Thread: New tests for new bugs (was Re: [BUGS] Re: backend dies on 7.1.1 loading large datamodel.)
New tests for new bugs (was Re: [BUGS] Re: backend dies on 7.1.1 loading large datamodel.)
From
Peter Eisentraut
Date:
Tom Lane writes: > Thomas Lockhart <lockhart@alumni.caltech.edu> writes: > > Will the patch include a case for the regression test? Or could someone > > (other than me??!!) volunteer to cover that? > > Seems like a good idea. As the embarrassee, I'm perhaps too close > to the problem to write a good addition to the regress tests; any > volunteers? The query that showed the bug would serve just fine. Actually, this practice should be much more widely deployed. For each bug, a test case should be added to guard against the bug coming back. At least when a suitable testing infrastructure exists. For instance, this would probably apply to each of the backend bug fixes that came in the last few days. Maybe it's too cumbersome to update the regression tests? Should the files be split into smaller pieces? -- Peter Eisentraut peter_e@gmx.net http://funkturm.homeip.net/~peter
Re: New tests for new bugs (was Re: [BUGS] Re: backend dies on 7.1.1 loading large datamodel.)
From
Tom Lane
Date:
Peter Eisentraut <peter_e@gmx.net> writes: > The query that showed the bug would serve just fine. Most of the bug reports we get are far too bulky to be appropriate to add to the regress tests as-is. IMHO anyway. We do need more extensive regress tests, but I don't think that slapping bug-report samples into them is the right way to get there ... regards, tom lane