Attaching a non-existant volume AttributeError

[Matthew D. Wood]

If a user tries to attach a non-existant volume (open two tabs, delete the volume in one, attach in the other) an AttributeError is raised here:

            volumes = self.get_object()
            attachments = [att for att in volumes.attachments if att]

volumes is None.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/78420

[Matthew D. Wood (woodm1979)]

I'm no longer interested in fixing this bug. I'm completely ok with someone culling it. Just not worth the effort to get a small fix through the bug-review process for such a small issue.

[Julie Pichon (jpichon)]

Hey Matthew, sorry you had a crappy experience with our contribution process. The review queue has been out of control for several weeks now and the core team is acutely, painfully aware of it. I think sometimes as well there's a bit too much "blocking" energy in the OpenStack review process, something we can hopefully become better at watching out for and discouraging. The core team is also keeping a keen eye out for people who do helpful reviews to add to the core team and help alleviate the delays. If you have additional ideas on how to improve the experience for contributors, please do share! I'm definitely interested in hearing them. For now there is no way around "everyone, please do more reviews."

[Matthew D. Wood (woodm1979)]

Oh, I certainly don't blame any person or any set of people. :-)

One thing that's happening, I think, is that because the core-review-team
is semi-absent (entirely justified, releases are a pain in the ass) that
leaves other folks to do reviews. And unfortunately, the best way to get
respect in the review community is to -1 people's stuff. It's just sooo
easy to drive-by -1 people's contributions with a "please add more tests"
type statement.

But part of the issue is an inescapable scattered nature of open-source
contributions. There's always one more person who can come along. That's
a good thing usually, but it's frustrating when you almost have a change
through (multiple +1s, maybe even a +2) and someone comes along and
complains that your commit message incorrectly includes the bug number.
 Sigh. So you make the change and then have to wade through the process
again.

I've got a bunch of other fixes up, and many of them have gone through. I
don't intend to abandon the project. But I was done fighting that
particular bug.

--
Matthew Wood

Big Jiu Jitsu Noob
Brand new daughter who's so cute
Soon she'll submit me

On Wed, Apr 9, 2014 at 8:28 AM, Julie Pichon <email address hidden>wrote:

> Hey Matthew, sorry you had a crappy experience with our contribution
> process. The review queue has been out of control for several weeks now
> and the core team is acutely, painfully aware of it. I think sometimes
> as well there's a bit too much "blocking" energy in the OpenStack review
> process, something we can hopefully become better at watching out for
> and discouraging. The core team is also keeping a keen eye out for
> people who do helpful reviews to add to the core team and help alleviate
> the delays. If you have additional ideas on how to improve the
> experience for contributors, please do share! I'm definitely interested
> in hearing them. For now there is no way around "everyone, please do
> more reviews."
>
> --
> You received this bug notification because you are subscribed to the bug
> report.
> https://bugs.launchpad.net/bugs/1288419
>
> Title:
> Attaching a non-existant volume TypeError
>
> Status in OpenStack Dashboard (Horizon):
> New
>
> Bug description:
> If a user tries to attach a non-existant volume (open two tabs, delete
> the volume in one, attach in the other) a TypeError is raised here:
>
> volumes = self.get_object()
> attachments = [att for att in volumes.attachments if att]
>
> volumes is None.
>
> To manage notifications about this bug go to:
> https://bugs.launchpad.net/horizon/+bug/1288419/+subscriptions
>

[Julie Pichon (jpichon)]

Thanks for the reply! I agree there is a perception that one needs to leave -1s to gain respect but it doesn't work as people think it does when it's not accompanied by a meaningful comment. I think reviewers should gently be told to back off if they -1 with "I wouldn't have done this way" types of comments, and it's totally ok for the patch author to reply with "this is why it's done this way" without submitting a new patchset, as long as the explanation sounds somewhat reasonable.

I'm going to mark this bug as a low hanging fruit since it has a suggestion for the solution and is quite small. Hopefully when a newcomer picks it up they'll have a more favourable experience. Thank you for your contributions to the project and to many more!

[Openstack Gerrit (openstack-gerrit)]

Fix proposed to branch: master
Review: https://review.openstack.org/91402

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/91402
Committed: https://git.openstack.org/cgit/openstack/horizon/commit/?id=41a24434d005f5b391de24d9c6ad62dc818a56d7
Submitter: Jenkins
Branch: master

commit 41a24434d005f5b391de24d9c6ad62dc818a56d7
Author: Veronica Musso <email address hidden>
Date: Tue Apr 29 19:33:13 2014 -0300

    AttributeError at attaching a non-existent volume

    When user tries to attach a non-existent volume, AttributeError
    exception is thrown in get_data function, where the right behavior
    is returning [].

    This patch is based in the solution proposed by Matthew D. Wood.

    Change-Id: Ib6c5f83845e0056865a8ace82d9e2f4351d4be3e
    Closes-Bug: #1288419