Add a github pages site for streamsx.inet about streamsx.inet

[hildrum]

I created such a page for my fork of this repository here:
http://hildrum.github.io/streamsx.inet/

Are you thinking of something different than this?

[ddebrunner]

Looks like you are well on the way to doing this, looks good.

Will it be automatically updated when new code is committed, e.g. a new operator will make the javadoc appear automatically?

[hildrum]

That page is for my fork of the inet repository, not for this one. I don't have permission to create a web page for this repository the same way I did for my fork, because I need to use "Settings", which I don't see here.

I'm looking into making the updates appear automatically; I can add a script to generate the new content on a commit, but getting it pushed automatically at the right time seems harder.

[hildrum]

Initial page created:
http://ibmstreams.github.io/streamsx.inet/

[ddebrunner]

That looks great!

[hildrum]

Closing this issue--I opened IBMStreams/administration#6 for figuring out a process to make updates easier.

[ddebrunner]

Where does the script live that generates the SPLDOC, I can't see it in the repository?

I'd like to chnage the SPLDOC to create composite images.

[hildrum]

There's no script--I just ran spl-make-doc separately for the toolkit and the sample and then did the github branches dance to get them pushed up. I'm mulling around in my mind how to do this smoothly via a script.

[ddebrunner]

It would be good to have a script to run spldoc to ensure consistent output, e.g. the title. I'm hoping to contribute the Ant tasks to run spl-make-doc, sc etc that old be used.

So if we got to a point where it was run a script to update docs & instructions on how too push to pages, it would be good.

[hildrum]

We definitely want a script, I'm just thinking about what I want it to do. I'll probably write one by Monday.

To give you an idea of what I'm thinking about, we have to

• make your docs in the master branch
• save your doc stuff somewhere (it doesn't get checked in to the master)
• switch to the gh-pages branch
• copy the saved doc stuff into the gh-pages branch (gh-pages branch doesn't have the code)
• commit to the gh-pages branch
• push to the gh-pages branch (?) -- but this requires a password (The question mark is because you might not want to push the doc stuff before you've pushed the relevant code changes, so it's not clear whether you'd want this as an automated script.)
• make sure you end up back in the master branch so the user isn't confused.

I can see a couple of different options--one is to save the docs to /tmp and just do all the work in the same repository clone using git checkout to switch branches. But this can be a problem. Another option is to have a special gh-pages clone--but do we create one in /tmp just for these updates, or do or assume the user already has one, perhaps passed in a command-line argument?