Chris McMahon wrote:
> I had another thought, namely that calling goto() inside an explicit 
> timeout is nicer than actually building a timeout value into Watir.  
> There is a timeout value set in the Ruby http-access2 library and I am 
> forever having to remember where the dang value is set and rejigger it 
> on all of my test machines.  Not to mention the heartbreak of when a 
> long-running process terminates because I didn't set the timeout high 
> enough, destroying some critical test or another. 
I used to think we needed to add the timeout mechanism requested by the 
original poster, but lately i've been more inclined to follow Chris's 
thinking -- use Ruby for things that Ruby is good for and it already has 
a general timeout mechanism.

My main reason for thinking it should be added to Watir is that it 
really sucks when you run a large testsuite and then find out that one 
of the early tests hung for a stupid or hard-to-reproduce reason and 
then you have to re run it. The test harness itself really should know 
to abort a hung test and move on to the next one. The problem is make it 
smart enough to "know" when it is hung.

I would appreciate more comments on this topic. Clearly the timeout 
functionality already exists. The question is how and whether to build 
this into Watir or perhaps Watir::TestCase in order to help testers make 
competent use of the tool.

I'd also like to whether there are any drawbacks to using plain 
"timeout". I spent some time working on "until_with_timeout" as a Watir 
method and never quite finished that because of these concerns.

Bret
_______________________________________________
Wtr-general mailing list
[email protected]
http://rubyforge.org/mailman/listinfo/wtr-general

Reply via email to