[Chromium API] Disambiguate allowJavaScript from didFailToExecuteJavaScript

• Component: New Bugs
• Status: RESOLVED FIXED

[Adam Barth]

[Chromium API] Disambiguous allowJavaScript from didFailToExecuteJavaScript

[Adam Barth]

Created attachment 49141 [details]
Patch

[Peter Kasting]

You might want to comment more (perhaps in comments on the function declarations?) on why a piece of code would care whether it can execute script if it doesn't actually intend to execute script.  There are other possible APIs here whose usefulness probably depends on the answer.

(I'm mostly worried about future code calling allowJavaScript() without calling didFailToExecuteJavaScript() when it should.  Perhaps we should create a single access point you call when you want to run script that checks if it's legal and calls this callback if not, or something.)

[Adam Barth]

Comment on attachment 49141 [details]
Patch

This patch is wrong because V8 cheatzors.

[Adam Barth]

Created attachment 49146 [details]
Patch

[Adam Barth]

Done and done.  This patch got a bit bigger and probably harder to merge, but I think you're right that we should force clients to be explicit about this bit so we do the right thing in the future.

[Peter Kasting]

Comment on attachment 49146 [details]
Patch

> +enum ReasonForCallingCanExecuteScripts {
> +    AboutToExecuteScript,
> +    NotAboutToExecuteScript
> +};

I wonder if there could be a better name for the second choice... "NotPlanningToExecuteScript"?  "JustCuriousAboutCurrentSetting"?  I dunno :(.  In any case it seems like rationale for this enum should be given in comments above it.

[Darin Fisher (:fishd, Google)]

Comment on attachment 49146 [details]
Patch

> +        [Chromium API] Disambiguate allowJavaScript from didFailToExecuteJavaScript

As discussed elsewhere, I'd still much prefer a name for didFailToExecuteJavaScript
that better indicates failure due to a policy decision as opposed to other problems
such as a syntax error.

How about renaming this method to didNotAllowScript?  That nicely corresponds with
allowScript.

[Peter Kasting]

Created attachment 49435 [details]
patch v3

Renamed didFailToExecuteJavaScript() to didNotAllowScript().  Updated patch to apply against current tree.

[Peter Kasting]

Created attachment 49440 [details]
patch v4

Last patch didn't fix everywhere because I didn't realize the first patch used two different function names (for no good reason)

[Peter Kasting]

Landed in r55207.