Re: [PATCH 1/2] remote: Use `catch' in killing pending force-kills

[Maciej W. Rozycki]

On Thu, 21 May 2020, Jacob Bachmeyer wrote:

> >  This is a rare corner case though and I would therefore recommend my 
> > minimal fix to be applied first, and then you can apply your code 
> > modernisation to all `exec' invocations throughout `close_wait_program'.  
> > This way changes will be functionally consistent and won't mix fixes with 
> > clean-ups (or indeed address independent issues at a time).
> 
> Even if you are using the previous exec(n) command exactly as is, the 
> catch(n) script should still be braced, since you are adding the 
> catch(n) command.  Including the exec(n) cleanup also has a benefit of 
> ensuring that the cleanup does not reintroduce the very bug you are 
> fixing or a similar bug.  When testing is as involved as verifying 
> DejaGnu currently is, grouping related changes is beneficial because it 
> allows one verification run to validate related changes.

 Fair enough.  I've posted v2 now with your suggested improvement.

  Maciej