[cvsnt] Re: CVS Use case

[Gerhard Fiedler]

Matt Schuckmann wrote:

> I suppose that you could infer that a file has no changes associated with
> a task if all the revisions on that Task's branch are the result of
> merges from the parent (identified by mergepoints) but that could get a
> little confusing and would require some sort of special script. 

I guess that's probably the best you can get: a script that compares the
diff of the branch with the diff on HEAD and strips out everything from the
branch diff that's also in the HEAD diff.

At this point, it seems to me that relying on floating branches for
critical procedures is kind of scary. And as you say, not being able to
control when new code becomes "active" on the branch is kind of like
somebody secretly changing code on my branch... All of a sudden, stuff
doesn't work anymore because a function I'm using changed, but I didn't
change it and a diff on my branch doesn't show any change... How do you
find the problem? The traditional diff review ("yesterday it worked, today
it doesn't, so let's look at what changed since yesterday") wouldn't work
anymore. Floating branches seem to be debugging hell, besides the not very
reliable implementation.

Gerhard