[RFC] Community driven branch extensions

[DeeJay]

I don't want to detract from Ian's overall point, but I have to mildly
take issue with this -

"you could scratch your itch like this:

1. Fetch the branch using lp:foo
2. Make the change, linking it to the bug.
3. Submit a merge proposal.

The upstream project may have it's own bug tracker and it's own VCS
server somewhere else but that wouldn't be stuff casual contributors
would need to care about. Launchpad would join all the dots by having
good code imports, smart syncing with the upstream bug tracker and have
a neat way of turning merge proposals into the required upstream "patch
bundle".

Right now, the above works really nicely when the upstream is Bazaar..  "

The feedback I want to give, having had experience yesterday of just
that process, is that at the moment the description "works really
nicely" is over-optomistic. Repository formats, and automatic
conversions between them, mean too much attention has to be paid to
the tools rather than the itch one is scratching.

Specifically. lp:qbzr is not yet converted to 2a. I branched that to a
local 'master' on my own machine, then made a 'feature' branch from
that in which I developed the change. When I tried to merge that
change back to my master it failed because the repository formats were
incompatible. Retried, ensuring that the feature branch didn't use the
current default format, and the local merge back was then OK. But then
when I pushed the resulting master branch [which should still have
been derived from a clone of lp:qbzr] back to my account on lp: the
resulting branch for merging ended up as 2a, which was unusable by the
qbzr devs. [I've no idea why that conversion took place - I suspect it
might have been a partial loss of responsiveness at lp?]

So although I 'got there', it took some patient hand-holding from
qbzr-devs (thanks Gary and Alexander) on irc which I think - ideally -
should not have been necessary. I came close to giving up and deciding
that I clearly wasn't smart enough to contribute. I wonder how many
potential contributors projects lose who just give up in silent
frustration? [I don't know either - I'm just guessing there must be
some, and that it would be a shame.]

This doesn't conclude with me going off to work with git or Hg instead
- I got hooked on the style, portability and ethos of bzr 3 years ago
- this is meant as friendly but critical feedback from out in
'userland' where things are not quite as smooth as might be hoped!

(I can provide extracts from .bzr.log for anyone who might want to see
what actually happened.)


Regards -
 DeeJay [aka smartpgx]