do our tinderboxen all run at 1280x1024?

Andrew Sutherland asutherland at
Mon Apr 26 21:38:27 UTC 2010

  On 04/26/2010 02:28 PM, Ben Bucksch wrote:
> Isn't it a good thing that they run at different resolutions? If 
> something breaks with a resolution != 1280x1024, we'd want to know 
> that, no?

The test in question tests that we collapse and uncollapse the text on 
the labels on the quick filter bar out of existence when appropriate.  
There is a minimum screen width of ~1200 pixels to be able to expand the 
labels when the folder pane is generally expanded.  The window APIs  
won't let us size the content area larger than the size of the screen as 
far as limited probing reveals.  If the screen isn't 1280x1024, the test 
fails because it can't size the window big enough.

I attempted to work-around the problem by collapsing the folder pane, 
but we still need to have some minimum screen size and we generally need 
to know what that value is in order to be able to write tests.

Any tests that need to test at different resolutions should size the 
window appropriately in order to test themselves and not rely on more 
random issues.

Another potential problem with weird sizes is that you can't generate 
click events on things that aren't visible because they are scrolled 
outside of the visible area in the window.  It makes for weird test 


More information about the tb-planning mailing list