On Sun, Jun 24, 2012 at 2:46 PM, Dawid Weiss <[email protected]> wrote: > > Anyway, his is one argument in a broader discussion. I don't know > which theoretical approach would be ideal but I know for sure even > typical randomization of parameters reveals bugs pretty damn > quickly... >
Some of it might be true: e.g. on my java6 there are 152 locales and 615 timezones. But in general these parameters are independent for our purposes: we just want to make sure we handle timezones correctly, and we handle locales correctly. If you don't do this, usually you find out very very fast because the code is totally broken in some way. I don't think we need to worry about the fact that we surely havent tested all 93,480 combinations, maybe there is a lurking bug out there, but what we do is way better than most projects, and at least we make a best effort to prevent bugs like http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=6240963 without actually slowing down test runs. -- lucidimagination.com --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
