On 11/23/11 12:01 PM, eric b wrote:
Hi Jürgen,
Le 23 nov. 11 à 11:30, Jürgen Schmidt a écrit :
The current step is :
- IP clearance (not completed if I'm not wrong)
- fix build issues and see what disasters caused the removing of
important tools, like dmake and some other.
yes and every helping hand is very much appreciated. I would really
like to see more people working on this
Next one will probably be :
- consolidate the build (on every OS)
what do you mean here
Currently, code is removed, and every commit, we are not sure the build
can finish. I'd consider this as undefined / chaotic state, that we need
to consolidate.
- optimize configure command line
again, can you explain what exactly do you mean
The configure command line uses to differ from one OS to another. For a
given OS, we need to define one "default" configure command line, saying
how to build the Official Apache OpenOffice.org, and reproduce, as
precisely as possible the same build (never possible, I know).
This is the sense of "optimized" I have in mind.
- optimize build dependencies for every OS
- find and welcome newcomers, and builders on every OS
should be an ongoing effort
I think we have several volunteers around, indeed :-)
As volunteer, I remember I stopped to commit any cws and to contribute
directly to OpenOffice.org, because of excessive, stupid and boring QA.
well that's your personal view
Yes, it is.
that i can't share and that we hopefully will not follow here.
No problem.
At the end of Oracle OpenOffice.org time, I remember 20% max of the my
devel time was learn the bug or the feature, write code and commit it.
And 80% of it was QA stuff,or whatever I didn't care, like click the
right blocking checkbox / button on EIS, being blocked, or redo several
times the same fail with bots, because the bot was broken, or being
blocked by other OS, not concerned by the cws itself :-) , or something
similar.
I think we should trash that, but I can perfectly understand people
disagree my point of view :-)
well i haven't had in mind what we had before. And i think you mix some
things here. QA is indeed often more work than specifying and
implementing a new feature. With QA i mean the development of tests (in
case of unit tests) as well as the final execution of these test or even
manual tests.
But we should combine these things with a broken or not optimal
framework to track QA efforts or so. You talk mainly about your
frustration with the provided tooling. I can share your view here but
then we should start to make it better or different than in the past.
It's up to us to define and build something new, that works better and
is satisfying for all. Nobody needs frustration, we want to have fun ;-)
Juergen
Regards,
Eric
