rabbah commented on issue #2473: Discussion: Should the catalog repository use a pinned version? URL: https://github.com/apache/incubator-openwhisk/issues/2473#issuecomment-314471186 I think the catalog should be installed like the alarm, cloudant, and Kafka feeds: from their own repos. I don't install the catalog from this repo because for day to day developelent it's not needed. And when I do need to, I don't run post deploy/catalog playbook and instead run out of the catalg repo (which then allows for controlled deployment like you're suggesting). I would package a post deploy playbook perhaps in an extras repository, and do the pinning there. My hestiation to pinning the catalog from this repo is we end up with lots of "bump" commits. In general we are striving to have as few circular dependencies in this repo to dependent repos. So I would sever dependence on the catalog. With the exception of the swift tests, I think all tests in this repo are self contained. ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected]
With regards, Apache Git Services
