It was Safari Mac. I just realized that my Safari dock icon is still pointed at a WebKit nightly to work around some dev mode crashes I was seeing in the stock 4.0.4 with earlier dev mode plugins. The specific version I'm using is: Version 4.0.4 (5531.21.10, r51062) on Leopard.
I can't repro it with the OSX default 4.0.4. I'll test against the latest nightly WebKit later today and see if it's still busted there. Sorry about that - I'll keep an eye on this issue and bring it up again if it starts happening in a shipping WebKit. Matt. On 4-Dec-09, at 12:30 PM, Scott Blum wrote: > Is this Safari on Mac, or Chrome on Windows, or what? > > On Fri, Dec 4, 2009 at 11:31 AM, Matt Mastracci > <[email protected]> wrote: > I'm seeing a strange dev mode error in WebKit. It started happening > recently and I can't put my finger on a specific change that would > cause it. > > I'm returning an non-null JS native array from a JSNI method, cast to > JsArrayInteger. When I call JsArrayInteger::shift() on it, I get: > > Result of expression 'this.shift' [undefined] is not a function. > > If I defer this operation with the scheduler after catching the > exception (or step through with the debugger), it works as expected. > It appears to be a race of some sort but I'm not really sure what it > could be. Everything else is working as it should. > > Safari breaks on an exception in that method, but the console shows a > valid value for 'this' and 'this.shift'. > > The same code runs successfully a number of times before it fails and > all of the arrays (successful or not) are part of the same large data > structure. > > Any ideas? > > Matt. > > -- > http://groups.google.com/group/Google-Web-Toolkit-Contributors > > > -- > http://groups.google.com/group/Google-Web-Toolkit-Contributors -- http://groups.google.com/group/Google-Web-Toolkit-Contributors
