This is the mail archive of the automake@gnu.org mailing list for the automake project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

automake and multiple inputs for config.status


I hope this is suitable for the cross-post.

I'm currently using automake-1.5 and autoconf-2.53 in this situation.

I have a case where I need to produce a Makefile from a Makefile.in
(generated from a Makefile.am) and other files.  As it happens, I need one
of these files to come before the Makefile.in but I *think* my patches
handle that case already.

One place I *am* seeing a problem is with config.status and the
automake-generated rule to generate the Makefile.

The instantiation template "documented" for Makefile by config.status is:

    Makefile:foo:Makefile.in

and if I do a "plain" config.status then Makefile is produced from foo
followed by Makefile.in (with appropriate substitutions).

However, if somebody incants "config.status --file=Makefile" then the
generated Makefile does not contain "foo".

Furthermore, the rules to generate Makefile look like:

 Makefile: ...
	...
	CONFIG_FILES=@ $(SHELL) ./config.status

so again the generated Makefile only contains Makefile.in instead of foo and
Makefile.in .

config.status clearly knows the template list for the files.  It looks like
if one specifies a target file with no template list that config.status
assumes the "simplest" case instead of using the "listed default" case.

Anybody have any solutions or thoughts?

H


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]