xfs
[Top] [All Lists]

Re: [PATCH] xfsprogs: Make the compile output cleaner V2

To: Christoph Hellwig <hch@xxxxxxxxxxxxx>
Subject: Re: [PATCH] xfsprogs: Make the compile output cleaner V2
From: Dave Chinner <david@xxxxxxxxxxxxx>
Date: Thu, 14 Jan 2010 21:11:28 +1100
Cc: xfs@xxxxxxxxxxx
In-reply-to: <20100114065123.GA32559@xxxxxxxxxxxxx>
References: <1263377007-4500-1-git-send-email-david@xxxxxxxxxxxxx> <20100113220814.GA17310@xxxxxxxxxxxxx> <20100113223001.GR17483@xxxxxxxxxxxxxxxx> <20100114065123.GA32559@xxxxxxxxxxxxx>
User-agent: Mutt/1.5.18 (2008-05-17)
On Thu, Jan 14, 2010 at 01:51:23AM -0500, Christoph Hellwig wrote:
> On Thu, Jan 14, 2010 at 09:30:01AM +1100, Dave Chinner wrote:
> > > >         include/builddefs include/platform_defs.h
> > > 
> > > >  distclean: clean
> > > > -       rm -f $(LDIRT)
> > > > +       $(Q)rm -f $(DISTDIRT)
> > > 
> > > What does this have to do with the rest?
> > 
> > That's the "make clean doesn't cause configure to run again" fix.
> 
> Which is not related to the clean output build system..
> 
> I'm pretty sure we did this intentionally, but I'm open to a discussion
> about changing it.

It seems completely spurious to me - normal practice is to blow away
the config on a distclean, not a plain clean. And seeing as the
patch is about removing noise during the normal build, I figured
this qualified as a bug causing noise.

I can separate it out into another patch, but I'd really prefer that
make clean doesn't result in running configure again. Especially
considering the build dependencies are very broken so the only way
to guarantee a build catches all changes is to do 'make clean;
make'.

Cheers,

Dave.
-- 
Dave Chinner
david@xxxxxxxxxxxxx

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