W3C home > Mailing lists > Public > www-dom@w3.org > April to June 2013

Re: [Futures] Possible to multiply-resolve a resolver in weird ways

From: Boris Zbarsky <bzbarsky@MIT.EDU>
Date: Thu, 06 Jun 2013 11:06:32 -0400
Message-ID: <51B0A578.9080407@mit.edu>
To: Anne van Kesteren <annevk@annevk.nl>
CC: Sean Hogan <shogun70@westnet.com.au>, DOM mailing list <www-dom@w3.org>
On 6/6/13 5:31 AM, Anne van Kesteren wrote:
> On Thu, Jun 6, 2013 at 2:25 AM, Sean Hogan <shogun70@westnet.com.au> wrote:
>> Without .done(), what is the recommended way of ensuring unhandled errors
>> get to window.onerror?
> The plan is for debugging tools to solve this. It seems better to have
> debugging tools solve this problem than require every developer to use
> the correct method.

I don't see how debugging tools help at all.

It's reasonably common (in that we've gotten a number of bug reports 
from sites doing this and discovering issues that way) for sites to 
install an onerror handler that phones home the error and then analyze 
the resulting reports to catch bugs in their code.  Or browser code, as 
the case may be.

 From the point of view of those sites, any exception that they do not 
explicitly catch and also doesn't end up calling window.onerror is a 
problem because it might be hiding bugs.

This use case is obviously not covered by debugging tools...

Received on Thursday, 6 June 2013 15:07:06 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 20 October 2015 10:46:21 UTC