On Thu, Jul 18, 2013 at 9:21 PM, Yashin Mehaboobe <[email protected]> wrote: > I've installed couchdb on a Linux VPS. I cloned the git repository and > compiled it from source but I haven't created any databases yet. I've looked > through the /share/doc/src rst files as well as the sphinx conf.py file. How > I plan on implementing this is to use the following to gather the changes. > -The rst files from /share/doc > -The git commit messages obtained using git-python
I'd personally go with dulwich or pygit2 for this rather than git-python, since they're implemented as native libraries rather than wrappers around shelling out to other processes. > -Information from JIRA tickets by using the python client for the JIRA REST > API > I'm thinking of implementing this a separate python script that will be > integrated with the Makefile in such a way that executing "make changes" > will execute the script. I don't think you should worry to much about the Makefile or other build system integration. > Once the required data is obtained from the rst files,the JIRA tickets and > the commit logs, it'll be written as an rst file. This intermediary rst file > can then be used by sphinx to create HTML. > > Here's the path I intend to follow: > -Create a python script that collects the data from multiple sources and > compiles it to an intermediary form (rst?) Output to .rst would probably be nice. > -Configure sphinx's conf.py so that it takes the rst file and creates an > HTML output. An alternative would be to use docutils to do the same. > -Modify the Makefile to call the python script whenever "make changes" is > executed. Again, don't worry to much about the integration with build systems or other stuff. IMO the important things to focus on are: - Which changesets from the repository to use - How to extract useful information from the bug tracker - How to combine those sources into something that's easy to read and polish Cheers, Dirkjan
