Bug 173757 – [WPE] Use JSC API to send script messages from web extension in tests

• Component: WPE WebKit
• Status: RESOLVED FIXED

[Carlos Garcia Campos]

The GTK+ port uses DOM bindings for that. This will fix /wpe/WebKitSecurityManager/file-xhr and all TestConsoleMessage

[Carlos Garcia Campos]

Created attachment 313694 [details]
Patch

[Zan Dobersek]

Comment on attachment 313694 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=313694&action=review

> Tools/TestWebKitAPI/Tests/WebKitGLib/WebExtensionTest.cpp:266
> +    GUniquePtr<char> escapedMessageString(static_cast<char*>(g_malloc(strlen(messageString.get()) * 2 + 1)));
> +    char* src = messageString.get();
> +    char* dest = escapedMessageString.get();
> +    while (*src) {
> +        if (*src == '"') {
> +            *dest++ = '\\';
> +            *dest++ = '"';
> +        } else
> +            *dest++ = *src;
> +        src++;
> +    }
> +    *dest = '\0';

Would it be possible to use something like g_strescape() here? And use g_strcompress() on the receiving end, if necessary or possible?

Apart from that, src++ should be ++src.

[Carlos Garcia Campos]

(In reply to Zan Dobersek from comment #2)
> Comment on attachment 313694 [details]
> Patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=313694&action=review
> 
> > Tools/TestWebKitAPI/Tests/WebKitGLib/WebExtensionTest.cpp:266
> > +    GUniquePtr<char> escapedMessageString(static_cast<char*>(g_malloc(strlen(messageString.get()) * 2 + 1)));
> > +    char* src = messageString.get();
> > +    char* dest = escapedMessageString.get();
> > +    while (*src) {
> > +        if (*src == '"') {
> > +            *dest++ = '\\';
> > +            *dest++ = '"';
> > +        } else
> > +            *dest++ = *src;
> > +        src++;
> > +    }
> > +    *dest = '\0';
> 
> Would it be possible to use something like g_strescape() here? And use
> g_strcompress() on the receiving end, if necessary or possible?

The problem is not the receiving end, but JSC. If we pass something like

"Foo"bar"Baz" 

the script fails to run, so it never gets to the other end. We don't receive anything escaped in the other end. Also we would need to add ifdefs in the UI process side too. I started using g_strescape(), but the problem is that it escapes all non ascii characters, making the test fail, because one of the error messages uses utf8. So, since what we only need is escaping " because we are passing "%s" to JSC, I thought it was simple enough to do it here.

> Apart from that, src++ should be ++src.

Ok, just curiosity, does it make any difference in this case?

[Carlos Garcia Campos]

Committed r218743: <http://trac.webkit.org/changeset/218743>

[Zan Dobersek]

(In reply to Carlos Garcia Campos from comment #3)
> > Apart from that, src++ should be ++src.
> 
> Ok, just curiosity, does it make any difference in this case?

Not in terms of correctness, but it's good practice to always use the prefixed increment for this.

[Zan Dobersek]

(In reply to Zan Dobersek from comment #5)
> (In reply to Carlos Garcia Campos from comment #3)
> > > Apart from that, src++ should be ++src.
> > 
> > Ok, just curiosity, does it make any difference in this case?
> 
> Not in terms of correctness, but it's good practice to always use the
> prefixed increment for this.

... where possible.

[Michael Catanzaro]

Maybe for with user-defined increment operators, but for a primitive type like int, I don't think so... all compilers should know that they are the same if the value is not used as part of another expression, so I'd be surprised if the generated code was different, and the postincrement form is more natural and familiar.

Doesn't matter, though.