[Qgis-developer] Failing tests consider blockers

[Régis Haubourg]

Hi All, 
I know I'm not in PSC or or core dev here. 
I feel uneasy about that thread, because both options are right from a
certain point of view. 
We strongly need fixed calendar to organize a fluid release workflow. We
also need releases with as few blockers as possible.
 
So, since QGIS is still extremely fast to get new features - thanks to all
of you - I'm afraid blockers solved after the release date, will let some
room to new blockers if we don't put strong quality rules in dev and tests.
I understand Radim when he says all releases will have big blockers then. 

As a funder, my reaction to this would be to go funding a stable branch with
bugfixing releases. But.. we partly choose QGIS because our fundings were
also profiting to all the small NGO's and local authorities we finance to
improve water management.  Then, I need a free public stable release
(financed upstream by funders), and not a paying service.  I can't agree to
have a good software, and pay for it with public funds, and do not afford
small authorities we finance-  to get the same stable software...

To solve that paradox, I'm trying to convince my whole hierarchy (2 levels
done, 2 remaining!) to go for sponsoring. I hope in this way encourage you
to have more resources for tests and bugfixing. 

For 2.0 , I tried to finance bug fixing, but it is almost impossible in
France to do that. We can't hire someone for days of work, we can only do it
for a clearly predefined work). 

Having planned release dates will also help me have someone do standard real
use cases with prerelease versions once feature freeze is started. It
remains difficult to me since users don't like testing and reporting yet
here, I feel a bit alone doing that. I'm pretty convinced this is a cultural
change and some power users could help, but most users won't contribute.. 
And I'm not sure my hierarchy will follow with such time consuming tasks if
it implies other workers. Again, agreed with Radim when he says funders
already funded things..

By myself, I think a release every 4 month is too much. I'm only in capacity
to get involve in a migration for users once a year (80 users to manage, +
Citrix deploiement tests, and docs..) . (That's a only my situation, others
could prefer a frequent release)

Anyway, keep going on a stable calendar, and we'll do our best testing and
funding. Valmeria will be an excellent release. 
Cheers!
Régis



--
View this message in context: http://osgeo-org.1560.x6.nabble.com/Failing-tests-consider-blockers-tp5103837p5104799.html
Sent from the Quantum GIS - Developer mailing list archive at Nabble.com.