gnuplot-py-users Mailing List for Gnuplot.py

On Sun, 4 May 2008, (CDT) Mark Alford apparently wrote:
> gnuplot-py version 1.8 is now available on sourceforge 


Thanks!
Alan Isaac

Mark Alford wrote:
> With a lot of hand-holding from Michael Haggerty, I have
> converted the current SVN code in to a sourceforge
> release. gnuplot-py version 1.8 is now available on sourceforge.

Great!  Thanks a lot, Mark!

Michael

With a lot of hand-holding from Michael Haggerty, I have
converted the current SVN code in to a sourceforge
release. gnuplot-py version 1.8 is now available on sourceforge.

Mark

[I'm sending a CC of this email to the gnuplot-py-users mailing list, so
that other people can keep track of what is going on.  Quick summary:
Mark Alford is working on making a 1.8 release of Gnuplot.py.]

Mark Alford wrote:
> Questions:
> 
> 0) In step 1 of RELEASES.txt, when I increment the __version__
> string in __init__.py, I assume that I should NOT have a "+" on
> the end at this stage. That will come later, in step 10, right?

Correct.

> 1) I tried installing happydoc from sourceforge.
> The INSTALL.txt says to type "python ./setup.py install"
> but the happydoc source directory does not have any file called
> "setup.py" !!

I haven't used happydoc in years.  I have no idea whether it is still
maintained etc.  But I don't think there is anything happydoc-specific
in Gnuplot.py.  IIRC, happydoc just parses the Python files looking for
class definitions and docstrings, and generates the documentation from
that.  There might be other documentation packages that can do a similar
job.

I notice that Ubuntu gutsy includes a "python-happydoc" package that is
installed on my computer.  Maybe your Linux distribution has such a thing?

> 2) How do I check my version of gnuplot-py in to sourceforge using
> subversion? Do I cd to the directory where I have my version of
> gnuplot-py and then type
>   svn commit -m "created official version 1.8"

Yes, but this only works if the directory was originally obtained from
sourceforge using SVN.  Probably it also has to be obtained under your
own userid (as opposed to "guest").  Something like

svn co --username=alford \
    https://gnuplot-py.svn.sourceforge.net/svnroot/gnuplot-py/trunk \
    gnuplot-py-trunk
cd gnuplot-py-trunk
# edit files
svn commit -m 'message'

> 3) In RELEASES.txt, why do steps 4,5,6 come after commiting the
> code in step 2? They only affect my local copy.
> How will those changes be communicated to the
> sourceforge release that people will then download?

The point of these steps is to generate a tarball and zip archive
containing the Gnuplot.py module.  Most people prefer to install from a
tarball rather than checking the code out of Subversion (this is pretty
much the whole point of making a release!)  The source code from which
the tarball was created is marked in Subversion by the tag created in
step 3, but of course the tarball and the generated documentation should
not be checked into Subversion because they can be regenerated at will.

The products of step 5 are uploaded to sourceforge in step 7.

> 4) Should I do step 6? Your comment there seems to be saying not
> to do it!

Yes, I think step 6 should be omitted.  You can delete it from the docs
if you like.  (Installing for end users is documented in README.txt anyway.)

By the way, feel free to add yourself to CREDITS.txt.

Michael