Launchpad reviews (was Re: Patch Pilot report)
Aaron Bentley
aaron at aaronbentley.com
Tue Nov 24 02:35:30 GMT 2009
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Martin Pool wrote:
> Let me add a few more things:
>
> The criticism is intended to be constructive; the page is already
> quite useful but it has the germ of an amazing system.
>
> The "wants review from/claim review" system is interesting, but not
> fully realized. It should be clear from the list "needs review from
> me in particular"
'Reviews I have to do'
> "needs review from someone on the team",
'Requested reviews I can do'
> "doesn't
> need more reviewers".
'Other reviews I am not actively reviewing'
It's not clear to me whether you're critiquing our terminology or didn't
understand this.
> When someone works on an mp the system should be asking them "what's
> the next action?" much more clearly than at present.
It sounds like you're suggesting a wizard-style approach. That's quite
different from most areas of Launchpad. For example, updating or
commenting on a bug report doesn't propose further actions.
> * I will do the changes and land it
I think that would be best expressed by creating your own branch and
creating a new merge proposal that supersedes this one. The 'supersede'
part isn't something we currently support.
> * I want someone to do the changes and land it, but not me
> * I want the submitter to do the changes and land it (not sure what
> status this is)
"needs-fixing"
> * I want the submitter to do the changes and resubmit it (I guess
> this is 'needs fixing'?)
It is "resubmit".
> * I need the submitter to do something other than to the code, eg to
> sign the contributor agreement or confirm that they tested the
> performance; nobody can progress this until then
Perhaps a review of "needs info", and resetting the status to "work in
progress". Certainly, confirming that they tested the performance is a
case of "needs info".
> * I need to come back and finish reviewing it
Claim review. Optionally, add a comment.
> * We need to discuss and agree on whether this is even a good idea
Sounds like a special form of "request review".
> These are kind of captured, but not optimally clearly or easily, and
> they're not clearly displayed on the overview page.
The code review system is not intended to enforce a particular workflow;
it's meant to provide enough flexibility to enable teams to use it as
they see fit.
Aaron
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
iEYEARECAAYFAksLRm8ACgkQ0F+nu1YWqI3pSQCfVcT74r4uIGzOZ/mauaLfLVVY
xgEAnArKFsCBTWN8AmgwVJYNSA/FGbLE
=xWOW
-----END PGP SIGNATURE-----
More information about the bazaar
mailing list