+1 (non-binding) tested on Apache Mahout trunk
Kind regards,
Stevo Slavic.
On Tue, Sep 10, 2013 at 7:09 PM, Daniel Kulp <[email protected]> wrote:
>
> On Sep 10, 2013, at 12:16 PM, sebb <[email protected]> wrote:
>
> > On 10 September 2013 16:33, Daniel Kulp <[email protected]> wrote:
> >>
> >> -1
> >>
> >> The src.tar.gz and src.zip files have lost their top level NOTICE
and
> LICENSE files. This is a regression from 3.1.0 (and 3.0.5). That
> definitely needs to be fixed. I don't have time today to look into
that,
> but might tomorrow if someone doesn't beat me to it.
> >>
> >> Ran a couple builds with the result of building the source and
things
> look OK.
> >>
> >> I checked the rest of the contents with the tag and everything
looks
> OK. There are three files in the git repo that aren't part of the
> release (/.gitignore, /.gitattributes, and
> /apache-maven/src/bin/.gitattributes), but those files really are
specific
> to our scm and thus don't need to be in the source release.
> >
> > OK, so is it necessary to check the release against the tag?
>
> That's one way, sure.
>
> Personally, I log the trunk/master/branch, find the appropriate commit
for
> "prepare release maven-3.1.1", check that out, then diff that with the
src
> tar ball as well as diff that with the tag to make sure all three
match
up.
> Likely not necessary with git where the tag would apply directly to
that
> commit, but with subversion, it is certainly possible that the three
don't
> match and the diffs make sure to check that. If any of the three
diffs
> find differences, it's an immediate red-flag for further review and
likely
> require a -1 on the vote until resolved.
>
> > Or is that just your personal take on what a reviewer should do?
>
> A reviewer should do whatever they feel is necessary to do a thorough
> review.
>
> > If it is necessary (for at least one reviewer) to do, then the SCM
> > coordinates need to be provided in a transparent manner so any
> > reviewer can do it, and the coordinates need to be part of the vote
> > e-mail "for the public record".
>
> No, each reviewer should do what they feel is appropriate for them.
If
> every reviewer did the exact same thing, we'd get the exact same
result
> from each of them. Some reviewers may checkout the tag, others may
troll
> back master, others may do something completely different.
>
>
> --
> Daniel Kulp
> [email protected] - http://dankulp.com/blog
> Talend Community Coder - http://coders.talend.com
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [email protected]
> For additional commands, e-mail: [email protected]
>
>
