Olli Pettay [:smaug][bugs@pettay.fi] Assignee
	Description • 9 years ago

This is a bit ugly, but I tried to follow the existing style where it is quite
strongly checked that the thing is a link.

	Trevor Saunders (:tbsaunde)
	Comment 1 • 9 years ago

Comment on attachment 8580298 [details] [diff] [review]
v1

actually this doesn't work, AtkObject and AtkHyperlink are different objects.  So what you need to do is first make MaiHyperlink sotre either an accessible* or a Proxy* and make nsMaiInterfaceHyperlinkImpl.cpp deal with proxies.  Then you can make this stuff work by grabing the MaiHyperlink like the current code does but use the pointer has to get at a proxy.

	Olli Pettay [:smaug][bugs@pettay.fi] Assignee
	Comment 2 • 9 years ago

Ok, I don't know how to make that work.

	Trevor Saunders (:tbsaunde)
	Comment 4 • 9 years ago

Comment on attachment 8581841 [details] [diff] [review]
no atk

>+DocAccessibleChild::RecvStartOffset(const uint64_t& aID,
>+                                    uint32_t* aRetVal,
>+                                    bool* aOk)
>+{

I think the extra checking is unneccessary, the link interface should only be exposed if IsLink() returns true, and It can kind of be wrong in that it can say false when the accessible went away.  However istm handling the cases of method called on dead accessible and called on wrong type differently doesn't make much sense.  So this is certainly fine for now and we'll so how it turns out.

	Olli Pettay [:smaug][bugs@pettay.fi] Assignee
	Comment 5 • 9 years ago

https://hg.mozilla.org/integration/mozilla-inbound/rev/a90178826816

	Ryan VanderMeulen [:RyanVM]
	Comment 6 • 9 years ago

https://hg.mozilla.org/mozilla-central/rev/a90178826816