> I think XPConnect still exposes this interface as SVGMpathElement

It's SVGMPathElement in the spec, so I think it should be exposed as such.

Comment 31

[Phil Ringnalda (:philor)]

Backed out in http://hg.mozilla.org/integration/mozilla-inbound/rev/9a315f39e03a for browser-chrome failures like https://tbpl.mozilla.org/php/getParsedLog.php?id=18479622&tree=Mozilla-Inbound.

Comment 32

[Boris Zbarsky [:bzbarsky]]

> so I think the easiest solution may be to convert HTMLScriptElement and land it at the
> same time as SVGScriptElement.

Makes sense to me.

> It's SVGMPathElement in the spec, so I think it should be exposed as such.

Yes, that's what this patch does.  So we have Components.interfaces.nsIDOMSVGMpathElement but window.SVGMPathElement.

The failing test is testing the code that does instanceof SVGTitleElement.  I'm fairly surprised that didn't work right, since the landed patch did have the hasInstanceInterface bit in it...

Comment 33

[David Zbarsky (:dzbarsky)]

SVGTitleElement didn't have a QI entry for nsIDOMSVGTitleElement =(

remote:   https://hg.mozilla.org/integration/mozilla-inbound/rev/1c9ec9784927
remote:   https://hg.mozilla.org/integration/mozilla-inbound/rev/786f26d1302e
remote:   https://hg.mozilla.org/integration/mozilla-inbound/rev/d42ed304bda1
remote:   https://hg.mozilla.org/integration/mozilla-inbound/rev/5b57d2f7af3d
remote:   https://hg.mozilla.org/integration/mozilla-inbound/rev/46e52ba374e0
remote:   https://hg.mozilla.org/integration/mozilla-inbound/rev/c47eb8f76e9d
remote:   https://hg.mozilla.org/integration/mozilla-inbound/rev/a0414feb9656

Comment 34

[David Zbarsky (:dzbarsky)]

Pushed SVGScriptElement
https://hg.mozilla.org/integration/mozilla-inbound/rev/9d028fec8318

Comment 35

[Phil Ringnalda (:philor)]

https://hg.mozilla.org/mozilla-central/rev/1c9ec9784927
https://hg.mozilla.org/mozilla-central/rev/786f26d1302e
https://hg.mozilla.org/mozilla-central/rev/d42ed304bda1
https://hg.mozilla.org/mozilla-central/rev/5b57d2f7af3d
https://hg.mozilla.org/mozilla-central/rev/46e52ba374e0
https://hg.mozilla.org/mozilla-central/rev/c47eb8f76e9d
https://hg.mozilla.org/mozilla-central/rev/a0414feb9656
https://hg.mozilla.org/mozilla-central/rev/9d028fec8318