Chris, Ahah, I didn't realise that of course Jekyll was also pulling the new Javadoc files into the generated site. That makes much more sense now.
That also explains why the .gitignore entry is there as well. But why the "git add" in the script to generate the Javadocs, what is that trying to catch? Since I've hit a few things in the commit/documentation over the past few days I'll update the contribute page to help the next person not look like such a clown. :) What are we doing for documentation changes, full Jira/RB review process or a subset? Thanks, Garry -----Original Message----- From: Chris Riccomini [mailto:[email protected]] Sent: 27 February 2014 23:47 To: [email protected] Subject: Re: Javadoc generation Hey Garry, Sadly, this isn't documented. :( Every time we run publish-site.sh, it also updates the javadocs. It's a manual process. The script you see just updated your javadocs locally. The publish-site.sh script is the one that actually does the publication. There's been some chatter about setting up a Hudson CI to do this every time a commit is executed, but no one has done it yet. Cheers, Chris On 2/27/14 12:42 PM, "Garry Turkington" <[email protected]> wrote: >So doubtless the cause of this is that I'm using the wrong version of >*another* build tool but let's see. :) > >I can't work out how the Javadocs are being kept up to date. In >.gitignore in the base dir of the project we have: > >devel@vm17:~/incubator-samza$ tail -1 .gitignore >docs/learn/documentation/0.7.0/api/javadocs > >Then in generate-javadocs.sh we set that directory as the destination >dir: > >JAVADOC_DIR=$BASE_DIR/docs/learn/documentation/$VERSION/api/javadocs > >In that script we remove the old files, generate new ones then do a git >add to add them to the set to be committed: > >./gradlew javadoc >rm -rf $JAVADOC_DIR >mkdir -p $JAVADOC_DIR >cp -r $BASE_DIR/samza-api/build/docs/javadoc/* $JAVADOC_DIR git add >$JAVADOC_DIR > >When I run the script I get the message that nothing was added to the >set due to the .gitignore entry. > >So how then are the Javadocs being kept up to date, the other script >to publish the site doesn't include the Javadoc dir when pushing to svn? > >As an aside why do we have the split of Javadoc in one place and >conceptually updated by one mechanism, the rest of the website content >through another and svn? > >Thanks >Garry ----- No virus found in this message. Checked by AVG - www.avg.com Version: 2014.0.4259 / Virus Database: 3705/7127 - Release Date: 02/26/14
