On Tue, 19 Mar 2002, Michael Cardenas wrote:
The problem with only testing releases is that so many patches go into a release, that you won't know what patch broke your app.
Well, you only have about two weeks worth of patches to check. It's already better than the current situation where applications are not tested on a regular basis so that when a bug is reported there may be months of patches to check out.
I think the real problem is, what would your cron job do after the "wine {my app}" stage?
how about doing a
"wine {my app} > myapp.log && tail -f myapp.log | grep 'segfault occurred' && mail me@mydomain.com "my app crashed" ^D^D^D
The problem is that if the application does not crash it is likely to stay up indefinitely waiting for you to do something... Rather than automating the testing, it may be better to automate the process of updating Wine. This can be done completely unattended and can thus be done daily. Maybe a procedure documenting how to do this would be useful. Then as soon as Wine breaks something the user will know about it, just by using the application for normal stuff. Add to this a weekly check to see if any of the known bugs has been fixed, update the Application database entry if that's the case, and we have something which I think is close to ideal. Of course it would be wise to keep a known good Wine release in store in case Wine does break stuff.
-- Francois Gouget fgouget@free.fr http://fgouget.free.fr/ May your Tongue stick to the Roof of your Mouth with the Force of a Thousand Caramels.