[Qgis-psc] Backporting?

Jürgen E. Fischer jef at norbit.de
Fri Mar 7 03:37:53 PST 2014


Hi Paolo,

On Fri, 07. Mar 2014 at 11:22:16 +0100, Paolo Cavallini wrote:
> If I interpret it correctly, the first stumbling block is a general agreement
> on a release strategy, then automatization may proceed.
> Is this correct? In case, we can discuss about this in the meeting of this
> afternoon.

Not at all.  IMHO it's the limited testing.   We need more testing before the
release to make the releases better.  That could avoid having to make bugfix
releases for trivial things - that only got in unnoticed, because of the lack
of testing.  Point releases for severe bugs is a different thing - not sure if
we had any yet.  So that doesn't really apply here.

If we had better pre-release testing, there would be less need for fixes to the
release, that would make everyone happier, make us look better (which is
usually not my line of reasoning) and avoid extra effort.

Most bugs are spotted by users and if we could get them somehow to test
voluntarily before the release instead of involuntarily after it would be for
our all good (users & devs).   I think the users have more resources to
contribute there, than the devs have resources to spend on point releases.
Money of course could shift the line, but I doubt that there are enough funds
to turn it around.

I don't really care how small the avoided extra effort might be.  Packaging is
admittely some effort - but for the most part just up/downloading sources and
binaries, running scripts and monitoring their progress.  But it's extra
releases are avoidable and IMHO it still a big waste (and image damage; also
that's just indirectly relevant to me), if a small bug requires a new point
release, before the machine can cool down from the last build.

We've also discussed renaming and annouce the weekly standalone snapshots to
"release candidates" after the feature freeze to encourage testing (again
without any significant extra work).  I think if that really yields more
testing, that would help much more, than spending dev time on point releases.

That's one idea to get more testing.  Others?
 
I hope this clear my line of reasoning more.  I don't think it has been
actually challenged.  But I might have missed some in all this discussion, if
so please repeat - I'm not getting tired of it either (well, another hour that
could have been spend on other things ;)


Jürgen 

-- 
Jürgen E. Fischer         norBIT GmbH               Tel. +49-4931-918175-31
Dipl.-Inf. (FH)           Rheinstraße 13            Fax. +49-4931-918175-50
Software Engineer         D-26506 Norden               http://www.norbit.de
QGIS PSC member (RM)      Germany                      IRC: jef on FreeNode                         

-- 
norBIT Gesellschaft fuer Unternehmensberatung und Informationssysteme mbH
Rheinstrasse 13, 26506 Norden
GF: Jelto Buurman, HR: Amtsgericht Emden, HRB 5502




More information about the Qgis-psc mailing list