<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 TRANSITIONAL//EN"> <HTML> <HEAD> <META HTTP-EQUIV="Content-Type" CONTENT="text/html; CHARSET=UTF-8"> <META NAME="GENERATOR" CONTENT="GtkHTML/3.32.2"> </HEAD> <BODY> On Tue, 2011-06-14 at 09:26 -0700, Dan Nicholson wrote: <BLOCKQUOTE TYPE=CITE> <TT>On the other hand, I believe non-recursive make adds to the complexity</TT> <TT>of the build rather than reducing it. With classic recursive automake,</TT> <TT>it's very easy to divide and conquer tasks into subdirectories. Since</TT> <TT>each Makefile is separate, you can tailor it specifically to what's</TT> <TT>going on in that directory (e.g. creating manpages) and you don't have</TT> <TT>to worry about variable or pathname collisions at all.</TT> </BLOCKQUOTE> That's my number one motivation. So nicely put. There will be more bugs in non-recursive makefiles. Automake is already a significant learning curve, and there have been plenty of bugs in any makefile with unrelated target types. There is a high turn over rate of maintainers, when there is one. When anyone makes a change in a non-recursive makefile, however simple it maybe, he has to unit test the 5 other targets in the makefile, not just the one he is familiar with. With subdirectories, there is no risk of breaking those other targets. This makes it possible for people with lower level of Automake skills to perform tasks that would otherwise be too risky. These are more project management than technical issues. We can't assume that everyone has all the skills to do everything. Anyway, I ran the test before and after the patch and there is no performance gain whatsoever. </BODY> </HTML>