1248681 - Add MOZ_WARN_UNUSED_RESULT to all of our OOM sensitive APIs

Status: RESOLVED FIXED

(Core :: JavaScript: GC, defect)

Description

[Terrence Cole [:terrence]]

I expect this will almost certainly surface some latent leak bugs.

Comment 1

[Terrence Cole [:terrence]]

Attachment: warn_unused_putWrapper-v0.diff

Added annotations to everything that looked likely in JSCompartment and JS::Zone. Found one missing check in JS_TransplantObject.

Comment 2

[Steve Fink [:sfink] [:s:]]

Comment on attachment 8719976 [details] [diff] [review]
warn_unused_putWrapper-v0.diff

Review of attachment 8719976 [details] [diff] [review]:
-----------------------------------------------------------------

Oh, wow. I guess somehow I had thought that we already had that warning on by default, and were casting to (void) or whatever the places where we were intentionally not checking. Out of curiousity, do you know how many warnings we get if we turn it on globally?

Comment 3

[Terrence Cole [:terrence]]

(In reply to Steve Fink [:sfink, :s:] from comment #2)
> Comment on attachment 8719976 [details] [diff] [review]
> warn_unused_putWrapper-v0.diff
> 
> Review of attachment 8719976 [details] [diff] [review]:
> -----------------------------------------------------------------
> 
> Oh, wow. I guess somehow I had thought that we already had that warning on
> by default, and were casting to (void) or whatever the places where we were
> intentionally not checking. Out of curiousity, do you know how many warnings
> we get if we turn it on globally?

I don't think there are any compilers that will do that, sadly. Maybe something for the analysis?

Comment 4

[Terrence Cole [:terrence]]

https://hg.mozilla.org/integration/mozilla-inbound/rev/ec2ea269b3de08dfc9827971d1c62b70a39bf496
Bug 1248681 - Warn if the result of putWrapper goes unused; r=sfink

Comment 5

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/ec2ea269b3de