Code review comment for lp:~cmars/juju-core/resolve-cs-series
Revision history for this message
| Casey Marshall (cmars) wrote | # |
© 2004
Canonical Ltd.
•
Terms of use
•
Data privacy
•
Contact Launchpad Support
•
Blog
•
Careers
•
System status
•
bbfa235
(Get the code!)
PTAL
https:/
File cmd/juju/
https:/
cmd/juju/
strings.
On 2014/04/03 13:34:15, fwereade wrote:
> not quite sure that config is the right package for these -- but I'm
not sure I
> can think of a better one. Unless you want to create a tiny,
hyper-focused,
> series package somewhere? I'm sure it'll grow excitingly as we deal
with other
> OSs...
LP: #1301999
https:/
File cmd/juju/
https:/
cmd/juju/
charm.InferRepo
On 2014/04/03 13:34:15, fwereade wrote:
> We still need ctx.AbsPath, I think. Would you change the relevant test
to use a
> path relative to the working dir so we check this properly please?
Ah, must have botched this in a prior merge. Restored it.
https:/
File environs/
https:/
environs/
On 2014/04/03 13:34:15, fwereade wrote:
> yeah, this feels a bit tacked-on here. Let's give it its own package
if we don't
> think of anywhere better.
LP: #1301999
https:/
environs/
On 2014/04/03 13:34:15, fwereade wrote:
> this is guaranteed to be a string if it's got this far, right?
Non-string values that make their way in here will be ignored with a
warning.
https:/
File provider/
https:/
provider/
EnsureBootstrap
cons.Arch)
On 2014/04/03 13:34:15, fwereade wrote:
> quibble quibble, we should probably allow bootstrap on any series we
can find
> tools for. Out of scope today, but maybe worth a bug?
LP: #1302005
https:/
File provider/
https:/
provider/
On 2014/04/03 13:34:15, fwereade wrote:
> wondering if this should be LatestLts? I presume you have a good
reason it
> isn't, but it's not immediately clear ;)
It's to separate what is used for default-series: in tests from the
preferred series selection. This allows us to test scenarios against
different configured series vs. what will be preferred.
Most of what's affected atm is the tools setup & bootstrapping for some
tests. There are still some areas lurking where its hard-coded to
precise.
I've opened LP: #1302015 to address this.
https:/
File state/api/client.go (right):
https:/
state/api/
...charm.Reference) ([]params.
On 2014/04/03 13:34:15, fwereade wrote:
> I think this client-side method might be cleaner as non-bulk -- I
don't really
> care about bulk calls and compatibility for state/api, just for
state/apiserver.
Done.
https:/
File state/api/
https:/
state/api/
On 2014/04/03 13:34:15, fwereade wrote:
> remind me, does this jsonify to a string?
It does now.
https:/
File state/apiserver
https:/
state/apiserver
config.
On 2014/04/03 13:34:15, fwereade wrote:
> is it rational to move all this block...
Done.
https:/
state/apiserver
On 2014/04/03 13:34:15, fwereade wrote:
> ...into here? and not hit the db for config unless it's really needed?
Done.
https:/
state/apiserver
On 2014/04/03 13:34:15, fwereade wrote:
> I suspect it would be nicer to move this stuff into ResolveCharm to
avoid the
> repeated db hits
Done.
https:/