If the configure script detects that configure.in or Makefile.am are newer, it'll attempt to regenerate the files- in attempting to regenerate the configure script it will *always* fail. autoreconf fixes this nicely. Reproducible: Always Steps to Reproduce: 1.ebuild ppmd-9.1.ebuild unpack 2.touch ${s}/Makefile.am ${s}/configure.in 3. ebuild ppmd-9.1.ebuild compile Actual Results: Attempts to regenerate configure, and dies in the process. Expected Results: Shouldn't be The applied patch creates Makefile.am and configure.in *after* it creates Makefile and configure. I think the forum poster got unlucky, and Makefile.am and configure.in ended up w/ a different timestamp then configure/makefile, thus exposing the broken aclocal.m4 . note, emerge info left out (this is a break in the source, not the ebuild);
Created attachment 21778 [details, diff] ppmd-9.1.ebuild patch Fixed a makefile error (over-riding the users cflags), along w/ a permenant fix for correcting the aclocal.m4 problem.
Pardon; the "expected results" I seem to have screwed up- expected result is that this ebuild emerges rather then possibly borking.
Created attachment 21815 [details, diff] ppmd-9.1.ebuild patch, v2... Should (hopefully) be the final tweak needed.
Comment on attachment 21815 [details, diff] ppmd-9.1.ebuild patch, v2... Contains the original aclocal.m4 fix, and a switch from using 'head -3' to 'head -n 3' due to the former being deprecated...
fixed in cvs, thanks brian