On 3/1/21 3:07 PM, Andres Freund wrote: > Hi, > > As part of trying to make the aio branch tests on cirrus CI pass with > some tap tests I noticed that "src/tools/msvc/vcregress.pl recoverycheck" > hangs. A long phase of remote debugging later I figured out that that's > not a fault of the aio branch - it also doesn't pass on master (fixed in > [1]). > > Which confused me - shouldn't the buildfarm have noticed? But it seems > that all msvc animals either don't have tap tests enabled or they > disable 'misc-checks' which in turn is what the buildfarm client uses to > trigger all the 'recovery' checks. > > It seems we're not just skipping recovery, but also e.g. tap tests in > contrib, all the tests in src/test/modules/... > > Andrew, what's the reason for that? Is it just that they hung at some > point? Were too slow?
I don't think speed is the issue. I probably disabled misc-tests on drongo and bowerbird (my two animals in question) because I got either instability or errors I was unable to diagnose. I'll go back and take another look to narrow this down. It's possible to disable individual tests. > > > On that last point: Running the tap tests on windows appears to be > *excruciatingly* slow. How does anybody develop on windows without a > mechanism to actually run tests in parallel? I think most people develop elsewhere and then adapt/test on Windows if necessary. > > I think it'd be good if vcregress.pl at least respected PROVE_FLAGS from > the environment - it can't currently be passed in for several of the > vcregress.pl tests, and it does seem to make things to be at least > somewhat less painful. +1 > > > This makes it even clearer to me that we really need a builtin > testrunner that runs tests efficiently *and* debuggable on windows. > "show me the code" :-) cheers andrew -- Andrew Dunstan EDB: https://www.enterprisedb.com
