Subject: Re: remove default of STRIPFLAG=-s ?
To: Frederick Bruckman <firstname.lastname@example.org>
From: James Chacon <jmc@NetBSD.org>
Date: 02/11/2004 13:01:03
On Wed, Feb 11, 2004 at 12:38:17PM -0600, Frederick Bruckman wrote:
> On Wed, 11 Feb 2004, James Chacon wrote:
> > On Wed, Feb 11, 2004 at 08:14:39AM -0600, Frederick Bruckman wrote:
> > > On Thu, 12 Feb 2004, Luke Mewburn wrote:
> > >
> > > > I've been considering for a while that we should remove the
> > > > default build setting of STRIPFLAG=-s (which causes programs to
> > > > be stripped (with strip(1)) as they're installed (by install(8))
> > > >
> > > > This change would provide far more useful debugging information
> > > > at the expense of slightly more disk space. Here's a comparison
> > > > of a stripped versus unstripped full "distribution + x11" build:
> > > > 291564 destdir.i386
> > > > 281409 destdir.i386.stripped
> > > > (These figures are in KB)
> > >
> > > I bet the difference will be a lot greater if the build is with "-g".
> > > Maybe just warn folks to add STRIPFLAGS=-s when using COPTS=-g?
> > Ummm..doesn't stripping kind of remove the usefullness of -g?
> Well, no. The unstripped one is still in the object directory. You
> could use it to analyze a core dump of the stripped one, since the
> image in memory is the same.
You're assuming folks keep all their object directories then which may not
be the case. So this advice needs caveats like this mentioned..