On Tue, 11 Jan 2022 at 08:55, Stefan Brüns stefan.bruens@rwth-aachen.de wrote:
On Freitag, 7. Januar 2022 18:10:05 CET you wrote:
On Fri, 7 Jan 2022 at 09:00, Stefan Brüns stefan.bruens@rwth-aachen.de
wrote:
In case real and expected segment count differ, don't compare the segment type/position for extra segments.
Unless it's e.g. a huge help when debugging failures, I'm not sure we care; if the segment counts don't match the test is going to fail, one way or another.
This is an invalid memory access (though read-only). This should never be done, even in tests.
That's not the position the Wine project has historically taken on this kind of thing, at least when it comes to the tests. The basic idea is that the test isn't supposed to fail, so when it does, we don't particularly care about any follow-up damage. (Provided that's limited to the test process.) I.e., think of ok() much like assert().