Bug 16265
| Summary: | run-webkit-tests should run failing tests so it's easier to spot fixed tests | ||
|---|---|---|---|
| Product: | WebKit | Reporter: | Adam Roben (:aroben) <aroben> |
| Component: | Tools / Tests | Assignee: | Nobody <webkit-unassigned> |
| Status: | NEW | ||
| Severity: | Normal | CC: | alice.barraclough |
| Priority: | P2 | ||
| Version: | 528+ (Nightly build) | ||
| Hardware: | All | ||
| OS: | All | ||
Adam Roben (:aroben)
The current mechanism for dealing with failing regression tests is to put them in a Skipped file, which means that under normal conditions the tests will never be run. This makes it hard to know when a test is fixed. It would be nice if we could check in the failing results for tests but mark them as known failures. After a test run, run-webkit-tests could tell you which tests failed (as it currently does), and which known failures had different results.
| Attachments | ||
|---|---|---|
| Add attachment proposed patch, testcase, etc. |
Adam Roben (:aroben)
Perhaps we could make the Skipped file use an INI format, with sections such as [fail] and [hang]. We'd probably want to rename Skipped to Failures or Exceptions or something similar.