[HEADSUP] A new policy. First victim: Gerrit

Corinna Vinschen vinschen@redhat.com
Thu Jun 10 16:10:00 GMT 2004

On Jun 10 10:15, Yaakov Selkowitz wrote:
> Corinna Vinschen wrote:
> | While it's not overly difficult to get votes, it's fairly difficult to get
> | a review.  So the idea is to ease getting new packages into the Cygwin
> | distro for well known good packagers.
> If I may say so, as a package maintainer myself, that unless I missed
> something I've never seen *clear* guidelines for making a good-to-go
> review.  One can always find problems :-) but saying about something

Sure.  But a good packaged package (gulp) isn't rocket science.  The
binary package should work, the paths should be correct, collisions with
other packages should be avoided, a bunch of README's and a Cygwin
specific README would be nice. 
The source package should be buildable if all build dependencies are
resolved, a description of the build process if it's actually complex
should be given in the Cygwin specific README. 
That's it, basically, isn't it?

> that it DOESN'T have problems is harder.  Having guidelines for how much
> you have to test the package would make it easier to call it GTG without
> feeling stupid later when a major bug arises -- either because you

You will never be able to gurantee that a package has no problems.
It's impossible to test all border cases.  If the major operations
work, the package is fine in the first place.  If users on the Cygwin
list find a problem 10 seconds after the announcement has been sent,
bad luck.  The serious maintainer will take the opportunity to either
convince the Cygwin list that it's a feature, not a bug, or s/he will
create a -2 package pretty soon.

> If the votes aren't the problem, then maybe it should be like this:
> either three votes without waiting, or if after a week there's only one
> aye vote and no vetos.  If a package has three votes and doesn't need a
> review, why should it have to wait a week?  Or is that what you meant?

A +3 within a day or two is sure ok as well.

> | Does that sound fair?
> Alright, but I think this should be made clear when people become
> review-exempt, coming from you and/or cgf, to avoid confusion.

It's not an automatism.  The maintainer should ask for the right.  Let's
try to find a way which is based on mutual trust.  If it doesn't work out,
we would have to stop this again.  But it would be a pity.

> ( BTW, what's the story with the PPL's???  I haven't got any response to
> my questions about this.  Is everyone in charge of tracking their own
> ITPs now, like Gerrit has done? )

Well, Daniel seemed to have lost interested unless the process can't
be entirely automated.  Igor has send two or three reports and then
stopped his effort again.  I, for one, would be very excited about
somebody stepping up to take over this list.


Corinna Vinschen                  Please, send mails regarding Cygwin to
Cygwin Co-Project Leader          mailto:cygwin@cygwin.com
Red Hat, Inc.

More information about the Cygwin-apps mailing list