Duplicity

Merge lp://qastaging/~mterry/duplicity/tarfile into lp://qastaging/duplicity/0.6

tarfile
Merge into 0.6-series

Proposed by Michael Terry on 2011-08-22

Status:

Merged

Merged at revision:

783

Proposed branch:

lp://qastaging/~mterry/duplicity/tarfile

Merge into:

lp://qastaging/duplicity/0.6

Diff against target:

4277 lines (+2257/-1404)

13 files modified

duplicity/diffdir.py (+10/-12)
duplicity/dup_temp.py (+12/-0)
duplicity/gpg.py (+14/-0)
duplicity/patchdir.py (+10/-9)
duplicity/path.py (+8/-8)
duplicity/tarfile.py (+2154/-1271)
duplicity/util.py (+27/-0)
rdiffdir (+0/-1)
tarfile-CHANGES (+3/-0)
tarfile-LICENSE (+1/-1)
testing/diffdirtest.py (+4/-2)
testing/patchdirtest.py (+10/-17)
testing/test_tarfile.py (+4/-83)

To merge this branch:

bzr merge lp://qastaging/~mterry/duplicity/tarfile

Related bugs:

Bug #485219: Fixed fields are not fixed, leading to buffer overflows and tar explosions.	Medium	Fix Released
Bug #690549: uid and guid setting adversely affects integrity	Medium	Fix Released

Link a bug report

Reviewer	Review Type	Date Requested	Status
duplicity-team		2011-08-22	Pending
Review via email: mp+72422@code.qastaging.launchpad.net

Revision history for this message

Michael Terry (mterry) wrote on 2011-08-23:

Due to conversations in the mailing list, I've retooled this branch to, instead of using the system tarfile.py, just updating our internal copy to python2.7's version (with changes for 2.4-compatibility).

Once we drop 2.4 support, it will be a simple change to drop our internal version and use the system version again.

I've run the test suite against this branch (under python 2.4 too), but of course it could use more testing.

Revision history for this message

edso (ed.so) wrote on 2011-08-23:

Does this solve
https://bugs.launchpad.net/duplicity/+bug/690549
?

Or is the patch there invalid, because tarfile.py does not support such big values?

ede

On 23.08.2011 20:36, Michael Terry wrote:
> Due to conversations in the mailing list, I've retooled this branch to, instead of using the system tarfile.py, just updating our internal copy to python2.7's version (with changes for 2.4-compatibility).
>
> Once we drop 2.4 support, it will be a simple change to drop our internal version and use the system version again.
>
> I've run the test suite against this branch (under python 2.4 too), but of course it could use more testing.

Revision history for this message

Michael Terry (mterry) wrote on 2011-08-23:

The patch there looks invalid, because it looks like it just assigns more bits in the tarfile for the gid/uid, which I don't think is part of the standard. Though maybe I'm misinterpreting the patch.

However, I do think this branch would solve that bug, as newer tarfiles seem to have explicit support for negative numbers (see the itn() function in this branch's tarfile.py).