Hi,
I just spent a wad of time figuring out why every single .Net regression
test was failing. Turns out that I had created a "t.pbc" in the same
directory as the Parrot executable. Each regression test script produces a
file of this name in the languages/dotnet folder as the output of the
translator, so that it can run tests on it (each test load_bytecode's it).
And the tests all failed because Parrot found the "t.pbc" in the same
directory as the Parrot executable, not the CWD.
Question is, what's the right behaviour here? I guess at the end of the day
we'll want to be able to modify the search order so as to provide for @*INC,
but in the meantime it makes more sense to me if we look in the CWD first.
Thoughts?
Jonathan
