[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
[cross-project-issues-dev] Neon.3 update problem status

Hi,

Cross-posting for better visibility.

Here is the current status of the Neon.3 problem and its potential
solutions:

On 06.04.2017 14:06, MÃlanie Bats wrote:
> During the call I listed all the possible solutions :
> 
> A. Respin against Oxygen Mx Orbit
> 
> B. Respin with the old version of httpclient
> 
> C. Respin with the two versions    of httpclient
> 
> D. Respin with only the new fixed version of httpclient

Solution A (least favorite) and D (possible wiring issues and required
respins of MPC, USS and Oomph) seem to be out of the race.

C was the preferred solution. Jeff did a respin of Linux Tools Docker
Tooling. Unfortunately in the aggregation build the fixed 4.5.2 version
of the HttpClient did not get pulled in (for unknown reasons). Thomas/Ed
found out that there are still wiring issues (even with the backported
Equinox fix). => Failed.

The Architecture council recommended to go for solution B. Jeff
downgraded the Linux Tools Docker Tooling to the old version of
HttpClient. In the next aggregation build HttpClient 4.5.2 and HttpCore
4.4.6 were removed from the repo as expected. This also eliminated the
wiring issue. Jeff fixed another issue with slf4j and ran successful
tests with Committers EPP. Ed successfully tested with ELWMS.
=> Looks promising.

Open questions:
- Do we need to use the uninstall command in p2.inf to remove old Http*
bundles for users that already upgraded to Neon.3 and have a broken
setup at the moment? How would that affect 3rd party plugins that might
bring their own HttpClient version (Martin's question)?

- Are we confident that solution B works for all affected components
(MPC, Oomph, USS, Code Recommenders, others)?

I'd like to run the Neon.3 respin next week, if possible. So please
provide test feedback!!

Latest aggregation repo for Neon.3 respin can be found here:
https://hudson.eclipse.org/simrel/job/simrel.neon.3_respin.runaggregator.BUILD__CLEAN/4/artifact/aggregation/final

Please let me know if you have any concerns and/or correct me if
something is missing or just plain wrong.

Regards,

Fred