Project Infrastructure

[John W. Eaton]

At OctConf, a few of us discussed project infrastructure and what we 
could do to consolidate some of it and maybe make things easier to 
manage, especially for the group web pages that must currently be 
uploaded using CVS(!).

I think some people also wanted to know just what resources we are 
using, who can manage them, and so on.  My best attempt to answer that 
question is here:

  http://wiki.octave.org/Project_Infrastructure

Could someone fill in more details about Octave Forge?

Some conclusions from our discussion, as I recall them:

  * no one likes the savannah trackers but moving bug reports to 
another system will be a lot of work

  * we need a better way to handle patches that will allow more 
detailed and transparent code review

  * uploading web pages with CVS is ridiculous and must be fixed

  * it's confusing to have our infrastructure spread over so many 
different systems (I'll try to add some info to the wiki page about why 
things are currently distributed the way they are)different )

  * we probably won't be able to fix all of this at once, but it would 
be nice to at least eliminate the problem of uploading web pages with CVS


So, the one action item that we agreed to start with is

  ==> Move the web pages from gnu.org back to octave.org where we can 
at least manage them however we choose.

I think the master version should remain in Mercurial.

What is the best way to deal with publishing?  I can imagine that it is 
best to manage editing and publishing separately, but maybe that isn't 
really important?  Should the pages be published when they are committed 
to the hg archive at octave.org?  Now that we are using Jekyll, should 
the processing steps be done locally or on the web hosting system?

As a related issue, maybe this would also be a good time to think about 
merging the Octave Forge site with the Octave web site?  Would that help 
users?  Would it be any easier to manage if we did this?

jwe