🚀 go-pugleaf

%% throopw@sheol.org (Wayne Throop) writes:

  wt> But it wouldn't be confusing, nor break any makefiles, to
  wt> introduce a syntax for expanding things at "dependency time", in
  wt> addition to the already existing alternatives of expanding things
  wt> at read-time, and expanding things at execution time (of an
  wt> action).

Well, it would be a new syntax and increase by half again the number of
ways that expansion can occur... judging by the number of requests for
help and "bug reports" which are not really bugs I answer every week
which basically boil down to not understanding the _current_ expansion
rules, I've grave concerns :).


Nevertheless, you might be interested in the new $(eval ...) function
available in 3.80 (unless something happens I expect to release
3.80 in the next week or so).

This function allows you to define whole new segments of makefiles
essentially anywhere and any time that variable evaluation is performed
(that includes within command rules when they're run, if you want to get
very bizarre).

The function expands its arguments and reads the results as makefile
commands.  It, itself, expands to the empty string.  Combined with the
ability to use define/endef to create variable values containing
newlines this is a powerful feature.

This capability should obsolete a very large percentage of the reasons
for using auto-re-exec.


Also available in 3.80 is the slightly bizarre SysV $$@ feature,
including $$(@D) and $$(@F)--but only for explicit and static pattern
rules (not pattern rules).  However this is provided mainly for
compatibility with SysV make, and is not intended to become a new
paradigm (IOW, the re-expansion of escaped items on the prerequisites
list is limited to only $$@ and friends, searched for explicitly by the
parser: it's not a general thing).

-- 
-------------------------------------------------------------------------------
 Paul D. Smith <psmith@gnu.org>          Find some GNU make tips at:
 http://www.gnu.org                      http://make.paulandlesley.org
 "Please remain calm...I may be mad, but I am a professional." --Mad Scientist