----- Original Message ---- > From: Matisse Enzer <[EMAIL PROTECTED]>
> > Should be fairly easy to implement with the new Test::Harness. > What > > syntax would be desired? Also, it seems to me that you've two > > different cases. Have the entire suite BAIL_ON_FAIL or have an > > individual test program halt on failure. The latter seems much > moreuseful to me. > > I had not thought of limiting BAIL_ON_FAIL to a single test program, > can you say more about this? It's a common idiom in the xUnit world. The assumption is that when a test fails, you don't want to run more as subsequent success or failure can't be trusted (their notion of a test is somewhat different, though). For example, here's something which bites people all the time: use_ok $module; If that fails, the test program keeps running. It's also possible to have much of that test program pass. If you don't notice that use_ok failed, it's easy to misdiagnose a problem. Also, how many times have people dropped an 'exit' in their tests just to get things to stop at a failure? I do it all the time. Often just fixing that first failure makes all the others go away. > Personally, I would set the entire test run to BAIL_ON_FAIL - I want > the whole test run to stop as soon as a single unexpected (non-TODO) > failure occurs. That sucks when you have a test suite which takes an hour to run :) (I've worked on more than one). Cheers, Ovid -- Buy the book - http://www.oreilly.com/catalog/perlhks/ Perl and CGI - http://users.easystreet.com/ovid/cgi_course/ Personal blog - http://publius-ovidius.livejournal.com/ Tech blog - http://use.perl.org/~Ovid/journal/
