619639 - ContentParent.cpp build warnings like "warning: ignoring return value of ‘bool mozilla::dom::PContentParent::SendFlushMemory(const nsString&)’, declared with attribute warn_unused_result"

Status: RESOLVED WORKSFORME

(Core :: IPC, defect)

Description

[Daniel Holbert [:dholbert]]

Filing bug on build warnings about ignored return values for functions declared with attribute warn_unused_result, in ContentParent.cpp:


> ContentParent.cpp: In member function ‘virtual nsresult mozilla::dom::ContentParent::Observe(nsISupports*, const char*, const PRUnichar*)’:
> ContentParent.cpp:368: warning: ignoring return value of ‘bool mozilla::dom::PContentParent::SendFlushMemory(const nsString&)’, declared with attribute warn_unused_result
> ContentParent.cpp: In member function ‘virtual nsresult mozilla::dom::ContentParent::HandleEvent(nsIDOMGeoPosition*)’:
> ContentParent.cpp:795: warning: ignoring return value of ‘bool mozilla::dom::PContentParent::SendGeolocationUpdate(nsIDOMGeoPosition* const&)’, declared with attribute warn_unused_result
> ContentParent.cpp: In member function ‘virtual nsresult mozilla::dom::ContentParent::OnAccelerationChange(nsIAcceleration*)’:
> ContentParent.cpp:843: warning: ignoring return value of ‘bool mozilla::dom::PContentParent::SendAccelerationChanged(const double&, const double&, const double&)’, declared with attribute warn_unused_result

Looks like the lines of code in question are from three different bugs/csets:
 Bug 618533 - http://hg.mozilla.org/mozilla-central/rev/888d64457836
 Bug 587414 - http://hg.mozilla.org/mozilla-central/rev/814f092ecaf9
 Bug 605341 - http://hg.mozilla.org/mozilla-central/rev/22121066081d

Presumably we should be doing one of the following:
 (a) checking the return value, if the return val always matters
 (b) drop the warn_unused_result, if return val doesn't matter
 (c) capture & drop the return value, if it sometimes matters but doesn't matter in this case, to silence the build-spam.

Comment 1

[Josh Matthews [:jdm]]

The typical thing to do here is use mozilla::unused << whatever if the return values for IPDL functions don't matter.

Comment 2

[Josh Matthews [:jdm]]

These have all been fixed, I'm pretty sure.