Hi Lily,
The release note generator reads reports generated by JIRA queries. In
my experience, neither report type follows a stable format. Other Apache
engineers keep improving the JIRA reports and their improvements break
assumptions that are hard-coded into the release note generator. The
tool is supposed to help you build the release notes--feel free to
change it if that will help you get your job done. Note that Kristian
has some ideas about how to get this information through a programmatic
api which may be more stable.
Note also that the detailed release notes themselves (the notes which
are attached to JIRA issues) are written in html. That is because they
are dropped verbatim into the Issues section of the final release page.
The html dialect is broadly understood and it helps contributors format
their detailed descriptions. We could have standardized on some more
obscure markup dialect for the detailed notes but we chose what seemed
to be the simplest format possible. Please do not revisit that decision
without a community discussion.
Hope this helps,
-Rick
Lily Wei wrote:
Hi:
Why the releasenotegenerator read .html instead of .xml and does it
make sense to change back to .xml so we don't have to hard code the
attachment id for the release notes?
Cheers,
Lily
