New issue
Advanced search Search tips
Note: Color blocks (like or ) mean that a user may not be available. Tooltip shows the reason.

Issue 832319 link

Starred by 1 user

Issue metadata

Status: Fixed
Owner:
Closed: Apr 2018
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: ----
Pri: 2
Type: Task



Sign in to add a comment

Flakey /external/wpt/content-security-policy/embedded-enforcement tests

Project Member Reported by danyao@chromium.org, Apr 12 2018

Issue description

Hi - I'm filing this bug as Web Platform Test sheriff.

The upstream WPT PR for https://chromium-review.googlesource.com/c/chromium/src/+/999182 failed to build for firefox:nightly. The log from https://travis-ci.org/w3c/web-platform-tests/jobs/365762341 shows flaky content-security-policy/embedded-enforcement tests.

```
ERROR:/home/travis/build/w3c/web-platform-tests/tools/ci:## Unstable results ##
ERROR:/home/travis/build/w3c/web-platform-tests/tools/ci:|                                                 Test                                                |                                                 Subtest                                                 |          Results           |                  Messages                  |
ERROR:/home/travis/build/w3c/web-platform-tests/tools/ci:|-----------------------------------------------------------------------------------------------------|---------------------------------------------------------------------------------------------------------|----------------------------|--------------------------------------------|
ERROR:/home/travis/build/w3c/web-platform-tests/tools/ci:| `/content-security-policy/embedded-enforcement/allow_csp_from-header.html`                          | `Same origin iframes are always allowed.`                                                               | **FAIL: 6/10, PASS: 4/10** | `assert_true: expected true got undefined` |
ERROR:/home/travis/build/w3c/web-platform-tests/tools/ci:| `/content-security-policy/embedded-enforcement/allow_csp_from-header.html`                          | `Same origin iframes are allowed even if the Allow-CSP-From is empty.`                                  | **FAIL: 1/10, PASS: 9/10** | `assert_true: expected true got undefined` |
ERROR:/home/travis/build/w3c/web-platform-tests/tools/ci:| `/content-security-policy/embedded-enforcement/allow_csp_from-header.html`                          | `Same origin iframes are allowed even if the Allow-CSP-From is not present.`                            | **FAIL: 3/10, PASS: 7/10** | `assert_true: expected true got undefined` |
ERROR:/home/travis/build/w3c/web-platform-tests/tools/ci:| `/content-security-policy/embedded-enforcement/subsumption_algorithm-general.html`                  | `If there is no required csp, iframe should load.`                                                      | **FAIL: 9/10, PASS: 1/10** | `assert_true: expected true got undefined` |
ERROR:/home/travis/build/w3c/web-platform-tests/tools/ci:| `/content-security-policy/embedded-enforcement/subsumption_algorithm-hashes.html`                   | `'sha256-abc123' is properly subsumed.`                                                                 | **FAIL: 9/10, PASS: 1/10** | `assert_true: expected true got undefined` |
ERROR:/home/travis/build/w3c/web-platform-tests/tools/ci:| `/content-security-policy/embedded-enforcement/subsumption_algorithm-host_sources-paths.html`       | `Matching paths.`                                                                                       | **FAIL: 1/10, PASS: 9/10** | `assert_true: expected true got undefined` |
ERROR:/home/travis/build/w3c/web-platform-tests/tools/ci:| `/content-security-policy/embedded-enforcement/subsumption_algorithm-host_sources-ports.html`       | `Returned CSP should be subsumed even if the port is not specified but is a default port for a scheme.` | **FAIL: 1/10, PASS: 9/10** | `assert_true: expected true got undefined` |
ERROR:/home/travis/build/w3c/web-platform-tests/tools/ci:| `/content-security-policy/embedded-enforcement/subsumption_algorithm-host_sources-protocols.html`   | `The reverse allows iframe be to be loaded.`                                                            | **FAIL: 1/10, PASS: 9/10** | `assert_true: expected true got undefined` |
ERROR:/home/travis/build/w3c/web-platform-tests/tools/ci:| `/content-security-policy/embedded-enforcement/subsumption_algorithm-unsafe_hashed_attributes.html` | `'unsafe-hashed-attributes' is properly subsumed.`                                                      | **FAIL: 9/10, PASS: 1/10** | `assert_true: expected true got undefined` |
ERROR:/home/travis/build/w3c/web-platform-tests/tools/ci:
full_results
INFO:/home/travis/build/w3c/web-platform-tests/tools/ci:## All results ##
```

Please help investigate.
 
Cc: creis@chromium.org nick@chromium.org dcheng@chromium.org
Status: Started (was: Untriaged)
Oh, this is probably because as I found in https://bugs.chromium.org/p/chromium/issues/detail?id=828529#c26, Firefox actually fires the iframe element onload after delivering the postMessage, and after my tweak to these tests, they assume that the onload comes before  postMessage.  The tests should probably deal with the postMessage or the onload coming in either order; I'll take a crack at fixing this.
Note: the original CL (https://chromium-review.googlesource.com/c/chromium/src/+/999182) was reverted for a different reason.  I'll aim to fix this when I reland it.
Status: Fixed (was: Started)
I've relanded the original CL in issue 828529 with the CSP tests fixed.  It's been a few days and nothing exploded, so I think this is safe to close.

Sign in to add a comment