Bram Moolenaar <Bram@...> schrieb am 20.11.02 13:49:12:
> Joerg Beyer wrote:
> > My RFC collects these compile time switches in a seperate list.
> > I assume it is enough to collect "what sould be written to the
> > define.h file", so the feature tests could collect the items and a
> > later step could dump the file.
> Yes. I think autoconf does the same thing, since it includes
> "confdefs.h" in test programs. This is needed for a following test to
> use the results of a previous test (e.g., including a header file that's
> not always available).
yes, I will test-implement it soon. An post it to this list.
> Hmm, perhaps we should do this the same way, and by the time all
> configure checks are done you automatically have the complete
> "confdefs.h" file. Can't think of a reason why another file would be
> needed. It works for the "test_compile" tests, thus why wouldn't it
> work for the actual program?
I thought about the genric algorithm to generate a file form it's
".in" file (the way the make family does it), when I said "one file among others"