| Summary: | Fix misc. leaks, improve valgrind foo | ||
|---|---|---|---|
| Product: | Wocky | Reporter: | Will Thompson <will> |
| Component: | General | Assignee: | Sjoerd Simons <sjoerd> |
| Status: | RESOLVED FIXED | QA Contact: | |
| Severity: | normal | ||
| Priority: | medium | Keywords: | patch |
| Version: | git master | ||
| Hardware: | Other | ||
| OS: | All | ||
| URL: | http://git.collabora.co.uk/?p=user/wjt/wocky.git;a=shortlog;h=refs/heads/trivia | ||
| Whiteboard: | |||
| i915 platform: | i915 features: | ||
(In reply to comment #0) > The former is a leak in test_cancel_iq_closing(); seems like at least the in > stream is leaked. I think it's because the call to _close_async() never ends, > and so keeps sched_in alive, and hence the connection. Threw in another patch that makes that call end. Now with a bunch of extra patches adding and using more helper macros in tests:
Move assert_nodes_equal to test-helper
Add test_assert_nodes_not_equal()
Use test_assert_nodes_[not_]equal() in tests
This seems to have been merged. |
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.
I noticed that a couple of places assumed that _finish() would be called, and fixed them; and then I decided to valgrind some tests and stuff. Will Thompson (11): Don't leak the WockyMuc in the disco CB PEP: don't leak if pep_service_get_finish() isn't called test-resolver: don't leak if _finish() aren't called Correct a source tag Import tp-glib and gabble's valgrind suppressions Make valgrind suppressions work in out-of-tree builds Don't leak stanzas in the extract_errors test Don't leak a stanza in node iterator test Support passing arguments to tests when valgrinding Add a README for the tests Ignore more vim swapfiles With this branch, 3 tests had leaks or errors under valgrind: wocky-porter-test wocky-connector-test wocky-test-sasl-auth The former is a leak in test_cancel_iq_closing(); seems like at least the in stream is leaked. I think it's because the call to _close_async() never ends, and so keeps sched_in alive, and hence the connection.