It looks like params like timeout don't make their way down into the
redoer.. so even though timeout might be set on the app(if you do
app.timeout = 1;) it doesn't get set in the redoer so there are two
time based loops that pay no attention to the timeout. I found this
using the debugger. I short circuited the loops to break after the
first iteration and I was able to fail a check in less than .5s.
Of course the way the loops are coded setting a low timeout could
cause it to never go into the loop(I think changing to a do while
instead of a while do would solve that race condition).
On Apr 21, 9:27 am, Pete Hodgson <
phodg...@thoughtworks.com> wrote:
> We saw a similar issue ourselves. IIRC it turns out that that timeout is
> actually applied each time you traverse down the view hierarchy, or
> something like that. So if you have a 10 layer view heirarchy it will take
> 10 seconds to *not* find something. I may be a little off on the details,
> but I'm pretty sure that was the basic problem.
>
> I started working on a patch to resolve this, but never got 'round to
> finishing it off nicely. If you manage to find a solution I'd love to hear
> about it :)
>
> Cheers,
> Pete
>