Merge lp://staging/~vila/uci-engine/integration-speed-up into lp://staging/uci-engine
Status: | Rejected |
---|---|
Rejected by: | Vincent Ladeuil |
Proposed branch: | lp://staging/~vila/uci-engine/integration-speed-up |
Merge into: | lp://staging/uci-engine |
Prerequisite: | lp://staging/~doanac/uci-engine/integration-speed-up |
Diff against target: |
190 lines (+29/-20) 8 files modified
tests/deployers.py (+7/-13) tests/test_image_builder.py (+5/-1) tests/test_integration.py (+3/-1) tests/test_ppa_assigner.py (+3/-1) tests/test_publisher.py (+3/-1) tests/test_test_runner.py (+4/-1) tests/test_ticket_system.py (+2/-1) tests/test_webui.py (+2/-1) |
To merge this branch: | bzr merge lp://staging/~vila/uci-engine/integration-speed-up |
Related bugs: |
Reviewer | Review Type | Date Requested | Status |
---|---|---|---|
PS Jenkins bot (community) | continuous-integration | Approve | |
Canonical CI Engineering | Pending | ||
Review via email: mp+223065@code.staging.launchpad.net |
Commit message
Teach AmuletDeployment() to skip the deployment when one is already available.
Description of the change
This changes lp:~doanac/uci-engine/integration-speed-up to minimize the scope of the workaround.
Andy, we've been talking past each other I think, let me restart the discussion:
1) We are bitten hard by https:/
IMO, It's the most important issue we face for the current "integration" tests: I thought you had some insight about what was going on hence my question. You don't. Damn. Me neither. Damn damn.
I don't want this issue to get out of our radar hence the comment in setUp() in this proposal. If we kick the can, let's make sure it's still in front of us.
2) We need a workaround in the mean time.
I'm fine with that but I want that workaround to be expressed in the right place: at the test level.
I want (like you), as a dev, to be in charge of the environment where my test run.
I *don't* want to have to tell the test runner about that.
You proposed to use an env var for that, there is no need for a class
attribute to duplicate that information.
bzr diffstat for this proposal is:
1 file changed, 14 insertions(+), 10 deletions(-)
yours was:
8 files changed, 26 insertions(+), 31 deletions(-)
I don't want to carry those additional changes.
Unmerged revisions
- 563. By Vincent Ladeuil
-
Simplify DEPLOYED handling. The only place we care about it is in AmuletDeploymen
t.setUp( ). - 562. By Andy Doan
-
integration tests: allow re-use of a deployed environment
This allows us to deploy once and then run our integration tests
against a deployed environment. This greatly speeds up local testing.This is disabled by default, but specifying DEPLOYED=1 in your
environment allows you take advantage of this. I'm running things
at home with:DEPLOYED=1 JUJU_REPOSITORY
=`pwd`/ tmp/charms PATH=$PATH: `pwd`/branches/ juju-deployer JUJU_DEPLOYER_ DIR=`pwd` /tmp ./tests/ test_webui. py I'm not a huge fan of "DEPLOYED=1" but unittest.main has its own argparse
logic, so we screw that up if we try to accept CLI args. Even if this
idea is terrible, its local to one place so we can easily fix it in
the future w/o touching our other code.This makes testing the webui go from 76s to 4s when using juju-lxc
PASSED: Continuous integration, rev:563 s-jenkins. ubuntu- ci:8080/ job/uci- engine- ci/860/
http://
Executed test runs:
Click here to trigger a rebuild: s-jenkins. ubuntu- ci:8080/ job/uci- engine- ci/860/ rebuild
http://