861880 - Move TreeColumns to WebIDL

Status: RESOLVED FIXED

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

Description

[:Ms2ger (he/him; ⌚ UTC+1/+2)]

Attachment: Patch v1

Comment 1

[Boris Zbarsky [:bzbarsky]]

Comment on attachment 737497 [details] [diff] [review]
Patch v1

>+addExternalIface('MozTreeBoxObject', nativeType='nsITreeBoxObject')

I think this needs notflattened=True

>+interface TreeColumns {

[NoInterfaceObject], please.  I'm surprised this didn't fail the "list of interfaces we're exposing on the global" test.. or does that not catch webidl things?

> NS_INTERFACE_MAP_BEGIN(nsTreeColumns)

NS_INTERFACE_MAP_BEGIN_CYCLE_COLLECTION

r=me with those issues fixed.

Fwiw, if you put the WebIDL methods _after_ the XPCOM ones in the .cpp, I bet your diff would have been smaller...

Comment 2

[:Ms2ger (he/him; ⌚ UTC+1/+2)]

(In reply to Boris Zbarsky (:bz) from comment #1)
> >+interface TreeColumns {
> 
> [NoInterfaceObject], please.  I'm surprised this didn't fail the "list of
> interfaces we're exposing on the global" test.. or does that not catch
> webidl things?

Sure does. But the test whitelists TreeColumns.

Comment 3

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

IIRC because TreeColumns has been exposed to global scope for ages.

Comment 4

[Boris Zbarsky [:bzbarsky]]

Huh.  Why?  It's not nsIDOMTreeColumns....  Is it just because it had classinfo?

Comment 5

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

IIRC yes. It uses DOMClassInfo and that ends up putting it to global scope.
(don't recall how that happens..)

Comment 6

[Boris Zbarsky [:bzbarsky]]

Hmm.  I guess we use it on the RHS of instanceof too.  :(

Check this in as-is, I guess.  :(

Comment 7

[:Ms2ger (he/him; ⌚ UTC+1/+2)]

https://hg.mozilla.org/mozilla-central/rev/0d50cb959c46