796903 - Move DOMImplementation to Paris bindings

Status: RESOLVED FIXED

(Core :: DOM: Core & HTML, defect)

Comment 1

[:Ms2ger (he/him; ⌚ UTC+1/+2)]

Attachment: Patch v1

Comment 2

[Boris Zbarsky [:bzbarsky]]

Shouldn't there be some nsDocument bits here too?

Comment 3

[:Ms2ger (he/him; ⌚ UTC+1/+2)]

Attachment: Patch v2

Let's try this again.

Comment 4

[Boris Zbarsky [:bzbarsky]]

Comment on attachment 668144 [details] [diff] [review]
Patch v2

>+++ b/content/base/src/DOMImplementation.cpp
>+  NS_INTERFACE_MAP_ENTRY(nsIDOMDOMImplementation)
>+  NS_INTERFACE_MAP_ENTRY_AMBIGUOUS(nsISupports, nsIDOMDOMImplementation)

This was there in the code you copied from, but it's clearly silly.  Make the nsISupports bit not use AMBIGUOUS?

>+++ b/dom/bindings/Bindings.conf
>+'DOMImplementation': {
>+},
>+

Now that bug 792980 is fixed, you don't need that.

Given that this is not prefable, why did you leave nsIDOMDOMImplementation and associated gunk?  Seems like it can go....

r=me with those addressed.  Though if you do remove nsIDOMDOMImplementation, I'd like a look at the result.

Comment 5

[:Ms2ger (he/him; ⌚ UTC+1/+2)]

Attachment: Part b: Pass an nsIContent to InternalIsSupported

This interface is pretty silly. I like this better. (Turns out I don't strictly need it, as DOMImplementation still inherit from nsISupports.)

Comment 6

[:Ms2ger (he/him; ⌚ UTC+1/+2)]

Attachment: Part c: Remove nsIDOMDOMImplementation

Comment 7

[Boris Zbarsky [:bzbarsky]]

Comment on attachment 668577 [details] [diff] [review]
Part b: Pass an nsIContent to InternalIsSupported

r=me

Comment 8

[Boris Zbarsky [:bzbarsky]]

Comment on attachment 668578 [details] [diff] [review]
Part c: Remove nsIDOMDOMImplementation

r=me

Comment 9

[:Ms2ger (he/him; ⌚ UTC+1/+2)]

Attachment: Part d: Followup fixes

This passes try, at least. (Should've thought about that earlier.)

Comment 10

[Olli Pettay [:smaug][[e-mail]]]

Do we really want to remove nsIDOMDOMImplementation? I would expect some
binary addons to use it to create documents.
(And in fact it is probably the safest way to create "data" documents to be used in web pages.)

Comment 11

[Boris Zbarsky [:bzbarsky]]

> I would expect some binary addons to use it to create documents.

How would they get their hands on an nsIDOMDOMImplementation?

Comment 12

[Olli Pettay [:smaug][[e-mail]]]

From nsIDOMDocument

Comment 13

[Boris Zbarsky [:bzbarsky]]

And in particular, note that the ability to create a DOMImplementation by contract went away in bug 675166.

So are you thinking about cases in which the binary addon would call GetImplementation on an existing document?

Comment 14

[Boris Zbarsky [:bzbarsky]]

I guess if we think binary addons are using it, we could keep it...  It seems somewhat unfortunate, though.  :(

Comment 15

[Boris Zbarsky [:bzbarsky]]

Comment on attachment 671128 [details] [diff] [review]
Part d: Followup fixes

r=me on this part, but I talked to smaug and we decided to keep nsIDOMDOMImplementation for now for C++ consumers.  :(  Still don't need classinfo, of course.

Comment 16

[:Ms2ger (he/him; ⌚ UTC+1/+2)]

https://hg.mozilla.org/mozilla-central/rev/b72413cdab5e
https://hg.mozilla.org/mozilla-central/rev/abe709bfc49a