[Experiment] A validation error is displayed in the "Browser Console" when trying to enroll in the "Import experiment: Image (new users)" experiments on Firefox 100 to 103 versions
Categories
(Firefox :: Messaging System, defect, P1)
Tracking
()
Tracking | Status | |
---|---|---|
firefox-esr102 | --- | unaffected |
firefox109 | --- | unaffected |
firefox110 | --- | unaffected |
firefox111 | --- | unaffected |
firefox112 | --- | fixed |
People
(Reporter: mcoman, Assigned: emcminn)
References
(Regression)
Details
(Keywords: regression)
Attachments
(1 file)
1.30 MB,
image/gif
|
Details |
[Notes]:
- The following validation error is displayed in the "Browser Console": Error
[Affected versions]:
- Firefox Release 100.0.2 - Build ID: 20220519220738
- Firefox Release 101.0.1 - Build ID: 20220608170832
- Firefox Release 102.0.1 - Build ID: 20220705093820
- Firefox Release 103.0.2 - Build ID: 20220808125904
[Affected Platforms]:
- Windows 10 x64
- macOS 12.6.1
- Linux Mint 20.2
[Prerequisites]:
- Have a Firefox en-* locale installed.
- Have a Firefox profile betwen 7 and 28 days old.
- Have the "[user.js](https://drive.google.com/file/d/1Smyn7FOByDTtgFWGsdgV_dOywUfgqTXx/view?usp=share_link" file saved to disk.
[Steps to reproduce]:
- Navigate to the profile folder from prerequisites and paste the "user.js" file.
- Open the browser using the profile from the prerequisites.
- Navigate to the "about:studies" page and observe the listed experiments.
[Expected result]:
- The "import-bookmarks-image-and-segmentation-en" experiment is displayed.
[Actual result]:
- The "import-bookmarks-image-and-segmentation-en" experiment is NOT displayed.
[Additional Notes]:
Comment 1•2 years ago
|
||
The schema for Firefox Messaging System did not support empty messages in 100-103.
To work around this, we can include a dummy message in the control branch that has targeting: false
so that it will never be shown.
We cannot use the schema validation opt-out because that was added in Firefox 107.
Updated•2 years ago
|
Assignee | ||
Comment 2•2 years ago
|
||
This has been fixed by adding the following dummy message to the control
branch:
{
"id": "import-bookmarks-image-and-segmentation-en:control",
"template": "spotlight",
"groups": [],
"content": {
"template": "multistage",
"transitions": true,
"screens": [
{
"content": {}
}
]
},
"frequency": {
"lifetime": 1
},
"trigger": {
"id": "defaultBrowserCheck"
},
"targeting": "false"
}
targeting: false
will ensure the message is not seen and no impression is sent.
Comment 3•2 years ago
|
||
Could have been fixed by increasing the minimum version of the experiment and keeping the {}
message value? If so, should it be updated to simplify for future experiments that clone it?
Assignee | ||
Comment 4•2 years ago
|
||
Yes I think so - this fix had already been seen by QA so I didn't want to give them another thing. However now it sounds like they'll be running another round of tests anyway, so I can update the fix for simplicity's sake.
Comment 5•2 years ago
|
||
Set release status flags based on info from the regressing bug 1752665
Updated•2 years ago
|
Reporter | ||
Comment 6•2 years ago
|
||
This was part of the testing performed on [QA-1733] import-bookmarks-image-and-segmentation request. Considering this I am marking this issue as Verified.
Comment 7•2 years ago
|
||
Set release status flags based on info from the regressing bug 1752665
Comment 8•2 years ago
|
||
marking as fixed even though it was fixed two months ago, to prevent the bot from picking it up again
Description
•