[kepler-dev] [Bug 5458] Moving between 2.2 and 2.3 and back requires deleting ~/KeplerData

[bugzilla-daemon at ecoinformatics.org]

http://bugzilla.ecoinformatics.org/show_bug.cgi?id=5458

--- Comment #18 from Derik Barseghian <barseghian at nceas.ucsb.edu> 2011-10-05 18:40:23 PDT ---
I don't like having to edit the configuration.xml file either, but it's
currently the only way that I know to allow us to test things without making
rcs prematurely available to the world. I could change it in the rc, but then
the problem is worse, you initially won't be testing what will actually go out.
Apparently the test-area work was never completed to get around this issue. 
Once you edit the configuration.xml, your last attachment, where the MM GUI
tells you to select a valid suite will be a way to get back to 2.3.

Document how to switch releases for whom? For testers right now? I can do that
(although maybe those few that will test know the situation at this point). I
don't want to make such documentation for users if it turns out it's possible
to patch 2.2 for the issue and we decide it's worthwhile to do so, because then
hopefully there shouldn't be any documentation necessary. Because if you're
keeping different copies of the kepler application around, and we patch 2.2,
the MM GUI will come up as you see it here, telling you there was a problem,
and you should switch to a valid suite.


(In reply to comment #17)
> Right, comment 11 says:
> "If they don't, the other thing that can happen is 2.2 can fail to start from
> the 2.2 icon, and 2.2 will not start its MM GUI -- the user has to figure out
> to do this on their own. "
> 
> I was able to start MM and select the kepler-2.2 suite.  It was not at all
> obvious.
> 
> I don't think we should go back and patch 2.2.
> 
> We should document how to move between the releases and be sure that it can be 
> done.  I can't get back to 2.3 right now.  Deleting ~/.kepler and ~/KeplerData 
> would probably do it.
> 
> 
> About hacking the configuration because the files are only visible in test
> area,
> I'm not so sure if that is such a good test to validate that the
> problem has been fixed.  It means that I'm not running what users will
> be running.
> 
> Sorry to be a stick in the mud, but these failures are mysterious and
> confusing to users.

-- 
Configure bugmail: http://bugzilla.ecoinformatics.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA contact for the bug.