> -----Original Message-----
> From: Steven Rostedt <rost...@goodmis.org>
>
> On Tue, 7 Jul 2020 08:33:33 -0700
> Randy Dunlap <rdun...@infradead.org> wrote:
>
> > >> I was thinking good-list / bad-list.
> > >>
> > >> /me that has been doing a lot of git bisect lately...
> > >
> > > I think it depends on the context. I'd prefer a grammatically awkward
> > > verb that described
> > > the action more specifically, than a grammatically nicer generic term.
> > > In other words,
> > > yes/no, good/bad don't mean that much to me, unless it's obvious from
> > > context
> > > what the effect will be. With something like allow/deny, I have a pretty
> > > clear mental
> > > model of what the code is going to do.
> >
> > That matches what I was about to say:
> > Just using yes/no does not tell someone what they are saying yes or no
> > about.
> > It should be more descriptive, like allow/block.
>
> After doing two days worth of git bisect, good/bad is hardcoded in my head :-p
Maybe I have the same bias, because good/bad there doesn't bother me at all. ;-)
Here is some 'motivated reasoning' on my part...
In the git case, the good/bad terms describe the result status of the test, not
the action that git
is going to take based on that status. It's pretty clear from context that a
'good'
result will cause that commit and other commits to be added to the 'good' set.
I think what
git actually does in constructing the sets is a bit too magical to describe
with a simple
verb.
As an aside I just looked up 'git-bisect' documentation, and found it has
support
for changing the terms used ('git bisect terms ..') so you can use words like
'fast/slow'
or 'fixed/broken'. That's something I didn't know about. :-)
-- Tim
