Re: Plotting with Octave 3.0

[John W. Eaton]

On 26-Sep-2008, Matthias Brennwald wrote:

| However, apart from knowing how to work with diff, texinfo and  
| mercurial, users als need to INSTALL the stuff and make it work. This  
| is propably not an issue for Linux advocates with long-term  
| experience in hacking around in their Linux boxes, but there are  
| other, "lesser" OSes (and users) which don't have diff, texinfo or  
| mercurial installed by default and installing is not always  
| straightforward.
| 
| While I for myself think that I _might_ be able to manage all this  
| (because I somehow like to poke around with this stuff), I am sure  
| that all my colleagues working with Octave wouldn't want to waste  
| their time with messing around all this, even if they'd like to  
| contribute to Octave. So, all in all, I still think contributing to  
| Octave should be easier in order to convince more users to contribute  
| their work.

You can contribute changes in the following ways (ranked most to least
helpful):

  * Send a complete Mercurial changeset.

  * Send a context diff relative to the sources from the current
    Mercurial archive (I imagine you might do this if you have diff
    and have downloaded a file from the Mercurial archive on the web,
    but you don't have Mercurial installed locally).

  * Send a context diff relative to some released version of Octave.

  * Send a report with a suggested change, but not in the form of a
    diff.

  * Send a report about a problem with no suggested change, and hope
    that someone else will be motivated to make the change for you.

I'm much more likely to install changes that arrive as complete
Mercurial changesets.  If all you can send are suggestions then that
is OK too, but I'm unlikely to apply them until someone generates a
changeset that I can review and apply.

In all cases, changes should include an explanation of why the change
is needed, and if it is code that is changing to fix a bug, the bug
report should include a self-contained test case that we can run to
reproduce the problem.

jwe