Traceback when the credentials storage returns unicode

The fix is to change the line to this:

credentials.load(StringIO(value.encode("utf8")))

I can't re-create this any more which is rather annoying. I've left the bug here anyway in case it happens again.

On 18 October 2011 23:21, Julian Edwards <email address hidden> wrote:
> The fix is to change the line to this:
>
> credentials.load(StringIO(value.encode("utf8")))

I'm not sure that would be the true fix: utf-8 won't have sequences of \0 bytes.

This looks a lot like a similar bug with garbage being written in to
the gnome wallet, which is also hard to reproduce.

Hello Julian, or anyone else affected,

Accepted python-launchpadlib into oneiric-proposed, the package will build now and be available in a few hours. Please test and give feedback here. See https://wiki.ubuntu.com/Testing/EnableProposed for documentation how to enable and use -proposed. Thank you in advance!

Hello Julian, or anyone else affected,

Accepted python-launchpadlib into natty-proposed, the package will build now and be available in a few hours. Please test and give feedback here. See https://wiki.ubuntu.com/Testing/EnableProposed for documentation how to enable and use -proposed. Thank you in advance!

I was having this problem with oneric and the proposed package fixes it. Thanks to this I can use syncpackage now.

I have tested this with gnome on natty. While I cannot reproduce the exact problem of unicode being returned (a kwallet problem)I have verified that any time the key is corrupt or unrecognized it is deemed invalid and the authorization dance is redone. So corrupt keys are treated as expired, allowing work to continue. That said, unicode should be handled properly as Scott showed under oneiric.

Brad, can you confirm that this is "verified" on natty? Your message seems to suggest a regression was created, though I suspect you are just noting a different bug that was already present.

FYI for SRU verifiers, when an SRU is verified in some, but not all releases, leave both the 'verification-done' and 'verification-needed' tags in place. This will make the bug appear in purple on the SRU report:

http://people.canonical.com/~ubuntu-archive/pending-sru.html

We use this to determine whether or not something is ready to move to -updates... sometimes its ok for a partially verified bug to move forward, sometimes its not, but if both tags aren't present it will just appear in blue and we won't even look at it.

Clint, I did not mean to imply a regression. I stated I have verified on gnome that key storage and retrieval works and that it recovers from a corrupted key as I intended. The bug was originally written against behavior with kwallet but since I don't have kubuntu installed I cannot test that scenario.

Thanks Brad!

This bug was fixed in the package python-launchpadlib - 1.9.7-0ubuntu2.1

---------------
python-launchpadlib (1.9.7-0ubuntu2.1) natty-proposed; urgency=low

  * Add 100_base64_encode_credentials.patch cherrypicked from upstream
    tree to fix issue causing launchpadlib scripts to fail due to problems
    associated with keyring corruption. We encode launchpadlib's entries so
    the keyring won't be confused by unexpected characters.
     - Convert credentials from unicode when retrieved from the keyring.
       (LP: #877374)
     - Use base 64 encoding for credentials stored in the keyring.
       (LP: #745801)
     - Properly handle decoding base 64.
       (LP: #900307)
 -- Bryce Harrington <email address hidden> Fri, 09 Dec 2011 15:03:01 -0800

This bug was fixed in the package python-launchpadlib - 1.9.8-2ubuntu0.1

---------------
python-launchpadlib (1.9.8-2ubuntu0.1) oneiric-proposed; urgency=low

  * Add 100_base64_encode_credentials.patch cherrypicked from upstream
    tree to fix issue causing launchpadlib scripts to fail due to problems
    associated with keyring corruption. We encode launchpadlib's entries so
    the keyring won't be confused by unexpected characters.
     - Convert credentials from unicode when retrieved from the keyring.
       (LP: #877374)
     - Use base 64 encoding for credentials stored in the keyring.
       (LP: #745801)
     - Properly handle decoding base 64.
       (LP: #900307)
 -- Bryce Harrington <email address hidden> Fri, 09 Dec 2011 14:55:16 -0800