Bug 69329 – [WebKit2][gtk] Fix warnings while generating documentation using gtk-doc.

• Component: WebKit2
• Status: RESOLVED FIXED

[Nayan Kumar K]

Fix warnings while generating WebKit2-GTK+ API documentation using gtk-doc.

[Nayan Kumar K]

Created attachment 109599 [details]
Warning fix

[Gustavo Noronha (kov)]

Comment on attachment 109599 [details]
Warning fix

View in context: https://bugs.webkit.org/attachment.cgi?id=109599&action=review

> Source/WebKit2/UIProcess/API/gtk/WebKitError.h:37
> + * @WEBKIT_NETWORK_ERROR_FAILED: Protocol breakdown is detected.

Protocol breakdown sounds unnecessarily techy, can we do better? =) is this what we get when there's failure resolving DNS, the connection is refused or drops? What's the difference from this to transport error?

> Source/WebKit2/UIProcess/API/gtk/WebKitError.h:40
> + * @WEBKIT_NETWORK_ERROR_CANCELLED: Requested operation is invalid.

Cancelled = invalid?

[Nayan Kumar K]

Created attachment 109621 [details]
Warning fix

Updated patch

[Nayan Kumar K]

Comment on attachment 109599 [details]
Warning fix

View in context: https://bugs.webkit.org/attachment.cgi?id=109599&action=review

>> Source/WebKit2/UIProcess/API/gtk/WebKitError.h:37

> 
> Protocol breakdown sounds unnecessarily techy, can we do better? =) is this what we get when there's failure resolving DNS, the connection is refused or drops? What's the difference from this to transport error?

:). Changed the description for WEBKIT_NETWORK_ERROR_FAILED to "Resource loading failed" in modified patch.

Please note that, there aren't any documentation for these enums anywhere in source code. WebKit1 lacks the documentation for these enums, mac and other ports doesn't expose them and it is not clear from source code when they get triggered. These documentation are the best I could make by referring Source/WebCore/platform/gtk/ErrorsGtk.cpp and http://doc.qt.nokia.com/4.7-snapshot/qnetworkreply.html

>> Source/WebKit2/UIProcess/API/gtk/WebKitError.h:40
>> + * @WEBKIT_NETWORK_ERROR_CANCELLED: Requested operation is invalid.
> 
> Cancelled = invalid?

Changed it to 'Load request cancelled', as per Source/WebCore/platform/gtk/ErrorsGtk.cpp.

[Martin Robinson]

Comment on attachment 109621 [details]
Warning fix

View in context: https://bugs.webkit.org/attachment.cgi?id=109621&action=review

I really, really, really, really appreciate this! Please accept my humble suggestion below.

> Source/WebKit2/UIProcess/API/gtk/WebKitError.h:41
> + * @WEBKIT_NETWORK_ERROR_FAILED: Resource loading failed.
> + * @WEBKIT_NETWORK_ERROR_TRANSPORT: Transport protocol error.
> + * @WEBKIT_NETWORK_ERROR_UNKNOWN_PROTOCOL: Can not honor the request because protocol is not known.
> + * @WEBKIT_NETWORK_ERROR_CANCELLED: Load request is cancelled.
> + * @WEBKIT_NETWORK_ERROR_FILE_DOES_NOT_EXIST: File does not exists error.

I think we should use a consistent phrase for all of these and avoid the use of periods (since they aren't complete sentences):

So:
Generic load failure
Load failure due to transport error
Load failure due to unknown protocol
Load failure due to cancellation
Load failure due to missing file

[Nayan Kumar K]

Created attachment 109629 [details]
Warning fix

[Nayan Kumar K]

(In reply to comment #5)
> (From update of attachment 109621 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=109621&action=review
> 
> I really, really, really, really appreciate this! Please accept my humble suggestion below.
> 

Thank you for the suggestion, modified the patch accordingly.

[WebKit Review Bot]

Comment on attachment 109629 [details]
Warning fix

Clearing flags on attachment: 109629

Committed r96606: <http://trac.webkit.org/changeset/96606>

[WebKit Review Bot]

All reviewed patches have been landed.  Closing bug.