Hi Sean, I agree with this approach. I dont have a lot of ideas on making a website, if there are some pointers I can may be do some transfer. Also, the ANT version on our website is old and upgrading is not working due to some dependency issues.
Suraj -Suraj Acharya On Sat, Apr 21, 2018 at 9:52 AM, Roy Lenferink <[email protected]> wrote: > Hi Sean, > > I have to say, it took me some time as well to find the sources of the > site. > > Because the site is still relying on Hadoop, it is either getting the > current site fixed or starting over with the current best-practice for > hosting a website, which is using the gitpubsub approach. > > As Avro is already using gitbox, I would indeed suggest to request a new > git repository for the site [1]. When Avro is at a point of having > something useful in the repository, infra can simply switch from the > current svn approach to gitpubsub. > In a lot of (new) Apache projects this approach, together with jenkins, is > already used. A member of the pmc-chairs group can grant users the > sufficient permissions for Jenkins [2]. > > Using a tool as jekyll [3] would be perfect if you want to update the site > using markdown. > > Roy > > [1] https://gitbox.apache.org/setup/newrepo.html > [2] https://cwiki.apache.org/confluence/display/INFRA/Jenkins > [3] https://jekyllrb.com/ > > 2018-04-21 16:53 GMT+01:00 Sean Busbey <[email protected]>: > > > Hi Folks! > > > > Today I tried to work through some helpful contributions to our > > website[1], and I ran smack into some pretty severe tech debt for our web > > presence. > > > > For starters, none of our documentation talks about how to build / deploy > > or contribute to the website. That's pretty bad. What's worse, our "How > To > > Contribute" guide is on the obsolete ASF wiki and points folks to SVN > from > > when we were a subproject of Apache Hadoop (!!!). > > > > Eventually, I chased down that our website stuff is run out of this old > > svn repo: > > > > https://svn.apache.org/repos/asf/avro/site > > > > 1. edits go into "author" > > 2. the website is built using ant and Apache Forrest > > 3. results will be in "publish" > > 4. both edits and their results are committed at once to svn > > 5 ... > > 6. website is updated > > > > What's worse, our website still relies on an old version of the Hadoop > > website infra. author/skins is an svn external reference to their website > > as it was in 2014. This is why our footer materials all refer to Hadoop > and > > not to Avro. > > > > Before any of us start digging into fixing some of this stuff, what do we > > want from our website? > > > > Humbly I suggest we just eat the up-front investment in switching to the > > current asf infra best-practice. I think that's using gitpubsub to > publish > > from a stand-alone git repo. In atleast one other project I'm on, the > > source material for the website lives in the main repo and a jenkins job > > automatically publishes updates. > > > > I'd love us to also switch to markdown, asciidoc, or anything else that > > isn't editing xml. > > > > > > [1]: AVRO-2170, AVRO-2171 > > >
