Follow-up Comment #4, bug #66030 (group make): I'm not trying to be argumentative, really, but I still feel like we're not quite coming to closure on the same topic so let me try one more time.
It seems to me there's a significant difference between these two rules:
foo.h foo.c: ; <recipe>
foo.%: ; recipe
In case #1 make is evaluating targets serially and comes to foo.h first.
Perfect agreement there. But in case #2 make knows, because it's been told,
that the recipe will update both targets. It's not a "in most cases these are
the same thing" *prediction*, it's a "because I said so" *promise*.
Thus it seems to me that even though make encounters foo.h first, in case #2
it has more information - valuable information - which is worth sharing. You
say 'The fact that "foo.c" is also updated by this recipe is not relevant to
make' but it's *known* to make and quite relevant to the user. So why not
share it?
If I tell my spouse I'm going out to buy a gallon of milk and in fact I buy a
gallon of milk and a vial of crack cocaine, I may not have lied exactly but
there's been rather a glaring omission, no?
Anyway, there's no doubt this is just a nice-to-have enhancement and we agree
on that so the details aren't critical. I just think they're a little more
than "technical semantics", that's all.
BTW I'd expect the same reasoning to apply to grouped targets but have not
been able to work with 4.3 yet.
_______________________________________________________
Reply to this item at:
<https://savannah.gnu.org/bugs/?66030>
_______________________________________________
Message sent via Savannah
https://savannah.gnu.org/
signature.asc
Description: PGP signature
