Bug 198133 – Set default WebsiteDataStore storage quota based on StorageQuotaManager

• Status: RESOLVED
• Resolution: FIXED
• Component: Service Workers

[youenn fablet]

Set default WebsiteDataStore storage quota based on StorageQuotaManager

[youenn fablet]

<rdar://problem/51031436>

[youenn fablet]

Created attachment 370422 [details]
Patch

[Geoffrey Garen]

Can you add a test case? Seems like an API test could detect whether a permission delegate message was sent or not. Or you could use an html test case, with WebKitTestRunner configured to log or fail when sent the permission delegate message.

[youenn fablet]

(In reply to Geoffrey Garen from comment #3)
> Can you add a test case? Seems like an API test could detect whether a
> permission delegate message was sent or not. Or you could use an html test
> case, with WebKitTestRunner configured to log or fail when sent the
> permission delegate message.

WebKitTestRunner is being used to test the WebsiteDataStore delegate.
The UIDelegate delegate is tested through API tests in StorageQuota.mm.
In both cases, we do not rely on the default quota value and set it to a lower value (400ko).

We can add another test in StorageQuota.mm which would use the default quota value.
This would require to write 1GB on disk to hit the limit and trigger the delegate.
This seems a bit expensive to me but can write the test. wdyt?

[Geoffrey Garen]

> We can add another test in StorageQuota.mm which would use the default quota
> value.

Sounds good.

> This would require to write 1GB on disk to hit the limit and trigger the
> delegate.
> This seems a bit expensive to me but can write the test. wdyt?

Hmmm... As a regression test, it would be sufficient just to verify that 50MB + 1b did not trigger the delegate. Of course, 1GB is an even more valuable test. Here's a suggestion: Let's do a 1GB test (and make sure to delete afterwards!) if the test can run in less than 1s, and otherwise do a 50MB + 1b test. How does that sound?

[youenn fablet]

Created attachment 370464 [details]
Patch

[youenn fablet]

> Hmmm... As a regression test, it would be sufficient just to verify that
> 50MB + 1b did not trigger the delegate. Of course, 1GB is an even more
> valuable test. Here's a suggestion: Let's do a 1GB test (and make sure to
> delete afterwards!) if the test can run in less than 1s, and otherwise do a
> 50MB + 1b test. How does that sound?

I tried the 1GB and we are far from 1s, maybe we should try to optimize.
Anyway, I added a test that stores 100MB without hitting the delegate.

[EWS Watchlist]

Attachment 370464 [details] did not pass style-queue:


ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/StorageQuota.mm:45:  Alphabetical sorting problem.  [build/include_order] [4]
ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/StorageQuota.mm:189:  Multi-line string ("...") found.  This lint script doesn't do well with such strings, and may give bogus warnings.  They're ugly and unnecessary, and you should use concatenation instead".  [readability/multiline_string] [5]
ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/StorageQuota.mm:212:  Multi-line string ("...") found.  This lint script doesn't do well with such strings, and may give bogus warnings.  They're ugly and unnecessary, and you should use concatenation instead".  [readability/multiline_string] [5]
Total errors found: 3 in 4 files


If any of these errors are false positives, please file a bug against check-webkit-style.

[Geoffrey Garen]

Comment on attachment 370464 [details]
Patch

r=me

[WebKit Commit Bot]

Comment on attachment 370464 [details]
Patch

Clearing flags on attachment: 370464

Committed r245698: <https://trac.webkit.org/changeset/245698>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.

[Valerii]

Can anyone confirm that by applying this change the default 50Mb cacheStorage limit has been bumped to 1Gb?

[youenn fablet]

(In reply to Valerii from comment #12)
> Can anyone confirm that by applying this change the default 50Mb
> cacheStorage limit has been bumped to 1Gb?

This is the default now for WebKit.
Some WebKit applications can change this value, see for instance WKWebsiteDataStoreConfigurationSetPerOriginStorageQuota.

[Valerii]

Anyway I can figure out whether iOS Safari is one of those applications changing the value or does it rely on default?

[youenn fablet]

(In reply to Valerii from comment #14)
> Anyway I can figure out whether iOS Safari is one of those applications
> changing the value or does it rely on default?

Nowadays, there is a unified quota for IDB and Cache.
After a given threshold, user will be prompted to allow more storage or not.
I believe the current threshold is indeed 1GB for MobileSafari 13.

You can test this using something like https://demo.agektmr.com/storage.

[Valerii]

Thanks a lot for the information provided!
I unfortunately was not able to find that information in any documentation I searched.