Torsten Curdt wrote:
...IMO status.xml and the svn logs serve different purpose.seriously? ...are the commit message much different from what is in the status.xml? I don't think so.
...I am always a big fan of as little maintenance as possible :) OfWhy not just do it the other way around? Formulate the status.xml
course it *does* require some discipline on the commit messages.
entry such that it can be pasted as is into the commit message.
...well, I think that's how most people do it
right now - but that's not the same. You still
need to make sure not to forget to change the
status xml. It's just an additional step that.
My personal approach regarding this is either to use the same text for commit and status.xml or (more often) to be more verbose in status.xml. Remember also that we can do some layout in status.xml (lists, code, hrefs) that cannot be done in commit messages, this making it a usable documentation source.
Or should we write some xml in commit messages to feed status.xml? Hmm... doesn't smell good ;-)
If input from a non-committer was used, mention his/her name (as is
already common practice today). I think it's futile to try anything
more elaborate to have a fair appreciation of the importance of each
individual contribution.
Yup.
So in the end, my feeling is that the full list of contributors should be the aggregation of the current author tags and the "due-to" attributes in status.xml, with an automatic filter to ensure uniqueness.
WDYT?
Sylvain
-- Sylvain Wallez Anyware Technologies http://apache.org/~sylvain http://anyware-tech.com Apache Software Foundation Member Research & Technology Director
