> d> In the latest O'Reilly Gnu Make book, the author > d> uses the $(eval $(call ...)) pattern often. It seems to > d> offer subroutine-like functionality. > > d> Is this a preferred pattern in the community? > > Well, if you need that functionality it's preferred. I wouldn't use it > just for the sake of using it, myself. The main problem is it makes > your makefiles much less portable. But, if you don't mind that it's > OK. Certainly there are certain things that it's difficult or > impossible to do without eval, in particular.
Thanks much for your responses to this and to my earlier posting. By 'less portable', I'll assume that you mean with respect to using our Makefiles with makes other than gnu make. I haven't yet experience enough to know, but I'm hoping that a small library of eval/call functions will enable our group to much more easily construct and maintain one makefile per project (typically a few hundred C/C++ source files in several libraries and applications). In using a single, large Makefile, again, I'm following the lead of Mr. Mecklenburg, who (I think) suggests that such an approach offers speed, a more holistic dependency tree, and easy ad hoc modifications. Best wishes. -- Don _______________________________________________ help-gnu-utils mailing list help-gnu-utils@gnu.org http://lists.gnu.org/mailman/listinfo/help-gnu-utils
