Open Bug 1817727 Opened 1 year ago Updated 19 days ago

Intermittent uriloader/exthandler/tests/unit/test_filename_sanitize.js | xpcshell return code: 0

Categories

(Firefox :: File Handling, defect, P5)

Product:
Firefox
Component:
File Handling
Type:
defect

Tracking

()

People

(Reporter: intermittent-bug-filer, Unassigned)

Details

(Keywords: intermittent-failure, intermittent-testcase, leave-open, Whiteboard: [retriggered][stockwell disabled])

Attachments

(1 file)

Bug 1817727 - Disable test_filename_sanitize.js on mac because of permafailures. r=#intermittent-reviewers,aryx
48 bytes, text/x-phabricator-request
Details | Review

Filed by: chorotan [at] mozilla.com
Parsed log: https://treeherder.mozilla.org/logviewer?job_id=406237464&repo=autoland
Full log: https://firefox-ci-tc.services.mozilla.com/api/queue/v1/task/aL4Rv_yzSXu78Y7l1S0Gdg/runs/0/artifacts/public/logs/live_backing.log

task 2023-02-20T08:50:00.224Z] 08:50:00     INFO -  TEST-START | uriloader/exthandler/tests/unit/test_filename_sanitize.js
[task 2023-02-20T08:50:00.342Z] 08:50:00  WARNING -  TEST-UNEXPECTED-FAIL | uriloader/exthandler/tests/unit/test_filename_sanitize.js | xpcshell return code: 0
[task 2023-02-20T08:50:00.342Z] 08:50:00     INFO -  TEST-INFO took 117ms
[task 2023-02-20T08:50:00.342Z] 08:50:00     INFO -  >>>>>>>
[task 2023-02-20T08:50:00.342Z] 08:50:00     INFO -  (xpcshell/head.js) | test MAIN run_test pending (1)
[task 2023-02-20T08:50:00.342Z] 08:50:00     INFO -  (xpcshell/head.js) | test run_next_test 0 pending (2)
[task 2023-02-20T08:50:00.342Z] 08:50:00     INFO -  (xpcshell/head.js) | test MAIN run_test finished (2)
[task 2023-02-20T08:50:00.342Z] 08:50:00     INFO -  running event loop
[task 2023-02-20T08:50:00.342Z] 08:50:00     INFO -  uriloader/exthandler/tests/unit/test_filename_sanitize.js | Starting validate_filename_method
[task 2023-02-20T08:50:00.342Z] 08:50:00     INFO -  (xpcshell/head.js) | test validate_filename_method pending (2)

Retriggered and backfilled up to here

Could not found the regression bug yet.

Whiteboard: [retriggered]

This is actually almost perma failling as it can be seen here. I did some backfills and retriggers on both autoland and central but didn't find anything relevant. The only thing that seems strange to me is the fact that this started at the same time with the backout of Bug 1855564 from mozilla-central, and this can be seen here. Now, it is falling even where it was green before.

Hello Gijs! Could you please take a look at this? Maybe you can figure it out about this bug.

Thank you!

Flags: needinfo?(gijskruitbosch+bugs)
Summary: Intermittent uriloader/exthandler/tests/unit/test_filename_sanitize.js | single tracking bug → Almost perma uriloader/exthandler/tests/unit/test_filename_sanitize.js | xpcshell return code: 0

(In reply to Serban Stanca [:SerbanS] from comment #36)

the backout of Bug 1855564 from mozilla-central,

Was this backout definitely complete? That is, were there maybe supporting bugs that landed in the same / slightly earlier push that stayed on central?

This bug is very smelly because it started with the macOS backout and the failures are macOS-only. So it would be useful to understand this better... is there any chance that there's any leftovers on the builders from downloaded different macOS SDKs or something?

Flags: needinfo?(sstanca)
Flags: needinfo?(mh+mozilla)
Flags: needinfo?(gijskruitbosch+bugs)

is there any chance that there's any leftovers on the builders from downloaded different macOS SDKs or something?

I don't think so.

Flags: needinfo?(mh+mozilla)

I'd suggest bisecting between the landing of bug 1855564 and its backout, applying the backout on each step, and see if something changed in the meantime that somehow works with the new SDK but doesn't with the old one.

Do the mac minis get re-imaged between jobs? That is, if (hypothetically...) the new SDK means some API call Firefox/xpcshell makes changes something in whatever the macOS equivalent of the Windows registry is, is it possible that that persists on the machine?

I ask because the failures seems to do with the mimetype-vs-file-extension lookups, and I wonder if running under the new SDK changed the OS mime-to-extension mapping for whatever reason.

Flags: needinfo?(aryx.bugmail)

Not re-imaged but there are cleanup script.

Masterwayz, could the mapping of mime types to file extension on macOS be checked / purged between tasks? This might also belong into the test suite.

Flags: needinfo?(aryx.bugmail) → needinfo?(mgoossens)
Flags: needinfo?(sstanca)
Assignee: nobody → csabou
Status: NEW → ASSIGNED
Assignee: csabou → nobody
Status: ASSIGNED → NEW
Keywords: leave-open
Pushed by csabou@mozilla.com:
https://hg.mozilla.org/integration/autoland/rev/b34142ce9ea5
Disable test_filename_sanitize.js on mac because of permafailures. r=aryx DONTBUILD
Summary: Almost perma uriloader/exthandler/tests/unit/test_filename_sanitize.js | xpcshell return code: 0 → Intermittent uriloader/exthandler/tests/unit/test_filename_sanitize.js | xpcshell return code: 0
Whiteboard: [retriggered][stockwell disable-recommended] → [retriggered][stockwell disabled]
Flags: needinfo?(michelle)

I can reproduce this locally all the time on OS13. The internal api we are using at https://searchfox.org/mozilla-central/rev/7ddf1affe2c1fe192fdf88b70a21872af543666f/uriloader/exthandler/mac/nsOSHelperAppService.mm#541 to map mime types to extensions returns 'ogm' as the preferred and only extension for video/ogg.

We could provide some kind of special case for this in the Mac mime handlinig, but the simplest fix is to just adjust the test to accept either ogm or ogv as the right value.

You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: