Discuss web page instructions

[hildrum]

This is an issue for discussing the wiki page: Creating a web page for your toolkit

[ddebrunner]

I tried to use --createTempRespository without --push and it said I must do a push.

I was hoping to run this without a push so I can manually check the resulting pages before pushing to the real-site.

So I tried the non-temp repository, but it's not clear what repository path I'm meant to use, since I haven't seen any evidence of gh-pages in my repository clone yet.

E.g. with these steps, what path do I need to provide as the argument to updateWeb?

git clone https://github.com/IBMStreams/streamsx.inet.git cd streamsx.inet ./updateWeb.pl _what_do_I_puthere

[hildrum]

The web page is updated. The way I run it is that I have $HOME/IBMStreams/streamsx.inet where I do my work. I didn't change the branch, so it's on the master branch. and $HOME/IBMStreams-web/streamsx.inet (created by a git clone in $HOME/IBMStreams-web). This starts out on the master branch too, but the first time you run the script, it will be set to the gh-pages branch.

From $HOME/IBMStreams/streamx.inet,

./updateWeb.pl ../../IBMStreams-web/streamx.inet

[ddebrunner]

Can the wiki provide an explicit set of commands, rather than just text saying you should do this and that?

E.g. something like this, probably is not correct as I don't know the URL for the web site.

mkdir IBMStreams cd IBMStreams git clone https://github.com/IBMStreams/streamsx.inet.git

cd .. mkdir IBMStreams-web cd IBMStreams-web # NOTE THIS CLONE DID NOT WORK FOR ME, SO WHAT'S THE CORRECT URL? git clone https://github.com/IBMStreams-web/streamsx.inet.git cd .. cd IBMStreams/streamsx.inet ./updateWeb.pl ../../IBMStreams-web/streamsx.inet

[hildrum]

1. I added a --nodeleteTemp option to the script. If specified, you can use --nopush. Your web pages will be in /tmp/streamx.inet.
2. You use the exact same URL as you did to make the first clone. Let's say you've created a directory $HOME/IBMStreams-web.

cd $HOME/IBMStreams-web
git clone https://github.com/IBMStreams/streamsx.inet.git
# now there is a streamsx.inet in IBMStreams-web

Now go to the workspace where you edit code for streamsx.inet, and run the script from there, with updateWeb.pl $HOME/IBMStreams-web/streamx.inet

[ddebrunner]

Is the invocation of system("ant spldoc") meant to commented out?

[ddebrunner]

With the ant spldoc being executed I did:

./updateWeb.pl --createTempRepository --nopush --nodeleteTemp

Which results in /tmp/streamsx.inet being modified, but since the code is already committed, I cannot easy see what changes have been made. E.g. I was expecting to be able to do git diff.

Any idea how I see the changes in the last git commit?

[ddebrunner]

Found this for seeing the changes in the last commit

git diff HEAD^..HEAD

[hildrum]

Ah, didn't know that. Useful. If you don't like what you see, you can delete the whole directory /tmp/streamx.inet .

Also, the script lets you say --nocommit.

[hildrum]

Oops, no, the system("ant spldoc") was not supposed to be commented out--I commented it out for testing and forgot to uncomment it.

[ddebrunner]

I got it to work and pushed changes to the pages repository (branch?). I fixed an issue in build.xml where ant spldoc left the info.xml checked out.

Great work in making it easy!

Since I have a Jenkins server running, should I add it as a post successful build task so that the web-site is updated automatically on any change?

[hildrum]

@ddebrunner I think it'd be great to have the Jenkins server update it automatically on any change!
(FWIW, the gh-pages is a branch in the same repository--it's a branch off the empty repository, which is why it doesn't have any of the same files.)

[ddebrunner]

@hildrum Could you modify the updateWeb.pl script to allow use of the ssh URL to clone the repository, rather than just supporting https?

E.g. the ssh url for streamsx.inet is:

git@github.com:IBMStreams/streamsx.inet.git

This makes it easier to integrate it into a build system, as it can use a SSH key instead of having to enter a user/password.

[hildrum]

Yeah, I can do that. Right now, it's pulling the url out of the repository info, ie, from git remote show origin.

[hildrum]

You can try it now.

[ddebrunner]

@hildrum that worked, thanks very much.

Now my Jenkins CI server is building streamx.inet and updating the SPLDOC on the pages site after any commit to the repository.

[chanskw]

I do not think we have anything more to discuss here... closing.