Changes for 1.15.1?

[Jonathan Lange]

On Wed, Jun 3, 2009 at 11:55 PM, John Arbash Meinel
<john at arbash-meinel.com> wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Andrew Bennetts wrote:
>> Jonathan Lange wrote:
>> [...]
>>> Woot! Thanks Andrew.
>>>
>>> At the UDS sprint, there was also talk of providing bbc stacking
>>> support in a theoretical 1.15.1. Is this still on the cards?
>>
>> Sadly, I don't think so.
>>
>> The revisions to cherrypick for that would be 4392 and 4396.  They're pretty
>> huge.  Here's the diffstats:
>>
>> 4392: 20 files changed, 962 insertions(+), 190 deletions(-)
>> 4396: 6 files changed, 131 insertions(+), 10 deletions(-)
>>
>> Now, some of 4392 is sort-of orthogonal locking fixes that are mainly for
>> the benefit of tests, but it would be hard to untangle, and still pretty
>> large.
>>
>> Also, those fixes alone still aren't enough to allow creation of stacked
>> dev6 branches via a smart server IIRC, due to bugs in the “require_stacking”
>> logic in bzrdir.py, so the benefit would actually be pretty minimal until
>> even more fixes are made...
>>
>> John, care to summarise where dev6 stacking is at?
>>
>> -Andrew.
>>
>>
>
> It works as long as you create the branches "manually". I'm going to be
> working on fixing the "require_stacking" code today. I was considering
> doing it via a new RPC since it would let us
>
> a) Create the Branch at the same time we create the Repository,
> eliminating yet another round trip.
> b) If it is only going into 1.16 anyway, I don't have to worry about a
> "minimal" patch.
> c) While we could get GC stacking into a 1.15.X if we *need* to, given
> that we want to change the Revision serializer anyway, there isn't
> really a fixed format for 2.0 yet, which is what Launchpad wants in a
> 1.15.x release. And that *certainly* would be beyond the scope of a
> 1.15.x release anyway.
>

Thanks, that's good to know.

Bob, are you still planning on making a 1.15.1 with the bug 381329 fix included?

jml