Re: CVS commit: src/usr.bin/make

[Roland Illig <roland.illig%gmx.de@localhost>]

10.12.2021 01:07:22 Robert Elz <kre%munnari.OZ.AU@localhost>:

>     Date:        Thu, 9 Dec 2021 22:25:58 +0000
>     From:        "Roland Illig" <rillig%netbsd.org@localhost>
>     Message-ID:  <20211209222558.CDF22FAEC%cvs.NetBSD.org@localhost>
>   | No functional change.
>
> That obviously wasn't true.   That means that you just guessed at that,

I cannot defend against this argument. Indeed, I guessed, but that guess was backed up by several hundred lines of test code, essentially all cond-*.mk and especially cond-short.mk.  These tests made me reasonably sure that the behavior of the code wouldn't change.  The behavior did change though, and the change went unnoticed due to my lack of testing.  So even 100 % code coverage does not prevent all possible bugs.

I guess there's really no way around running the whole build before each commit, to reach a build success rate of 99.9 %.

Roland