On Wed, Jun 29, 2011 at 03:05:09PM +0100, Julio Merino wrote: >>> Perhaps if atf were less intrusive...? >> >> What do you mean? I think it needs to be quite intrusive >> (sandboxing, etc.). Unquestionably the old /regress-style is not >> the way to go. Even if you dislike some parts of the API, already >> the consistency is a big win. > > Yep, I had the same question here. I don't know what the original > comment meant.
In just about every other test suite I've used (which includes some very large ones with turing complete/scripted test harness programs and other fancy stuff) there are test programs and test driver scripts, but all the pieces have clearly defined and clearly separated roles, and moreover are self-contained so they can be worked with easily. ATF, however, insists on sticking its fingers and APIs in everywhere ("intrusive") with the net result that everything becomes vastly more complicated and working with individual tests directly becomes awkward and difficult. One of the predictable consequences is not being able to fold in preexisting test programs. But apparently I'm not entitled to an opinion since I haven't been working on the test suite. (I was at one point some time back, actually, but got fed up with all the unnecessary hoops to jump through and with how straightforward usability problems like PR 38746 got ignored.) -- David A. Holland dholl...@netbsd.org