After I upgraded this morning (6h30 AM +0000) make to its latest version in portage, many applications failed to get merged, for instance: sys-apps/busybox-1.19.2 dev-libs/gobject-introspection-0.10.8 sys-libs/glibc-2.13-r4 I got very strange errors from make. The first one I saw was caused by failing to glob (on gobject-introspection), and I worked around by changing the Makefile to explicitly do wildcard expansion on each line. The typical message was make[2]: *** No rule to make target `/usr/include/glib-2.0/glib/*.h', needed by `GLib-2.0.gir'. Stop. which clearly shows that make did not expand the dependency with *.h correctly (and I do have files to glob). As I found, the makefile is quite involved, using "generating rules" and so on, but it compiles flawlessly with make-3.81-r1. Next, I thought it was a deeper problem, and tried to rebuild the glibc. Again, it crashed, and again on a "file not found" situation. The build log for the glibc is attached. Finally, I tried just to "continue" with the merge, and found that busybox also had a problem, and the Kbuild system was simply not working: even the generating script for the numerous Kbuild files on each directory did not do its task, and the make process stoped. Reverting to make-3.81-r1 compiled busybox and gobject-introspection. As my glibc was untouched and needed no upgrade (and 2 cases "confirmed" sounds good) I did not try to recompile it. Reproducible: Always Steps to Reproduce: 1. upgrade to make-3.82-r2 2. emerge -va1 =busybox-1.19.2 Actual Results: The make step of the compilation of busybox does not completes, exiting with error Expected Results: build the package.
Created attachment 286827 [details] Build log for glibc
Created attachment 286829 [details] Build log for llvm I found also that I've tried to compile llvm after the make failed (don't even recall why...) but anyway here's the build log, with a further make error to get a feeling...
It looks like make has been in flux over that last couple of days. Does make-3.82-r3 work for you?
*** This bug has been marked as a duplicate of bug 383311 ***
