[Top] [All Lists]

Re: [PATCH] xfstests: rework "dmapi" subtree build mechanism

To: Christoph Hellwig <hch@xxxxxxxxxxxxx>
Subject: Re: [PATCH] xfstests: rework "dmapi" subtree build mechanism
From: Alex Elder <aelder@xxxxxxx>
Date: Thu, 10 Mar 2011 16:16:12 -0600
Cc: xfs@xxxxxxxxxxx
In-reply-to: <20110310172943.GD13390@xxxxxxxxxxxxx>
References: <201103041937.p24Jbk5W004828@xxxxxxxxxxxxxxxxxxxxxx> <20110310172943.GD13390@xxxxxxxxxxxxx>
Reply-to: aelder@xxxxxxx
On Thu, 2011-03-10 at 12:29 -0500, Christoph Hellwig wrote:
> >  LIB_SUBDIRS = include lib
> >  TOOL_SUBDIRS = ltp src m4
> > +ifeq ($(HAVE_DMAPI), true)
> > +TOOL_SUBDIRS += dmapi
> > +endif
> Does this work with the way we generate source tarballs on make dist?
> In xfsprogs we have special variables for source files that don't get
> build.

Right now, xfstests has no "make dist".  I'm trying
now to straighten things out a bit in preparation for
making xfstests follow the same model used for xfsprogs
(and the rest).  So, I guess the answer is "no" but
ultimately it should be "yes," though more changes
will be needed to get there.

> > @@ -0,0 +1,30 @@
> > +#
> > +# Copyright (c) 2011 SGI  All Rights Reserved.
> > +#
> I think a license would be more useful than all rights reserved here.
> Except for that the patch looks sane to me, and having a consistent
> build system is a good thing.

I'll add a boilerplate GPL statement in these new files.
What I used was simply mimicing what was present in the
other Makefiles.

In order to avoid re-sending this big patch, may I get
your signoff/reviewed-by on that plan?



<Prev in Thread] Current Thread [Next in Thread>