How I Patch Piloted lately and some tricks about merge proposals

[Martin Pool]

On 12 February 2010 08:12, Gordon Tyler <gordon at doxxx.net> wrote:
> On 11/02/2010 4:08 PM, Martin Pool wrote:
>>  * Bugs with patches, as I posted about last night: currently about
>> 36.  They are not all the same but many have half-baked patches that
>> need tests added or tests fixed.  There is a lot of value we can pull
>> out of them.
>
> How many of these are languishing because of a lack of a signed agreement?

Some, but it's far down the list of reasons why things are blocked.

As I posted a while ago we (especially Karl) are working with legal
people to improve the agreement/faq.

I want very much to get away from having big piles of stuff blocked
for various reasons, because it tends to inhibit people from looking
at the pile at all.  If it is at all possible to finish something off
and land it, let's do that.  If it just cannot proceed in its current
state, let's get it out of the list.  The main causes for things being
removed other than being landed are in rough order of frequency:

 * the patch is not a fix
 * bug is already fixed and/or a dupe
 * fix is no longer relevant
 * patch is just so wrong that there is no real value in it

Some of these patches were sent under the earlier Bazaar-specific
assignment process and I think it is reasonable to grandfather them
in.

One interesting case is where someone has posted a patch that
reproduces the failure.  Arguably we should turn this into an xfail
test.  However I think that's not worthwhile unless we're then going
to actually fix the bug.

-- 
Martin <http://launchpad.net/~mbp/>