Another advantage: since the geode-site/ directory would not be included in the geode source release, we can move a number of the javascript and font references out of the geode LICENSE.
Anthony > On Feb 16, 2017, at 5:06 PM, Anthony Baker <aba...@pivotal.io> wrote: > > Yes, please. Let’s call the repo geode-site. Use two branches: master and > asf-site. If we can auto-build and push to asf-site that would be awesome. > > Anthony > >> On Feb 16, 2017, at 4:38 PM, Dan Smith <dsm...@pivotal.io> wrote: >> >> +1 >> >> I think the current setup is confusing, because the website is supposed to >> include docs that are generated from the last release, but the site >> instructions say the site should be generated from develop. A separate repo >> with a single branch will probably reduce confusion. >> >> We also need to script the website building and publishing, and ideally >> have the publishing done by a CI system based on commits. It looks like >> some other projects are talking about doing this with jenkins jenkins - see >> INFRA-10722 for example. >> >> -Dan >> >> On Thu, Feb 16, 2017 at 4:10 PM, Karen Miller <kmil...@apache.org> wrote: >> >>> I think that the website content that is currently in geode/geode-site >>> ought to be moved to its own repository. The driving reason for this is >>> that changes to the website occur on a different schedule than code >>> releases. We often want to add a new committer's name or a new >>> event, and these items are not associated with sw releases. A new website >>> release that comes from the develop branch may have commits that >>> should not yet be made public. >>> >>> Are there downsides to separating the website content into its own repo? >>> >
