Merge lp://staging/~abentley/launchpad/fix-592319 into lp://staging/launchpad
Status: | Merged | ||||
---|---|---|---|---|---|
Approved by: | Māris Fogels | ||||
Approved revision: | no longer in the source branch. | ||||
Merged at revision: | 11012 | ||||
Proposed branch: | lp://staging/~abentley/launchpad/fix-592319 | ||||
Merge into: | lp://staging/launchpad | ||||
Diff against target: |
71 lines (+16/-7) 2 files modified
lib/lp/soyuz/tests/test_archive.py (+1/-1) lib/lp/testing/factory.py (+15/-6) |
||||
To merge this branch: | bzr merge lp://staging/~abentley/launchpad/fix-592319 | ||||
Related bugs: |
|
Reviewer | Review Type | Date Requested | Status |
---|---|---|---|
Māris Fogels (community) | Approve | ||
Review via email: mp+27552@code.staging.launchpad.net |
Commit message
Use a random starting integer instead of a UUID.
Description of the change
= Summary =
Fix bug #592319: test_match_
== Proposed fix ==
Append an integer, rather than a UUID, to the base of generic strings. This
allows us to retain our assertions that generic strings do not contain specific
non-numeric strings. However, while the integer will increment normally,
the starting point of the integer is random, so that multiple runs can be done
against the same database without conflicts.
== Pre-implementation notes ==
Discussed parameters of a solution with BjornT
== Implementation details ==
Used thread-local storage so that this approach is thread-safe. Each thread
will use its own random starting point.
Used a range between 0 and 1,000,000 for the starting point, because using
maxint/2 exceeded maximum database values.
== Tests ==
bin/test -v
== Demo and Q/A ==
None. This is a fix for a transient failure, so you can never be sure that it
*would* have failed.
= Launchpad lint =
Checking for conflicts. and issues in doctests and templates.
Running jslint, xmllint, pyflakes, and pylint.
Using normal rules.
Linting changed files:
lib/lp/
Hi Aaron,
This looks good. Brilliant use of Python builtins. r=mars
Maris