On Fri, Mar 18, 2011 at 9:39 AM, Tim Penhey <email address hidden> wrote:
> On Fri, 18 Mar 2011 08:17:54 Robert Collins wrote:
>> On Thu, Mar 17, 2011 at 4:56 PM, Ian Booth <email address hidden> wrote:
>> > Review: Approve *code
>> > This is great, especially the no sql bit, and it fits in with other work
>> > already done to optimise and improve the security around retrieving bug
>> > tasks.
>>
>> > I'm wondering how the order of looking for contexts was determined in
> getLinkedBugTasks():
>> its arbitrary and the order doesn't matter - see the check constraints
>> on bugtask.
>
> It isn't entirely arbitrary. Blueprints use the associated productseries or
> distroseries to indicate goals. If a particular bug had a bug task for the
> goal series, it made more sense to return that one than to return the more
> generic distro or product task.
Bugtasks are constrainted to have only one target: one of (product
series, product, etc etc). So the code that was changed which is
examining only one task, could be in any order and have the same
outcome.
On Fri, Mar 18, 2011 at 9:39 AM, Tim Penhey <email address hidden> wrote: ks():
> On Fri, 18 Mar 2011 08:17:54 Robert Collins wrote:
>> On Thu, Mar 17, 2011 at 4:56 PM, Ian Booth <email address hidden> wrote:
>> > Review: Approve *code
>> > This is great, especially the no sql bit, and it fits in with other work
>> > already done to optimise and improve the security around retrieving bug
>> > tasks.
>>
>> > I'm wondering how the order of looking for contexts was determined in
> getLinkedBugTas
>> its arbitrary and the order doesn't matter - see the check constraints
>> on bugtask.
>
> It isn't entirely arbitrary. Blueprints use the associated productseries or
> distroseries to indicate goals. If a particular bug had a bug task for the
> goal series, it made more sense to return that one than to return the more
> generic distro or product task.
Bugtasks are constrainted to have only one target: one of (product
series, product, etc etc). So the code that was changed which is
examining only one task, could be in any order and have the same
outcome.