Thank you. Unfortunately, it looks like Splunk is non-free. Since I
want the Octave devs to check my work or replicate it, they won't like
this. It's also a huge binary download, probably a lot more complex
than I need for this purpose.
I can write short a Perl program to parse the Changelog and output...
what's the format? Also, you guys probably prefer Python. This is a
good opportunity for me to practice my Python to do some simple
parsing.
- Jordi G. H.
> I can write short a Perl program to parse the Changelog and output...
> what's the format?
Oh, sorry, I should have looked around before I looked. The output is
a very simple XML file, nice. Hm, I'll just add the Changelog format
to convert_logs.py, send you guys a diff, and hope you like it.
- Jordi G. H.
> Yes, do so :
Done, see attached. I also managed to produce the following:
http://platinum.linux.pl/~jordi/movies/octave-code-swarm.avi
For some strange reason it died on me around 2004, which is a shame
since Octave development really exploded around 2007. :-(
Btw, the Changelog format is described in the GNU coding standards, here:
http://www.gnu.org/prep/standards/html_node/Style-of-Change-Logs.html
I am sure we can expect that other GNU member projects will adhere to
this format.
> It would be great if you could take the time to validate it with a few
> different changelog form different origins.
I tried it with the Octave Changelogs which go back to 1992 and
changed formatting more than once. Also, the Changelogs are scattered
amongst various directories; for that I manually went to each
directory, ran a quick sed on each Changelog so that all the files in
each subdirectory had that subdirectory as a prefix, catted all the
Changelogs together into one, and then ran convert_logs.py on that.
I'm not sure if this procedure should also be automated by the Python
script, recursing into subdirectories. Maybe.
Anyways, HTH
- Jordi G. H.