Re: [PATCH 2/4] bisect: replace hardcoded "bad|good" by variables

[Matthieu Moy <Matthieu.Moy@grenoble-inp.fr>]

Christian Couder <christian.couder@gmail.com> writes:

> "old/new" is not more generic than "good/bad".

I disagree with this. In any case, we're looking for a pair of commits
where one is a direct parent of the other. So in the end, there's always
the old behavior and the new behavior in the end.

In natural language, I can write "terms good/bad correspond to the
situation where the new behavior is a bug and the old behavior was
correct" and "terms fixed/unfixed correspond to the situation where the
new behavior does not have a bug and the old one does", so I can
describe several pairs of terms with old/new. When looking for a bugfix,
saying "NAME_GOOD=new" seems backward. I would read this as "the good
behavior is to be new", while I would expect "the new behavior is to be
good".

> and as "good/bad" is older and is the default we should keep that in
> the names.

I agree with this part though. If people working with the bisect
codebase (which includes you) are more comfortable with good/bad, that's
a valid reason to keep it.

IOW, I still think old/new is more generic, but that is not a strong
objection and should not block the patch.

-- 
Matthieu Moy
http://www-verimag.imag.fr/~moy/