On Tuesday 07 June 2011, Stefano Lattarini wrote: > On Tuesday 07 June 2011, Stefano Lattarini wrote: > > On Tuesday 07 June 2011, Peter Rosin wrote: > > > >> This test no longer checks if $AUTOMAKE -a copies over compile, as > > > >> that is done manually now. I assume this aspect of $AUTOMAKE -a is > > > >> tested elsewhere. Or is it? > > > >> > > > > Yes, in 'subobj.test'. > > > > > > The same argument could be made about the other instances where the > > > script is brought in explicitly. Seems like a bit of a fluke that > > > subobj.test covered the compile script. > > > > > Agreed. I now think we should have a centralized test where to check > > for files installed with `--add-missing', not to risk reduced coverage > > anymore. Patch coming up soon ... > > > Done in the attached patch. It is more complex than I'd like, so I'll > wait until sunday or so before pushing, to allow more time for reviews. > > Regards, > Stefano > A couple of minor fixes I'd like to squash in. First, a typofix in a comment; second, an internal sanity check looking out for bogus input passed to the `check_' subroutine.
Regards, Stefano -*-*-*- diff --git a/tests/add-missing.test b/tests/add-missing.test index a8742f6..e460a95 100755 --- a/tests/add-missing.test +++ b/tests/add-missing.test @@ -94,6 +94,7 @@ check_ () '== Makefile.am ==') what=Makefile.am;; '== configure.in ==') what=configure.in;; '== Files ==') what=LIST;; + '==.*') framework_failure_ "invalid input line: $line";; ''|'#%'*) : empty line or ad-hoc comment, ignore;; *) @@ -210,7 +211,7 @@ check_ --run-aclocal <<'END' compile == configure.in == # Using AM_PROG_CC_C_O in configure.in should be enough. No need to -# use AC_PROG_CC too, not to define xxx_PROGRAMS in Makefile.am. +# use AC_PROG_CC too, nor to define xxx_PROGRAMS in Makefile.am. AM_PROG_CC_C_O END