Exporting task_age breaks editing over the API unless you PATCH within a second of GET

[James Westby]

Bug 510985 exported task_age, but this means that it is part of the etag
calculation. This means that the etag changes every second. This means
that if you cross a second boundary between doing the GET on a bug task
and doing a PATCH on it you get a 412. This makes editing bugs with the
API all but impossible.

Thanks,

James

[Eleanor Berger (intellectronica)]

In hindsight, I think adding BugTask.age wasn't such a good idea after all. It's easy to calculate it locally anyway. Let's get rid of it to unblock editing bugs using the API.

[Eleanor Berger (intellectronica)]

This is a candidate for CRITICAL, since it's possible many users of the API are blocked.

[Deryck Hodge (deryck)]

Agreed with Tom, we should get a CP for this branch ASAP.

[Diogo Matsubara (matsubara)]

Fixed in devel r10248 <http://bazaar.launchpad.net/~launchpad-pqm/launchpad/devel/revision/10248>

[Brian Murray (brian-murray)]

It isn't necessary to fix this bug but there is still a task_age property in the bugtask model. It might be a good idea to remove this to prevent it from being exported again in the future.

[Eleanor Berger (intellectronica)]

Brian, good idea! Feel like creating a brunch?

[Graham Binns (gmb)]

On 2 February 2010 17:40, Brian Murray <email address hidden> wrote:
> It isn't necessary to fix this bug but there is still a task_age
> property in the bugtask model.  It might be a good idea to remove this
> to prevent it from being exported again in the future.

Could we maybe export a method to get the task age rather than
importing task_age itself? I know it's a bit of a kludge, but it
solves the use-case for retrieving task age without breaking things.