1335825 - [e10s] provide full accessible state via xpc in windows

Status: RESOLVED FIXED

(Core :: Disability Access APIs, defect)

Comment 1

[Eitan Isaacson [:eeejay]]

Attachment: Provide full state bitfield to xpcom accessibles in Windows r?yzen

Comment 2

[Yura Zenevich [:yzen] (please use "needinfo")]

Comment on attachment 8832535 [details] [diff] [review]
Provide full state bitfield to xpcom accessibles in Windows r?yzen

Review of attachment 8832535 [details] [diff] [review]:
-----------------------------------------------------------------

looks good, was there a reason to not use the same style as in the original method - setting state variable to 0 in the beginning and returning state (not 0) in 2 places.

Comment 3

[Eitan Isaacson [:eeejay]]

I just copied that pattern from another method. No good reason. (but in hindsight, its probably better not to return 'state' after passing it's address to a failing function. No promise that it didn't assign it.

Comment 4

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/mozilla-inbound/rev/f106e5720eb2
Provide full state bitfield to xpcom accessibles in Windows. r=yzen

Comment 5

[Trevor Saunders (:tbsaunde)]

(In reply to Eitan Isaacson [:eeejay] from comment #3)
> I just copied that pattern from another method. No good reason. (but in
> hindsight, its probably better not to return 'state' after passing it's
> address to a failing function. No promise that it didn't assign it.

technically I believe COM rules say you should only modify it if you return success.  That said personally I think its clearer to explicitly return 0 where you can so I'd do that anyway.

Comment 6

[Phil Ringnalda (:philor)]

https://hg.mozilla.org/mozilla-central/rev/f106e5720eb2