940323 - Move the strictMode flag from ContextOptions to RuntimeOptions

Status: RESOLVED FIXED

(Core :: JavaScript Engine, defect)

Comment 1

[Bobby Holley (:bholley)]

This probably isn't going to be looked at until bug 939562 lands. Marking the dep so I'll see it fly by when that bug lands.

Comment 2

[Bobby Holley (:bholley)]

Attachment: Mode strictMode to RuntimeOptions. v1

Comment 3

[Jan de Mooij [:jandem]]

Comment on attachment 8455487 [details] [diff] [review]
Mode strictMode to RuntimeOptions. v1

Review of attachment 8455487 [details] [diff] [review]:
-----------------------------------------------------------------

::: js/xpconnect/src/XPCComponents.cpp
@@ +3243,5 @@
>      }
>  
>  GENERATE_JSCONTEXTOPTION_GETTER_SETTER(Strict, extraWarnings, setExtraWarnings)
>  GENERATE_JSRUNTIMEOPTION_GETTER_SETTER(Werror, werror, setWerror)
> +GENERATE_JSRUNTIMEOPTION_GETTER_SETTER(Strict_mode, strictMode, setStrictMode)

Could this break code/websites, due to some addon toggling "use strict" for the entire runtime? I did some grepping and I don't think we set Cu.strict_mode ourselves. If you think this is okay to change I'll trust your judgement.

Comment 4

[Bobby Holley (:bholley)]

(In reply to Jan de Mooij [:jandem] from comment #3)
> Comment on attachment 8455487 [details] [diff] [review]
> Mode strictMode to RuntimeOptions. v1
> 
> Review of attachment 8455487 [details] [diff] [review]:
> -----------------------------------------------------------------
> 
> ::: js/xpconnect/src/XPCComponents.cpp
> @@ +3243,5 @@
> >      }
> >  
> >  GENERATE_JSCONTEXTOPTION_GETTER_SETTER(Strict, extraWarnings, setExtraWarnings)
> >  GENERATE_JSRUNTIMEOPTION_GETTER_SETTER(Werror, werror, setWerror)
> > +GENERATE_JSRUNTIMEOPTION_GETTER_SETTER(Strict_mode, strictMode, setStrictMode)
> 
> Could this break code/websites, due to some addon toggling "use strict" for
> the entire runtime? I did some grepping and I don't think we set
> Cu.strict_mode ourselves. If you think this is okay to change I'll trust
> your judgement.

Yeah. In general, the old model of having them on the JSContext wasn't really all that fool-proof. That definitely presents a risk, but we don't really have an alternative way to preserve the same semantics. If we encounter breakage on beta, we can just make this flag a no-op (though Jesse might not like that).

Comment 5

[Bobby Holley (:bholley)]

https://hg.mozilla.org/integration/mozilla-inbound/rev/d6289d614a70

Comment 6

[Wes Kocher (:KWierso) (Not reading bugmail; email directly if needed)]

https://hg.mozilla.org/mozilla-central/rev/d6289d614a70