Re: Timeout::Error

Daniel DeLorme wrote:
> What about the reverse problem? A lazy programmer might do a catch-all 
> rescue which would then catch unexpected inner exceptions which were not 
> meant to be rescued. i.e.
> 
> begin
>    Timeout.timeout(1) do
>      1 / 0 #oops!
>    end
> rescue
>    #silently ignore timeout
>    #...but ignores division by zero too!
> end
> 
> *That* might be the reason why Timeout::Error was not made a subclass of 
> StandardError.

This is no different from any other exception. If you're silently 
swallowing errors, you're asking for pain no matter what gets thrown. At 
some point, you need to treat programmers like adults and give them the 
tools they need without trying to hold their hands or strap them in.

If you take this thought to its logical conclusion, all exceptions will 
need to be caught individually and then /anything/ you didn't think of 
catching could filter up the stack and crash your app.

--J

--