[Bioperl-l] bioperl-dev or branch?

Robert Buels rmb32 at cornell.edu
Thu May 21 23:55:26 EDT 2009

Mark A. Jensen wrote:
> Also wanted to chime in briefly here--for me as a new developer, commit
> access to The Trunk is a little scary, but bioperl-dev seems friendly,

The point of a version control system is so that mistakes can be traced
and undone.  However, this is no substitute for writing good tests and
running them before you commit.  If your test coverage is good, you have
  a reasonable chance of catching a goof before it's committed, but if
it gets by, relief is just an svn-reverse-merge away.

> so I find I'm more comfortable putting my hare-brained schemes there
> where I know I won't break anything, but experienced folks can monitor,
> comment, ignore, get excited, emit raspberries, etc. the whole time. So I'm
> committing and developing where I might otherwise have shied away. In
> this way bioperl-dev may be an encouragement to the liberal tradition you
> describe.

These things are what a branch is for.  "Hey, svn switch to my branch,
run the tests, does this look sane?", "Could you diff this part to trunk
and tell me what you think about those changes".

But of course changes should not be merged back into the trunk until
they have test coverage and the rest of the test suite in that branch is


More information about the Bioperl-l mailing list