On 23.11.2012 18:17, Данило Шеган wrote:
> I assume the missing bit is to create artifact grants for all the
> subscribers that might not have them now? FWIW, Linaro can work around
> that once this hits production by unsubscribing and subscribing people
> again (I assume you don't have too many private blueprint users, so it
> might not be worth the effort to come up with migration plan).
Yes, this is an issue, and I'd appreciate your offer to do the
unsubscribe/subscribe cycle.
But my mistake is something else: Even people with an artfact grant
can't yet access blueprint pages. They'll get a 403 error when
product.getSpecifcation() is called during traversal. Nice example of
"too narrow" testing: create_inititalized_view(proprietary_blueprint)
works fine -- but this does not involve traversing to the blueprint...
> Will this also include assignees who are not explicitely granted
> permissions to view a blueprint?
>
No, I'm not addressing this right now. I see your point: Having to
explicitly subscribe an assignee is somwhat weird.
But:
- we have the at present same problem with bugtask assignees, so for
consistency sake both "assignment steps" should behave identically
- we will have to discuss a few issues around the details. The most
important one: Should we simply subscribe the assignee
automatically, or should we issue the artifact grant independently?
On 23.11.2012 18:17, Данило Шеган wrote:
> I assume the missing bit is to create artifact grants for all the
> subscribers that might not have them now? FWIW, Linaro can work around
> that once this hits production by unsubscribing and subscribing people
> again (I assume you don't have too many private blueprint users, so it
> might not be worth the effort to come up with migration plan).
Yes, this is an issue, and I'd appreciate your offer to do the subscribe cycle.
unsubscribe/
But my mistake is something else: Even people with an artfact grant getSpecifcation () is called during traversal. Nice example of inititalized_ view(proprietar y_blueprint)
can't yet access blueprint pages. They'll get a 403 error when
product.
"too narrow" testing: create_
works fine -- but this does not involve traversing to the blueprint...
> Will this also include assignees who are not explicitely granted
> permissions to view a blueprint?
>
No, I'm not addressing this right now. I see your point: Having to
explicitly subscribe an assignee is somwhat weird.
But:
- we have the at present same problem with bugtask assignees, so for
consistency sake both "assignment steps" should behave identically
- we will have to discuss a few issues around the details. The most
important one: Should we simply subscribe the assignee
automatically, or should we issue the artifact grant independently?