Oops accessing bug without a default bugtask

[Diogo Matsubara]

As seen in OOPS-1409C2606 a NoCanonicalUrl: No url for None because None broke the chain. was raised due to the fact the the default_bugtask for the given bug is None

This is similar to bug 403131 and may be caused by bug 403182

This bug report is about fixing the bad data since the root cause for this was fixed in bug 403182

Steps to reproduce:
1. Open https://launchpad.net/bug/268895
2. OOPS-1409C2606

[Abel Deuring (adeuring)]

I don't think that OOPS-1409C2606 is related to one of the bugs 403131 or 403182.

1. This OOPS occurred on the main LP bug page, https://launchpad.net/bugs, while the OOPSes from the other bugs occurred on bug pages.
2. I tried to reproduced this locally by manually creating bugs without associated bug tasks. Accessing https://launchpad.dev/bugs worked fine; the new bugs appeared without any hitch on this page (links look like https://bugs.launchpad.dev/bugs/16), while accessing them leads to an OOPS like that described in bugs 403131 and 403182.

[Diogo Matsubara (matsubara)]

This bug report is not about accessing https://launchpad.net/bugs, it's about accessing the canonical URL for a bug report (i.e. https://launchpad.net/bugs/268895) and the redirection code OOPSing due to the fact that the bug.default_bugtask is None. I think those OOPSes are similar because of the fact that the bug.default_bugtask attribute is None while it should return a bugtask object.

As Bjorn explained on bug 403131, the error occurred because LP allowed bugs to be created without a bugtask. Since that was fixed already, I think this one is about cleaning up the bad data in the DB so bug 268895 will have a default bug task.

Does that make sense?

[Abel Deuring (adeuring)]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 10.11.2009 20:11, Diogo Matsubara wrote:
> This bug report is not about accessing https://launchpad.net/bugs, it's
> about accessing the canonical URL for a bug report (i.e.
> https://launchpad.net/bugs/268895) and the redirection code OOPSing due

Ouch, yes, you are right. I only looked at the URL shown on the side of
the OOPS page, but did not notice the PATH_INFO request variable.

> to the fact that the bug.default_bugtask is None. I think those OOPSes
> are similar because of the fact that the bug.default_bugtask attribute
> is None while it should return a bugtask object.
>
> As Bjorn explained on bug 403131, the error occurred because LP allowed
> bugs to be created without a bugtask. Since that was fixed already, I
> think this one is about cleaning up the bad data in the DB so bug 268895
> will have a default bug task.
>
> Does that make sense?

Yes, sure. Thanks to mbarnett, I already have the set of bugs without
bugtasks (ca 30), and 268895 is indeed one of them.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFK+cP0ekBPhm8NrtARAqdRAJsGbCGVK0IiPnXS1ySafMNdT6klWwCfeq6G
zJbLOhB6xnsCUKBHUD0XnXg=
=dpDv
-----END PGP SIGNATURE-----

[Abel Deuring (adeuring)]

I think we can safely delete the broken bug reports. We have 28 bug reports without any bugtask:

2488
4111
 227978
 229619
 234151
 234152
 264296
 268888
 268889
 268890
 268893
 268894
 268895
 268896
 314252
 340360
 340364
 343056
 375262
 375263
 385579
 385587
 394535
 396358
 402275
 402280
 402289
 402640

All of them are meanwhile a bit older (bug is 402641 was filed 2009-07-21), nobody could really use the reports due to the canonical_url bug, so I see no point to keep them.