[webkit-dev] SVG pixel test failures (recent regressions?)

[David Hyatt]

On Jan 26, 2009, at 12:02 PM, Darin Adler wrote:

> On Jan 26, 2009, at 9:58 AM, David Hyatt wrote:
>
>> Yes.  I'd have done it myself, but I have Perian installed.  run- 
>> webkit-tests is refusing to allow me to generate new pixel test  
>> results (despite the fact that Perian affects only tests that use  
>> <video> or Quicktime).  Hopefully that pointless restriction on run- 
>> webkit-tests can be removed so that I can generate pixel test  
>> results again.
>
> That's my fault. I asked Pierre to put that in because people were  
> checking in incorrect pixel test results and we were getting a lot  
> of spurious failures because of it.
>
> We can easily take the check out, but then we’re going to have those  
> failures again.
>
> Do you have any ideas about how we should deal with this? Maybe I  
> can teach run-webkit-tests which tests depend on <video> and  
> QuickTime?

I think a strongly-worded warning is sufficient.  A warning lets  
someone using run-webkit-tests know that they should be careful if  
they are generating results for <video> tests.

dave
(hyatt at apple.com)