On Wed, Mar 18, 2015 at 02:09:49PM +0100, Ludovic Courtès wrote: >Tomas Cech skribis: > >> commit 87bafa07ae68084819bb2648bd1042c354d11317 >> Author: Tomáš Čech >> Date: Wed Mar 18 08:57:50 2015 +0100 >> >> gnu: Add agg. >> >> * gnu/packages/graphics.scm (agg): New variable. >> * gnu/packages/patches/agg-am_c_prototype.patch: New file. >> * gnu-system.am (dist_patch_DATA): Add it. > >[...] > >> +++ b/gnu/packages/patches/agg-am_c_prototype.patch >> @@ -0,0 +1,10 @@ >> +--- a/configure.in 2015-03-15 18:59:12.557417149 +0100 >> ++++ b/configure.in 2015-03-15 18:59:29.273416518 +0100 > >Patches must have a comment at the beginning to explain what they do, >why, and what the upstream status is. Will do. >Please pay more attention to these rules before pushing. In general, I >would recommend that you wait for some review before pushing. (This is >sometimes slow, but hopefully others will help reduce latency, >hint hint! ;-)) Yeah, I can see what I'm causing lately. I'll relax with pushing. > >> +@@ -8,7 +8,6 @@ >> + AC_PROG_CC >> + AC_PROG_CXX >> + AC_ISC_POSIX >> +-AM_C_PROTOTYPES >> + if test "x$U" != "x"; then >> + AC_MSG_ERROR(Compiler not ANSI compliant) >> + fi > >It is best to avoid modifying .ac or .am files because that requires >adding a dependency on the autotools.] I will remember that but I'm afraid that in this case there is no configure in tarball anyway and has to be generated. >What problem does this patch solve? I’m confident this can be worked >around without modifying the .ac file, for instance by passing the right >ac_cv_ variable as a ‘configure’ argument. AM_C_PROTOTYPES has been removed from automake. https://lists.gnu.org/archive/html/automake-patches/2011-06/msg00082.html and the reconfigure leads to failure: starting phase `autoreconf' aclocal: warning: autoconf input should be named 'configure.ac', not 'configure.in' configure.in:11: error: automatic de-ANSI-fication support has been removed /gnu/store/2h4lnk22bj479fnr0l2cyx2rvisv9cg7-automake-1.15/share/aclocal-1.15/obsolete.m4:26: AM_C_PROTOTYPES is expanded from... configure.in:11: the top level I'd gladly take your recommendation about passing 'the right ac_cv_ variable' on this. The agg package is old, last it's release was 8 years ago, sourceforge marked that as '(historical)'. It is unlikely that anyone would accept patch. The only reason I spent time on it is because it is required dependency for GNU Gnash. Thanks for review. Best regards, S_W