Fixing a regression test ...

[Duane Clark]

Paul Millar wrote:
> On Fri, 12 Jul 2002, Duane Clark wrote:
>>But the test should have tested for 
>>the existence of drive D: before using it.
> 
> Ok, so if I've understood this correctly, some tests check behaviour with
> whatever the current drive is (assumed to be C:), and some others check
> behaviour with some other drive (currently D: is assumed to be valid).
> 
> I'll submit another patch that should remove these assumptions. The patch
> works for me, but you might like to check I've not negated any of the
> tests: I'm not sure I follow exactly what's being tested ..

I should have tried a few more things.

First of all, the current directory is $TEMP, not where the test is 
being run from. The test changes it. But of course, $TEMP might not be 
on drive C:, so more needs to be done there.

In Windows, if a path "D:somedir\file.c" is supplied, then there are two 
possibilities. If D: is a drive for which Windows has a current 
directory (rather than being the current drive), then the returned path 
should be "D:\currentdir\somedir\file.c". If D: does not have a current 
directory, then the returned path should be "D:\somedir\file.c".

By the way, it does not matter whether the drive exists or not, so there 
is no need to test for the existence of it. I see that Wine returns the 
wrong value when a non-existent drive is supplied. I am not sure that is 
very important, though. It appears to return the correct value in all 
other cases.

So when running the test, it appears that Wine has a known "current 
directory" for the drive that the test is being run from, and for the 
drive that is part of $TEMP. To do the test properly would require that 
another configured drive be determined. A mapped CDROM drive would be a 
good choice, if it were easy to determine what drive that was.