Every year ... there's always a contest
to see who gets to be last to request a respin. I really thought webtools
had won for once.
Cedric, you are a Planning Council Member
... so, that's good, that's one requirement. Can you confirm you've discussed
with Modeling PMC and they agree this should incur the cost of a respin?
Don't know why they wouldn't ... but, that is the steps of the Exception
process .... we want to make sure they buy in.
If they agree, I suggest we get the
EPP packages "in hand" and then respin the repo later in the
day, on Friday, since (if I am reading right) it won't effect the packages,
and then we have those done. There is, after all, a chance a respin of
the repository will not necessarily work (if someone else changed something
to break it).
Thanks for the quick testing,
Cédric Brun <cedric.brun@xxxxxxx>
02/18/2011 05:58 AM
Helios SR2 maintenance staging repository is done
It does not affect the building of the package as Acceleo
is not included in it, it does affect though the ability to install the
3.0.2 version from the package once it is built as we are referencing the
Le 18/02/2011 11:45, Markus Knauer a écrit :
Does this affect any of the packages? In particular the
Eclipse Modeling Tools package?
We just found a serious problem, with the acceleo contribution. The version
which is included is not the 3.0.2 as expected but the 3.1, i.e. the Indigo
We have fixed the problem by specifing a maximum range value, see : https://bugs.eclipse.org/bugs/show_bug.cgi?id=337544
We think that the severity of this issue justifies a re-spin of SR2, and
so request one.
Le 18/02/2011 07:58, David M Williams a écrit :
That last aggregation rebuild is now
complete, and has been promoted to /releases/maintenance. Thanks all for
your patience and especially your contributions to improving the quality
of this service release.
Some time next week I'll promote to /releases/helios to allow mirroring
to begin, and then "make visible" on Friday, 2/25, 9:00 AM (Eastern)
... assuming no serious issues require other actions.
I have disabled the helios.runAggregator job in Hudson which means 1) any
rebuilds will need to be explicitly requested (and approved, etc.) and
2) if any of you move your SR2 repositories to a more "released"
location next week, you can safely update your contribution files to have
an accurate final URL, without fear of triggering a rebuild.
Do use care, though, since we want to the ability to do a rebuild on a
moments notice ... just in case something seriously wrong is discovered
in the next few days.
I have also tagged the CVS contribution files with "Helios_SR2"
so serve as a reference, for those that provide specific features versions
and URLs in their contribution files (as we all should, IMHO ... but I
know others have different opinions and processes). The tag is there, if
you find it handy.
Thus begins quite week. It is a good time for continued testing, since
even if you can not fix a bug for the literal release, it is always best
to find serious issues before hand, so you can prepare update sites or
workarounds for your users before they find the issues.
We used to say to update to final locations on the "day before"
release, but we have found, in the past few maintenance releases, this
can cause a bit of a bandwidth bottleneck that last day (we've gotten too
big for one day of mirroring ... apparently) so as long as you have the
ability to keep your stuff "invisible" feel free to move your
big artifacts to their final locations beginning Monday or so, so allow
several days of mirroring before we make things visible Friday morning.
As always ... questions welcome. I'll keep you updated as things change