Re: CVS psmith make: Various cleanups reported by people using the alpha release.

[Greg McGary]

"Paul D. Smith" <address@hidden> writes:

> Anyway, this is from memory from last night... what I think should
> happen in this case is that the order-only-ness of that prereq should be
> removed and it should be treated as just a normal prereq: after all,
> order-only prereqs are a strict subset of normal prereqs.

I implemented them as a disjoint sets, where normal and order-only
prereqs with the same name are considered unequal, so both are
retained.  The normal prereq goes into $^, and the order-only goes
into $|.

If you wish to consider order-only as a subset of normal prereqs, then
a normal prereq should subsume an order-only prereq, and if both are
present, uniquize_deps should toss an order-only prereq, so that
$^ gets it, and $| doesn't.

> Then the question becomes, should $| be empty?  I don't know for sure
> but I think that would be OK (and it's better than $? being empty, IMO).

I have no opinion on which behavior is desirable, since I haven't yet
thought of any use for the value of $|.

Greg