Resurrecting an old thread, for reference see:

<http://debbugs.gnu.org/cgi/bugreport.cgi?bug=8360>


On Saturday 09 April 2011, Ralf Wildenhues wrote:

> Hello Pitamila,

>

> let's keep the mailing list in Cc:, please. I would've seen your reply

> earlier then.

>

> * pitamila@free.fr wrote on Mon, Apr 04, 2011 at 04:02:53PM CEST:

> > ----- "Ralf Wildenhues" a écrit :

> > >

> > > thanks for the report.

> >

> > thanks for the support, I hope my answer is not too late.

>

> There's (almost) never a "too late" when giving useful feedback!

>

> > > * pitamila@free.fr wrote on Mon, Mar 28, 2011 at 09:17:09AM CEST:

> > > > I'd like to make a rpm buildpackage in my distcheck-hook rule.

> > > > My 'distcheck-hook' rule in my root Makefile.am is never called.

> > > > I tryed to define a 'distcheck-local' too without success.

> > > > Is there something more to know than 14.4 of automake manual?

> > >

> > > Can you show the Makefile.am file? distcheck-hook should work.

> >

> > Find attached the Makefile.am of my doc/ subdir

> > In fact I try to call a clean-local without success (I suppose

> > it's the same bug than my distcheck-local not beeing called

> > in my top Makefile.am)

>

> Yep. And you have indeed stumbled over a limitation in automake;

> I could call it a bug, but I'm inclined to document it rather than

> fix it as it is worked around easily.

>

> Quoting your Makefile.am:

>

> clean-local :

> -rm -rf html/* html @PACKAGE@.pdf latex/* latex

>

> There is a literal TAB after 'clean-local'. automake's parser is pretty

> dumb when it comes to distinguishing rule commands from rule lines from

> variable assignments, mostly because it is hard to get things right when

> hard-core GNU make code is used.

>

> Anyway, in this case it gets things wrong due to the TAB. Solution is

> to remove the TAB before the colon, maybe replacing it with a number of

> spaces. Then everything will work out fine.

>

I've ended up writing a documentation patch for this, as well for another

unrelated limitation encountered by Roberto Bagnara in an older thread on

the automake list, see:

<http://lists.gnu.org/archive/html/automake/2010-08/msg00074.html>

Since I was at it, I've written two tests verifying the documented

behaviour.


I will wait a few hours before pushing (to maint), to allow for reviews.


Regards,

Stefano