ant-ivy-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Xavier Hanin" <xavier.ha...@gmail.com>
Subject Build promotion?
Date Thu, 04 Jan 2007 14:17:00 GMT
Other message I forgot to send to the list:

---------- Forwarded message ----------
From: Xavier Hanin <xavier.hanin@gmail.com>
Date: Jan 4, 2007 12:05 PM
Subject: Re: Build promotion?
To: andreas@sahlbach.com

On 1/4/07, Andreas Sahlbach <andreas.sahlbach@gmail.com> wrote:
>
> On 1/4/07, Xavier Hanin <xavier.hanin@gmail.com> wrote:
> > Thanks for sharing these useful tips and opinion. I agree  that as often
> > it's a matter of taste, but I think you point something interesting: you
> say
> > that you have a build which is reproducible, but you don't want to take
> the
> > risk. So there is a risk. So you are not confident your build is
> > reproducible.
>
> The problem is: there is always a risk. If you and I would be the only
> developers in our development team, the risk would be very low I
> guess. But that's not the case. My build is used from around a dozen
> developers, if you add the people Matthias Kilian is advising, call
> that around 20 developers. So we now have 20 opinions regarding build
> standards and these opinions include ideas like "I've made a last
> minute fix that missed the build, let's move the release tag".


I agree that you always have to deal with a lot of opinions, and consider
them very carefully, especially when they come from people who pays you :-)
But when people suggest things like moving a tag, I spend time (usually
after things have calmed down) explaining why it's dangerous and what the
consequences could be. Then if the development team still want to be able to
introduce unreliable tricks in the release, it's up to them...

On the other hand: if something goes wrong during a release, I am in
> deep, deep trouble. A lot of people will become very angry with me
> little freelancer, because if we have made a mistake, a lot of stuff
> does not work here and this could possibly end up in production
> problems (which is a very expensive thing to do if you work for an
> automotive enterprise like I do).


I understand, that's why introducing automatic comparison of the builds can
ensure that you end up with exactly what you want, and be confident people
won't be angry about you. But I agree that this is not suitable in all
contexts.

Xavier


> So my build is reproducible because I automatically store svn-url and
> svn-version-number into every build I make, allowing only to publish
> into the enterprise repository, if the svn wc, the build is based of,
> is fully checked in and is not a mixed wc. But still I don't want to
> take the risk, call me a coward if you want :)
>
> Regards,
>
> Andreas
>
> --
> Andreas Sahlbach
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message