svn commit: r385202 - head/java/junit

[Kurt Jaeger]

Hi!

> > > Thanks for update! Could you also commit the patch for libreoffice
> > > from this PR?
> > 
> > The patch is submitted to the relevant PR:
> > 
> > https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=199819
> 
> This PR shows the problem with junit-4.11_1, I don't see why it's
> relevant to libreoffice problem with the newest junit.

Because junit was updated because it no longer built ? I see that one
argue both ways, but what's the gain ?

> > If someone from office@ approves it, I can commit it.
> 
> Too late, you have already committed update for junit, so please
> just fix all dependent ports now.

Done for libreoffice.

Does that mean that I have implicit approval to commit the fix that
danilo proposed for java/berkeley-db as well ? So why isn't danilo
doing this 8-} ?

https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=199869

I'm getting confused on those rules about maintainer approval
and other 'approval' rules, and more and more I come to the conclusion
that it's no longer the simple, rational rulebase that one can easily
grok.

If committing to one port implicit makes one responsible for other
ports beyond a certain 'depedency horizont', in a complex dependency
tree this might lead to huge amounts of additional work that
might become so huge as to not do it in the beginning. Is that
a sensible way to handle this ?

> Any comments on the second question in my previous mail?

Ups, I have not seen your second question, probably cutting
down on the quote would have made it easier to find ?

>From what I understand, junit used/included hamcrest in the past and
no longer does. I have not analysed whether there are cases
where junit is needed and hamcrest is not. If danilo wants to
comment on that ?

-- 
pi at FreeBSD.org         +49 171 3101372                5 years to go !