21 Jun
2024
21 Jun
'24
6:49 a.m.
Larry McVoy <lm(a)mcvoy.com> writes:
I've been up since 12:22am (psyched for fishing,
couldn't sleep)
so
maybe I'm not on point, but what is the problem that this
discussion
is trying to solve?
The complexity of the autoconf-based build process contributed to
the xz-utils backdoor attempt. (Here's Russ Cox's writeup:
https://research.swtch.com/xz-script) So, to what extent is the
complexity of autoconf _needed_ nowadays? For some cases, it's not
needed (and might never have been needed). For others, it seems
like it might still be needed. What about the in-between cases?
Can we do something different that gets us 90% of what autoconf
provides in those cases, but with only 10% of the complexity (to
use those commonly-provided figures)?
Alexis.