See https://dvcs.w3.org/hg/speech-api/raw-file/tip/speechapi.html#speechreco-error

This patch creates an enum for the error codes for SpeechRecognitionError. It also removes nsIDOMSpeechRecognitionError, since that turned out to be possible (but let me know if you'd actually like to keep it), and updates the tests accordingly.

Comment on attachment 766141 [details] [diff] [review] Bug 884407 - Use an enum in SpeechRecognitionError >+DOMCI_DATA(SpeechRecognitionError, mozilla::dom::SpeechRecognitionError) No need for DOMCI_DATA >+namespace mozilla { >+namespace dom { >+ >+SpeechRecognitionError::SpeechRecognitionError(mozilla::dom::EventTarget* aOwner, nsPresContext* aPresContext, nsEvent* aEvent) >+: nsDOMEvent(aOwner, aPresContext, aEvent), >+ mError(), >+ mMessage() There shouldn't be need for mMessage(). (codegen may create such). >+{ >+ SetIsDOMBinding(); No need for this anymore. These days all the events use DOM bindings, so SetIsBOMBinding() is called in the nsDOMEvent ctor. >+ nsresult >+ InitSpeechRecognitionError(const nsAString& aType, >+ bool aCanBubble, >+ bool aCancelable, >+ SpeechRecognitionErrorCode aError, >+ const nsAString& aMessage); So if we don't have .idl, why do we need this version? Wouldn't the one with ErrorResult be enough?

Refreshing the patch after addressing review comments. Keeping r+ from smaug.

It seems this breaks compilation with --disable-webspeech ... Can you please fix it?

Sorry about that. This patch should fix it.

Comment on attachment 768102 [details] [diff] [review] Only build SpeechRecognitionError when MOZ_WEBSPEECH is enabled I guess that works (I'm not familiar with the syntax moz.build uses)

Seems to work, thanks.