On Wed, 12 Jul 2000, Paul D. Smith wrote:
> Not in the makefile you sent in your message. There's a target %.gif
> that can be built from %.ppm, but there's no implicit rule that shows
> how to build anything (including %.ppm) from %.gif.
>
> Either you overlooked it, or the makefile I received got truncated
> somehow.
Ooops!
> ch> Anyway, yes, perhaps I'm stretching 'make' too far, but I'm strongly
> ch> encouraged by the fact that this works some of the time!
>
> >> As far as I can tell, make works all the time when it should work.
> >> What is broken is that in some situations where make can't do what
> >> you want, it will try to do something bogus instead of just
> >> printing an error.
>
> >> Am I correct here, or am I missing something?
>
> ch> It's actually doing the right thing about half the time. I guess the
> ch> code determining which dependency to drop is picking an arbitrary
> ch> dependency, instead of useless dependencies.
>
> It's not clear how make would determine "useless". What I would assume
> is that make would pick the last dependency that caused the loop, and
> throw it out.
Yes, that's exactly it. If adding a dependency to the graph would
form a loop, don't add the dependency. The result will be a tree.
It's like flood-filling a region in computer graphics; if the program
didn't turn a very loopy graph into a tree, it would be stuck in an
infinite loop, painting the same pixels over and over again.
--
Charles Howes -- [EMAIL PROTECTED]
