On 20.06.2013 15:35, A. Rothman wrote:
Hi,

I finished fixing all the major bugs I've come across in the past couple of months. Code review and some more testing by others would be greatly appreciated - I'm sure there's room for improvement, but that's hard to do without feedback.

A few things remain on my todo list, none of them blockers (mostly cleanup):

1. https://issues.apache.org/jira/browse/DOSGI-108 - awaiting verdict (cschneider/dkulp?)
I commented in the issue. As a summary I would prefer keeping the comma separated config for now.

2. Upgrading dependencies (and specifically https://issues.apache.org/jira/browse/DOSGI-191) - what is the policy on this? When do all the dependencies get upgraded?

Basically the dependency upgrades should match the kind of version we release. For a bugfix release there should only be bugfix upgrades for a minor release bugfix or minor upgrades... So I propose we upgrade to 3.3.2. We can also upgrade to 3.4.0 but as the release is planned quite soon I prefer the smaller upgrade. If you want to upgrade dependencies create an issue for it and if it may have big impact it makes sense to discuss on the list.

I created https://issues.apache.org/jira/browse/DOSGI-192 to track this.

3. Refactoring to create utils module (still awaiting response to the mail from a few days ago)
I would prefer to wait with this till the next release.

4.  Split zookeeper discovery into listening/publishing subpackages
Looks like a smaller change so I agree we can do this for 1.5.0. Can you create an issue and do the change if you have time?

5. Standardizing variable names to make the code more consistent and readable
This sounds like a bigger change. At least it may affect a lot of code. Can you propose a new naming scheme on the list?

I can probably do them all in time, but would like to hear feedback if anyone is for/against them or has better ideas before I proceed.

Finally, among the 34 unresolved issues, there seem to be a bunch related to distribution/deployment, possibly several are no longer relevant (e.g. single-bundle stuff). If anyone can sift through them and close the low hanging fruit (at least) that would be nice.
I regularly go through the issues and ask for feedback. I then close them if there is no feedback after some time. Of course anyone feel free to do the same.

Christian

--
Christian Schneider
http://www.liquid-reality.de

Open Source Architect
http://www.talend.com

Reply via email to