automake15-1.5-18 has the most common review issues fixed, please check...

rpmbuild trashing config.guess and config.sub 
(c.f. https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=225296#c3)
also shows effect here:

It causes the config.guess/config.sub being distributed though the automake15
package to be replaced.

Instead of the ancient versions dated 2001-<something> contained in the
automake-1.5 tarball, the resulting rpm contains the a bit less ancient versions
being distributed by rpm (2003-<something> on FC6).

IMO, automake15 should either ship the original version from the automake-1.5
tarball or an up-to-date version of config.guess/config.sub (from gnu.org).

automake15-1.5-19 uses ./configure instead of %configure. Maybe I'll replace
config.{guess,sub} with more recent ones when I have some spare time as I need
to do a test build of some packages then.

What's the review status ? Is it approved ?

Why are you asking me? 

I never formally reviewed this package, nor do I know how merge reviews are
supposed to work.

probably because your comment looked like some kind of review to me ;-)

I'll probably review formally when I have time.

I attach a patch for configure to avoid rerunning autoconf.

I patched as the first patch with
%patch10 -b .versioning
which allowed to restore timestamps by doing
touch -r configure.versioning configure
touch -r configure.in.orig configure.in

I think it should be part of automake15-versioning.patch
but I let that to you, unless you agree with those changes (including
adding the -b .versioning) in that case I could do a patch against
the cvs version changing automake15-versioning.patch.

Created attachment 150101 [details]
patch configure and Makefile.in to avoid rerunning autotools

The patch now applies with 
%patch10 -p1 -b .versioning

In this patch there are changes that should be in 
automake-1.5-subdirs-89619.patch
(lines like
-       @set fnord $(MAKEFLAGS); amf=$$2; \
+       @set fnord $$MAKEFLAGS; amf=$$2; \
)

And also the patch automake-1.5-info.patch leads to a 
Makefile.in which isn't what automake would generate based
on the patched Makefile.am. Not a big deal, but i'll propose
an updated patch once this has been digested.

As a side note automake was run during the make. Is it 
on purpose? it seems to me that the automake run was the
automake from the package itself. 

I've added your changes in -22

Created attachment 153256 [details]
complete patches of Makefile.in to avoir rerunning autoconf

This patch allows not to run autoconf (and the autoconf call 
is commented out). As a side note it would be better
to cal autoconf-2.13. But yet better not call anything...

Package has been deprecated. Closing.