Merge lp:~maxb/udd/lockcontention into lp:udd
| Status: | Merged |
|---|---|
| Approved by: | Martin Packman on 2012-05-03 |
| Approved revision: | 588 |
| Merged at revision: | 588 |
| Proposed branch: | lp:~maxb/udd/lockcontention |
| Merge into: | lp:udd |
| Diff against target: |
40 lines (+7/-7) 1 file modified
udd/scripts/import_package.py (+7/-7) |
| To merge this branch: | bzr merge lp:~maxb/udd/lockcontention |
| Related bugs: |
| Reviewer | Review Type | Date Requested | Status |
|---|---|---|---|
| Martin Packman | 2012-05-03 | Approve on 2012-05-03 | |
|
Review via email:
|
|||
Description of the Change
Fixes:
80 packages failed with key bzrlib.
The importer appears to be contending with its own locks
aqsis armagetronad aspell-sv at-spi billard-gl cgiemail cl-md5 fraqtive gcc-4.5 gpe-login gpe-what gtk2-engines-aurora gw-fonts-ttf heartbeat icoutils isight-
| Max Bowsher (maxb) wrote : | # |
Thanks for the review; merged. Even though this code *looks* like a library module, as it's udd.scripts.*, the code there is only used within that single file. As such, the only pressing concern for create_branch is that it does whatever the code in the same file wants it to. Indeed, as you say, the importer only wants these branches with trees, because it needs to import changes into them.

Changes look fine.
Is it expected that this create_branch helper also makes a tree as a side effect?
If not, that step could be done just before the DistributionBranch only.
If so, you may still want to only create a tree `if repo.make_ working_ trees() `, even though I imagine the importer just uses one style.