test_source harmful?

[Martin Pool]

I'd tend to agree with Robert that a pqm rejection, if people ever hit
it, is far too strong medicine.

I prefer not to have it, as it does make diffs noisier if it's added
or removed incidentally, but a rejection is a big whack.

If we are going to have a test, it musn't be something that depends on
the working state, as Marius's original patch (that looked essentially
for whitespace in 'bzr diff') did.  Though it was well intended it's
actually very disruptive in practice when you're trying to merge
things.  Either a strict rejection or no test at all is better.

Probably what we need is an --ignore-trailing-whitespace option to
diff, and also to annotate.  It would be nice if they could be turned
on by rule preferences too.

>> Isn't the solution simply to configure your text editor to remove
>> trailing whitespace before you save the file?
>
> Thats one approach.

I'd treat that as a 'should' but not a 'must'.  It can be fairly
easily done in emacs and vim at least, and in fact I think many people
have it on already.

-- 
Martin <http://launchpad.net/~mbp/>