Re: Intentionally smaller pbulk builds

[Taylor R Campbell <campbell+netbsd-tech-pkg%mumble.net@localhost>]

> Date: Mon, 30 Dec 2024 13:02:36 +0100
> From: Jörg Sonnenberger <joerg%bec.de@localhost>
> 
> On 12/26/24 1:04 PM, Greg Troxel wrote:
> > This is a clue that our insistence on having the bulk build completely
> > fail if there are scan errors is not a reasonable strategy.   Packages
> > that can't find deps should just be marked failed (with some new token,
> > perhaps), and what can be done should be done.
> 
> We've been there and you should be old enough to remember. The result 
> was a lot of silent breakage that no one ever cared to clean up.

Why would this lead to silent breakage?  Surely the broken packages
should wind up in the bulk build results as failed, and anything that
depends on them should wind up as indirect-failed, no?