[webkit-reviews] review granted: [Bug 189455] Streamline JSRetainPtr, fix leaks of JSString and JSGlobalContext : [Attachment 349279] Patch

[bugzilla-daemon at webkit.org]

Keith Miller <keith_miller at apple.com> has granted Darin Adler
<darin at apple.com>'s request for review:
Bug 189455: Streamline JSRetainPtr, fix leaks of JSString and JSGlobalContext
https://bugs.webkit.org/show_bug.cgi?id=189455

Attachment 349279: Patch

https://bugs.webkit.org/attachment.cgi?id=349279&action=review




--- Comment #11 from Keith Miller <keith_miller at apple.com> ---
Comment on attachment 349279
  --> https://bugs.webkit.org/attachment.cgi?id=349279
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=349279&action=review

r=me with comments. I think you need to rebase though.

> Source/JavaScriptCore/API/JSObjectRef.cpp:796
> +	  
propertyNames->array.uncheckedAppend(adopt(OpaqueJSString::create(array[i].stri
ng()).leakRef()));

What's the point of adopting the string then leaking it?

> Source/JavaScriptCore/API/JSRetainPtr.h:69
> +    // FIXME: Make this private once Apple's internal code is updated to not
rely on it.

nit: Could you file a bugzilla for this and link to it in the FIXME?

> Source/JavaScriptCore/API/JSValue.mm:159
> +    JSStringRef string = JSStringCreateWithCFString((__bridge
CFStringRef)description);
> +    auto value = [JSValue valueWithJSValueRef:JSValueMakeSymbol([context
JSGlobalContextRef], string) inContext:context];
> +    JSStringRelease(string);

Aww man, I'm so used to RAII lifetime management that I completely missed this!
:( Good catch.

Maybe we should use RetainPtr in this file so we don't make these mistakes
again?