Re: Should this makefile really hang make?

[Paul Smith]

On Sun, 2011-11-20 at 08:26 -0800, Bryan Ischo wrote:
> Is there any validity to the argument that SECONDEXPANSION implies 
> "second variable expansion", which implies that the variable expansion
> would happen before the pattern substitution (even for the purposes of
> creating variable names using the pattern)?

Well, without second expansion none of this ambiguity is possible so it
doesn't really help to use that as a flag to choose between the two
orders.

However, with "normal" behavior it IS true that the variable expansion
comes first (during makefile read-in), before stem expansion (which
happens in the second phase of the make process).  This might be a good
argument for changing the order used by GNU make today.  On the other
hand, we've had releases in the field for a long time which use the
current order and it would be a backward-compatibility break to switch
them now.

> If this strict ordering is not desired, then maybe it makes sense to
> add a substitute character that can be used in place of % for either
> the pattern rule or the function

The option I was suggesting was to allow "%" to be escaped with a
backslash, when we wanted to avoid having it interpreted as a stem.

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