Bug 162547 – [GTK][EFL] imported/w3c/web-platform-tests/fetch/api/basic/accept-header.html is failing

• Component: WebCore Misc.
• Status: RESOLVED FIXED

[youenn fablet]

It seems GTK WTR is not setting Accept-Language header.

[youenn fablet]

Using minibrowser, the test is passing.
This may be an issue with GTK WTR setup?

[youenn fablet]

Created attachment 289814 [details]
Patch

[youenn fablet]

(In reply to comment #2)
> Created attachment 289814 [details]
> Patch

The soup session accept-language is well initialized, making minibrowser work as expected.
But we are creating a new soup session for testing in WRT.
This patch takes the simple approach to set Accept-Language for testing session with a fixed default value.

[Carlos Garcia Campos]

So, maybe the bug is that we are using SoupNetworkSession::defaultSession() directly in NetworkProcess::userPreferredLanguagesChanged() and we should use NetworkStorageSession::defaultStorageSession().soupNetworkSession()?

[youenn fablet]

AIUI, the default soup session gets updated after the network process is launched and properly set.

When we launch tests using WTR, test controller asks to use the testing network session. At that time, we lose all soup session specific options previously set and we are back to the default values.

[Carlos Garcia Campos]

(In reply to comment #5)
> AIUI, the default soup session gets updated after the network process is
> launched and properly set.
> 
> When we launch tests using WTR, test controller asks to use the testing
> network session. At that time, we lose all soup session specific options
> previously set and we are back to the default values.

Ok, there's probably more places where we are using the default soup session directly, we need to check that carefully, for now it make sense to use a sane default value for the accept-language property.

[Carlos Garcia Campos]

Comment on attachment 289814 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=289814&action=review

> Source/WebCore/platform/network/soup/SoupNetworkSession.cpp:69
> +    auto cookieJar = adoptGRef(createPrivateBrowsingCookieJar());
> +    auto newSoupSession = std::unique_ptr<SoupNetworkSession>(new SoupNetworkSession(cookieJar.get()));
> +    g_object_set(newSoupSession.get(), "accept-language", "en-us", nullptr);
> +    return newSoupSession;

Maybe we could simply add a fixme here.

[Michael Catanzaro]

Please do add a FIXME

[youenn fablet]

Created attachment 289842 [details]
Patch for landing

[WebKit Commit Bot]

Comment on attachment 289842 [details]
Patch for landing

Clearing flags on attachment: 289842

Committed r206387: <http://trac.webkit.org/changeset/206387>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.

[youenn fablet]

Patch is not working according bots.

[youenn fablet]

Created attachment 289931 [details]
Patch

[youenn fablet]

(In reply to comment #13)
> Created attachment 289931 [details]
> Patch

This time, it should work...

[Csaba Osztrogonác]

(In reply to comment #10)
> Comment on attachment 289842 [details]
> Patch for landing
> 
> Clearing flags on attachment: 289842
> 
> Committed r206387: <http://trac.webkit.org/changeset/206387>

It made all perf tests fail on EFL and GTK performance bots:
- https://build.webkit.org/builders/GTK%20Linux%2064-bit%20Release%20%28Perf%29/builds/6718
- https://build.webkit.org/builders/EFL%20Linux%2064-bit%20Release%20WK2%20%28Perf%29/builds/10018


(process:23353): GLib-GObject-CRITICAL **: g_object_set: assertion 'G_IS_OBJECT (object)' failed

...

[WebKit Commit Bot]

Comment on attachment 289931 [details]
Patch

Clearing flags on attachment: 289931

Committed r206431: <http://trac.webkit.org/changeset/206431>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.