[kde-edu]: [was: Re: Linker error in marble tests)] - Communication problems
tackat at t-online.de
tackat at t-online.de
Sat Jul 12 11:28:30 CEST 2008
> > Although I more or less introduced this error by enabling unit
> > testing in marble, I dislike the way I am addressed for
> > errors like this. I sometimes feel that people are disappointed
> > because I am working on windows more than on linux
> > and as such people behave different towards me and any errors.
Well the actual problem here is this:
Patrick has introduced some unit tests during feature freeze
which I consider a pretty perfect thing to do: We'd like to fix bugs and
prevent them. Introducing unit tests to prevent further bugs is
as important as fixing bugs as I don't want to fix issues on one hand
and introduce others at the same time. And this can happen even
during feature freeze as it's an act of bug hunt and prevention.
Due to working on Windows it got unnoticed by Patrick that the same
source code wouldn't compile on Linux.
I personally do see it as a requirement that any changes that might
potentially introduce new issues with Linux need to get
safely get verified by a different person to compile on Linux.
I even usually ask myself other people at this stage whether my source
code changes compile immediately after I've committed the source
code. This is even more important to do if you develop on Windows:
Don't understand me wrong: I consider the Windows plattform
very much important to support. However most people contributing
to KDE right now are using Linux, so it's very much important not to
make those angry.
That all being said I'd like to emphasize here that Patrick has done
an unbelievable outstanding job during his GSoC activities and exceeded
my expectations in several ways.
I'm deeply thankful for all things he has done so far.
He is always helpful and as you might have noticed helps other
newcomers and GSoC students to get started -- often enough in
ways that go way beyond what I'd expect from a normal GSoC student.
> That's why I asked you yesterday to help this user, as tackat and
> ingwa are not there, I thought as a Marble person you would know about
Yeah sorry, I'm currently in Limerick Ireland at the OpenStreetMap
> someone else at any time. This role is only to ensure that the general
> Release rules are followed and this position is intended to ease the
> Release Dude work load. I feel I have quite failed for this release.
I don't think you have failed and I'd encourage and ask you to do your
work in the
I do admit that for the Marble project things were a bit difficult due
major refactoring we had to do on all ends to get Marble in a state
where it could
get used as a library by other applications (and to stay future-proof
I don't want to go into all details ( I guess this is more easy to
discuss during aKademy )
but I've already talked to Jens-Michael at the conference and we'd like
we can!) do a lot better for the next release as it will be more of a
(than this release which has seen a lot of introductions of refactoring
introduction of new frameworks to allow a future proof development after
I plan to create a special QA squad for Marble that checks for package
and that checks for other things. I feel that together with a dedicated
team this is one of the most important things that need to get started
> The problem is with Marble development in general as they were many
> changes in the code while we are on freeze.
Yes. I believe that Marble is probably the youngest and most rapidly
application in KDE-EDU. This is partially reason for the issues we have
Also we had decided to start a lot of new frameworks and groundwork for
was needed to get Marble away from ugly workarounds, hardcoded stuff and
solutions in core parts of Marble development. I admit that we have
underestimated the amount of work in lots of places, so we were a bit
schedule when the feature freeze set in and we had to do some catchup in
pragmatic ways to finish things right in time for the release. The only
solution would have been to revert Marble 0.5 for KDE 4.1 which wouldn't
been a real alternative either.
Then there was also a vitally important feature that got only completed
after feature freeze (OSM-support with a delay of a week) which was
enough that I wouldn't have been able to justify postponing its
only after KDE 4.1. Anyways, I think we are on a good way to
deliver a Marble version which will be of good quality for KDE 4.1.
And I see the chance that we'll be able to do a lot better for the
Marble 0.7/KDE 4.2 release. And we should discuss these things
in depth during Akademy.
> even dependency chain... I don't think they were necessary during
> freeze and I said so both via mails and on IRC.
> kde-edu mailing list
> kde-edu at mail.kde.org
More information about the kde-edu