prebuild sanity checks

[Brooks Davis]

On Thu, Aug 18, 2005 at 01:58:45AM +0200, Benjamin Lutz wrote:
> [...]
> > Another option might be a new variable (or variables) that ports that
> > tend to break spectacularly and unobviously can set like:
> >
> > BUILD_DEVS=	null zero
> 
> As a potential user of such a variable, I wonder how I'm supposed to
> figure out which basic system facilities are required by a given piece
> of software.

Either by having it fail and debugging it or by doing a build with one
of the common culprates missing from devfs.  In theory it would see that
you could do a periodic sweep using the package cluster.

> I think the right thing to do here would be to have the software react
> more sensibly to such a problem, ie bail out with an error message. In
> other words: have the people upstream change their software.

In theory yes.  In practice, I'm sure a lot of software authors won't
care about supporting this environment.

-- Brooks

-- 
Any statement of the form "X is the one, true Y" is FALSE.
PGP fingerprint 655D 519C 26A7 82E7 2529  9BF0 5D8E 8BE9 F238 1AD4
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
Url : http://lists.freebsd.org/pipermail/freebsd-ports/attachments/20050817/2a805fcc/attachment.bin