[Webkit-unassigned] [Bug 193919] [GTK] Implement back/forward touchpad gesture

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=193919

--- Comment #9 from Michael Catanzaro <mcatanzaro at igalia.com> ---
(In reply to Alexander Mikhaylenko from comment #7)
> Oops. Also, looking at it now, isn't "if ... will" a grammatical error?

Nope. Why do you think it would be?

I'd put a comma before ", or %FALSE otherwise."

> > WebKitSettings* settings
> Huh, I assumed the style checking script caught all of these :o

It normally does; not sure why it missed this one.

(In reply to Alexander Mikhaylenko from comment #7)
> Oops, it's definitely leaked. Though I get an error with
> cairo_surface_destroy() in ViewSnapshotStore.cpp being called on already
> destroyed surface if I do this. Is that one not necessary? I admit I don't
> completely understand when it gets ref-d/unref-d.

Without investigating the actual refcounting issue here: the RefPtr/GRefPtr is basically "safer g_autoptr" where unref occurs automatically at the bottom of the scope. The main difference is that you have to adopt the initial ref with adoptRef/adoptGRef, otherwise you wind up with an initial refcount of 2.

> Though it's a different warning, and I don't completely understand where
> it's coming from:
> > /home/exalm/Projects/WebKit/Source/WebKit/UIProcess/gtk/BackForwardGestureController.cpp:229:10: warning: variable ‘sinceStart’ set but not used [-Wunused-but-set-variable]
> >     auto sinceStart = MonotonicTime::now() - m_startTime;
> 
> I assume RELEASE_LOG() is a macro is substituted with nothing, which would
> also explain a lack of format warning. What is the preferred way of logging
> in gtk parts?

Hm, you might be building with logging disabled? I think logging is automatically enabled in debug builds, but I'm surprised the release log would not be enabled in release builds. Anyway, the RELEASE_LOG(foo) should ultimately boil down to a fprintf(stderr, foo), but it takes me half an hour every time I investigate all the various layers involved.

Maybe you could do something like

#if !RELEASE_LOG_DISABLED
    auto sinceStart = ...;
    RELEASE_LOG(...);
#endif

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20190130/6f1a383a/attachment-0001.html>