Hello all, I wrote this on Slack [on March 3] to get a rough feeling and initial feedback. I got a non-binding +1 and no other comments, so I'll just copy/paste it all here:
- Jim just asked [elsewhere] about the asfbot, and I suggested that the code should likely [be] brought up to current ASF Infrastructure patterns: py3, git, and a pipservice-based daemon. - obviously, STeVe is not part of Infra - but I do believe that we have created a nice pattern for daemons, and a little bit for webapps (still fleshing out a good way to build flask-based apps, which I think is the correct long-term path for $reasons) - we defo want to drop elasticsearch, and switch to sqlite, too - should there be energy, then I would suggest moving to git as a first step. Bring up a website on the asf-site branch (per ASF standard) and use Pelican and .asf.yaml to build that site from a "site" directory (which the svn repo already has, but is empty) - (not a separate repo; there is no need for that; git peeps tend to create a repo when the wind shifts; it makes no sense, from a version control perspective) - we can recover the old site's .md files from svn. It was rm'd last August, but is easy to bring back (I replied to sebb with the revision#, so the hard archaeology is already complete) - I also suggest rm'ing all the obsolete stuff from the repository. We have several generations in there, mostly unused and only serving to distract. Thoughts welcome. I suggest the above as steps that I think we can shift Apache STeVe towards, fix up some basic structure to provide a more inviting community (eg. shifting some items to GitHub to create better interactions). I hadn't thought about it, when I wrote the above, but would also suggest a move to using GitHub Issues rather than Jira. (and yes, this implies a need to handle the existing Jira issues, or to move them). Ideas? Concerns? Cheers, -g
