Bug 190196 – [GTK] Enable more Cache Storage and Service Workers tests

• Component: WebKitGTK
• Status: RESOLVED FIXED

[Frédéric Wang (:fredw)]

Currently most of them are skipped but we actually already pass most of them.

[Frédéric Wang (:fredw)]

Created attachment 351379 [details]
Patch

[Michael Catanzaro]

Comment on attachment 351379 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=351379&action=review

> LayoutTests/platform/gtk/http/wpt/cache-storage/cache-quota.any-expected.txt:6
> +FAIL Hitting cache quota for non opaque responses assert_unreached: Reached unreachable code
> +FAIL Hitting cache quota for padded responses assert_unreached: Reached unreachable code

Yeah, I'm not really comfortable with having this in the expected results. I think this should be marked as Gailure

> LayoutTests/platform/gtk/imported/w3c/web-platform-tests/service-workers/service-worker/registration-script.https-expected.txt:3
> +FAIL Registering invalid chunked encoding script assert_unreached: Should have rejected: Registration of invalid chunked encoding script should fail. Reached unreachable code
> +FAIL Registering invalid chunked encoding script with flush assert_unreached: Should have rejected: Registration of invalid chunked encoding script should fail. Reached unreachable code

Ditto

[Michael Catanzaro]

(In reply to Michael Catanzaro from comment #2)
> Yeah, I'm not really comfortable with having this in the expected results. I
> think this should be marked as Gailure

Um, Failure, of course.

[youenn fablet]

(In reply to Michael Catanzaro from comment #2)
> Comment on attachment 351379 [details]
> Patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=351379&action=review
> 
> > LayoutTests/platform/gtk/http/wpt/cache-storage/cache-quota.any-expected.txt:6
> > +FAIL Hitting cache quota for non opaque responses assert_unreached: Reached unreachable code
> > +FAIL Hitting cache quota for padded responses assert_unreached: Reached unreachable code
> 
> Yeah, I'm not really comfortable with having this in the expected results. I
> think this should be marked as Gailure

Having consistent failure is better than having some changes in the behavior without noticing it. Generally, we go with "FAIL" and tests marked as PASS.

[Frédéric Wang (:fredw)]

Created attachment 351381 [details]
Patch

[Frédéric Wang (:fredw)]

Created attachment 351382 [details]
Patch

[Michael Catanzaro]

Comment on attachment 351382 [details]
Patch

I mean, I know that's your preference... but I'd rather have [ Failure ] expectations so we can track what's broken... this way the expectations function as a TODO list of things to fix.

[WebKit Commit Bot]

Comment on attachment 351382 [details]
Patch

Clearing flags on attachment: 351382

Committed r236742: <https://trac.webkit.org/changeset/236742>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.