Subject: Re: CVS commit: basesrc
To: Simon Burge <simonb@netbsd.org>
From: Bill Studenmund <wrstuden@nas.nasa.gov>
List: tech-userlevel
Date: 12/03/1999 18:14:37
On Sat, 4 Dec 1999, Simon Burge wrote:

> Bill Sommerfeld wrote:
> 
> > hmm.  i certainly would leave it in there, since new dirs get added to
> > the source tree all the time, and i rarely if ever switch around to
> > using a different obj tree with the same source tree..
> 
> I meant the same obj tree with different source trees.  Ok, I agree that
> this might not be a common case.  Is it easier to just not handle it?

Probably. Though couldn't you just fix things with a make obj when you
decide to use the new tree?

> It removes and re-creates the symlink (if you're using symlinks) for
> each obj directory.  This is noticebly slow especially when using a NFS
> v2 fileserver for the source tree.  To make things run faster, it'd be
> nice to have a clean-obj-includes or obj-clean-includes target, but this
> is obviously overboard.

Actually an "objdependall" target might be reall cool. Charles was telling
me what needs to happen for it, but I didn't do it. It seemed quite
simple..

> How about (leaving out all the .if/.endifs):
> 
> 	make share/mk
> 	make cleandir
> 	make obj
> 	make share/tmac
> 
> With this, any new tmac's get created in the tmac obj dir (if it exists)
> and not in the source directory, which is one thing I was trying to
> stop.  In fact, because we build the libraries with MKMAN=no we could
> move the "make share/tmac" to after the libraries are built...

That would probably be good. :-)

Take care,

Bill