Re: [Boost-testing] Problems using --monitored on Windows

From: Reece Dunn (msclrhd_at_[hidden])
Date: 2006-03-06 18:55:43

Rene Rivera wrote:
>Reece Dunn wrote:
> > This leads me to the following question: isn't there a better way of
> > performing a monitored build and detecting non-responsive applications?
>
>Yes, don't use "--monitored", instead add the bjam option "-l300" with
>"--bjam-options=-l300" to regression.py. The bjam time limit doesn't
>have the drawbacks that the monitor exe has, it does have others though :-)

Cool, thanks!

Out of interest... what are the limitations of -l600?

>You might want to use more than 5 minutes, for me 10 minutes has worked
>well experimentally. At 5 minutes it would sometimes kill long running,
>but not indefinite, compiles.

Sure - I have set it to 10 minutes and am doing a run-through of the cw-9.4 
tests.

- Reece