Re: AC_CONFIG_HEADERS vs. AC_CONFIG_FILES

[David Fang]

> The problem with doing these things in configure is that one must rerun
> configure to regenerate the file.
> Sometimes it is better do produce these things in config.status.

Doesn't autoconf/automake generate makefile dependencies for the output
files it configures?  When I have maintainer-mode turned off, the make
rules automatically rerun config.status when an input file has changed.

e.g. cd .. && config.status src/foo.h

In response to Ralf's suggestion:
AC_CONFIG_FILES([foo.h.tmp:foo.h.in],
        dnl ... move-if-change foo.h.tmp foo.h
)

This will work except that there will be no foo.h: foo.h.tmp dependency
automatically generated to auto-reconf, which would be nice.  It's not
difficult to add the rule in the Makefile.am manually, but it's just one
more thing to remember (and explain to someone reading through my files).
I'll play around some more...


David Fang