<?xml version="1.0" encoding="UTF-8" standalone="yes" ?>
<!DOCTYPE bugzilla SYSTEM "https://bugs.webkit.org/page.cgi?id=bugzilla.dtd">

<bugzilla version="5.0.4.1"
          urlbase="https://bugs.webkit.org/"
          
          maintainer="admin@webkit.org"
>

    <bug>
          <bug_id>142334</bug_id>
          
          <creation_ts>2015-03-05 03:20:47 -0800</creation_ts>
          <short_desc>[GTK] Apps Using Webkit Under Wayland Do Not Render Pages (GNOME 3.15)</short_desc>
          <delta_ts>2015-07-01 09:00:05 -0700</delta_ts>
          <reporter_accessible>1</reporter_accessible>
          <cclist_accessible>1</cclist_accessible>
          <classification_id>1</classification_id>
          <classification>Unclassified</classification>
          <product>WebKit</product>
          <component>WebKitGTK</component>
          <version>528+ (Nightly build)</version>
          <rep_platform>Other</rep_platform>
          <op_sys>Linux</op_sys>
          <bug_status>RESOLVED</bug_status>
          <resolution>FIXED</resolution>
          
          
          <bug_file_loc></bug_file_loc>
          <status_whiteboard></status_whiteboard>
          <keywords></keywords>
          <priority>P2</priority>
          <bug_severity>Normal</bug_severity>
          <target_milestone>---</target_milestone>
          
          <blocked>81456</blocked>
          <everconfirmed>1</everconfirmed>
          <reporter name="Antoine Saroufim">Antoine.Saroufim</reporter>
          <assigned_to name="Nobody">webkit-unassigned</assigned_to>
          <cc>cgarcia</cc>
    
    <cc>clopez</cc>
    
    <cc>mcatanzaro</cc>
    
    <cc>unixman</cc>
    
    <cc>zan</cc>
          

      

      

      

          <comment_sort_order>oldest_to_newest</comment_sort_order>  
          <long_desc isprivate="0" >
    <commentid>1074464</commentid>
    <comment_count>0</comment_count>
    <who name="Antoine Saroufim">Antoine.Saroufim</who>
    <bug_when>2015-03-05 03:20:47 -0800</bug_when>
    <thetext>Apps running under Wayland on openSUSE Factory using GNOME 3.15 do not render pages at all. Gnome-maps would display an empty UI and Epiphany would keep loading the pages forever but wouldn&apos;t display anything. Please note that everything works fine under x11 on the same system.</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>1075601</commentid>
    <comment_count>1</comment_count>
    <who name="Michael Catanzaro">mcatanzaro</who>
    <bug_when>2015-03-09 08:02:58 -0700</bug_when>
    <thetext>FWIW: WebKitGTK+ 2.6.5 works fine running under Wayland with GNOME 3.14 in Fedora. At least, it works well enough for me to post this Bugzilla comment using it!

FWIW: The current plan is to use Wayland as the default session in Fedora 23 (release: November) but WebKitGTK+ is important enough to delay that if it&apos;s not working.</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>1075733</commentid>
    <comment_count>2</comment_count>
    <who name="Zan Dobersek">zan</who>
    <bug_when>2015-03-09 13:37:09 -0700</bug_when>
    <thetext>I can both post Bugzilla comments and reproduce the problem on maps.google.com with tip-of-tree WebKitGTK+, GTK+ 3.15.10 and Weston master.</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>1075735</commentid>
    <comment_count>3</comment_count>
    <who name="Zan Dobersek">zan</who>
    <bug_when>2015-03-09 13:41:45 -0700</bug_when>
    <thetext>A crash is occurring on maps.google.com. Only reproducible under Wayland.

Backtrace:
#0  _fill_xrgb32_lerp_opaque_spans (abstract_renderer=0x7ffcc3c5dc90, y=110, h=0, spans=0x1baee60, num_spans=8) at cairo-image-compositor.c:2249
#1  0x00007f11a15bd196 in blit_a8 (xmax=101, xmin=0, height=1, y=110, spans=0x1baee50, renderer=0x7ffcc3c5dc90, cells=&lt;optimized out&gt;) at cairo-tor-scan-converter.c:1635
#2  glitter_scan_converter_render (renderer=0x7ffcc3c5dc90, antialias=1, winding_mask=&lt;optimized out&gt;, converter=&lt;optimized out&gt;) at cairo-tor-scan-converter.c:1786
#3  _cairo_tor_scan_converter_generate (converter=0x19faae0, renderer=0x7ffcc3c5dc90) at cairo-tor-scan-converter.c:1849
#4  0x00007f11a15afc99 in composite_polygon (extents=extents@entry=0x7ffcc3c5f5a0, polygon=polygon@entry=0x7ffcc3c5f160, fill_rule=fill_rule@entry=CAIRO_FILL_RULE_WINDING, 
    antialias=antialias@entry=CAIRO_ANTIALIAS_DEFAULT, compositor=0x7f11a1868780 &lt;spans&gt;, compositor=0x7f11a1868780 &lt;spans&gt;) at cairo-spans-compositor.c:801
#5  0x00007f11a15b066a in clip_and_composite_polygon (compositor=compositor@entry=0x7f11a1868780 &lt;spans&gt;, extents=extents@entry=0x7ffcc3c5f5a0, polygon=polygon@entry=0x7ffcc3c5f160, 
    fill_rule=CAIRO_FILL_RULE_WINDING, antialias=antialias@entry=CAIRO_ANTIALIAS_DEFAULT) at cairo-spans-compositor.c:967
#6  0x00007f11a15b11f7 in _cairo_spans_compositor_stroke (_compositor=0x7f11a1868780 &lt;spans&gt;, extents=0x7ffcc3c5f5a0, path=&lt;optimized out&gt;, style=0x7ffcc3c5f9b0, ctm=0x7ffcc3c5f9e0, 
    ctm_inverse=0x7ffcc3c5fa10, tolerance=0.10000000000000001, antialias=CAIRO_ANTIALIAS_DEFAULT) at cairo-spans-compositor.c:1083
#7  0x00007f11a156c1ee in _cairo_compositor_stroke (compositor=0x7f11a1868780 &lt;spans&gt;, surface=0x1d85000, op=CAIRO_OPERATOR_CLEAR, source=0xbe00ac, path=0xe46768, style=0x7ffcc3c5f9b0, 
    ctm=0x7ffcc3c5f9e0, ctm_inverse=0x7ffcc3c5fa10, tolerance=-0, antialias=CAIRO_ANTIALIAS_DEFAULT, clip=0x0) at cairo-compositor.c:157
#8  0x00007f11a157d302 in _cairo_image_surface_stroke (abstract_surface=&lt;optimized out&gt;, op=&lt;optimized out&gt;, source=&lt;optimized out&gt;, path=&lt;optimized out&gt;, style=&lt;optimized out&gt;, 
    ctm=&lt;optimized out&gt;, ctm_inverse=0x7ffcc3c5fa10, tolerance=&lt;optimized out&gt;, antialias=CAIRO_ANTIALIAS_DEFAULT, clip=0x0) at cairo-image-surface.c:964
#9  0x00007f11a15b4556 in _cairo_surface_stroke (surface=0xc99930, op=CAIRO_OPERATOR_OVER, source=0x7ffcc3c5fa40, path=0xe46768, stroke_style=0x7ffcc3c5f9b0, ctm=0x7ffcc3c5f9e0, 
    ctm_inverse=0x7ffcc3c5fa10, tolerance=0.10000000000000001, antialias=CAIRO_ANTIALIAS_DEFAULT, clip=0x0) at cairo-surface.c:2270
#10 0x00007f11a157422a in _cairo_gstate_stroke (gstate=0xee, path=0x1d85000, path@entry=0xe46768) at cairo-gstate.c:1194
#11 0x00007f11a156dbc9 in _cairo_default_context_stroke (abstract_cr=0xe46400) at cairo-default-context.c:1010
#12 0x00007f11a15670b5 in INT_cairo_stroke (cr=0xee) at cairo.c:2150
#13 0x00007f11a47ce171 in WebCore::CanvasRenderingContext2D::strokeInternal(WebCore::Path const&amp;) () from /data/webkit/git/WebKitBuild/Release/lib/libwebkit2gtk-4.0.so.37
#14 0x00007f11a506529b in WebCore::jsCanvasRenderingContext2DPrototypeFunctionStroke(JSC::ExecState*) () from /data/webkit/git/WebKitBuild/Release/lib/libwebkit2gtk-4.0.so.37
#15 0x00007f118dffe1a8 in ?? ()
#16 0x00007ffcc3c5fdc0 in ?? ()
#17 0x00007f114e35a68c in ?? ()
#18 0x0000000000000000 in ?? ()</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>1078591</commentid>
    <comment_count>4</comment_count>
    <who name="Zan Dobersek">zan</who>
    <bug_when>2015-03-19 12:35:18 -0700</bug_when>
    <thetext>Somehow I can&apos;t reproduce the crash anymore. Can anyone try to confirm this? After maps.google.com is loaded, the relevant WebKitWebProcess should still be alive, and moving mouse across the window should change the cursor, indicating there&apos;s actual content there, just not displayed.

As far as I&apos;ve looked into this, the issue is that even under Wayland WebKitGTK+ is entering accelerated compositing mode, for which there&apos;s no support for just yet. This results in content not being displayed at all.</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>1105369</commentid>
    <comment_count>5</comment_count>
    <who name="nick">unixman</who>
    <bug_when>2015-06-29 04:28:01 -0700</bug_when>
    <thetext>it still do not render sites like vimeo, youtube.

i am able to play youtube videos but only sound, video part not seen.

i attacked a log after playing finished.

gtk-3.16.4   webkit-2.8.3  epiphany-3.16.1 mesa-10-6.0

everythings stock arch-linux with latest upgrade.

HW: intel ivybridge (HD 4XXX)</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>1105370</commentid>
    <comment_count>6</comment_count>
      <attachid>255743</attachid>
    <who name="nick">unixman</who>
    <bug_when>2015-06-29 04:36:27 -0700</bug_when>
    <thetext>Created attachment 255743
playing video ended.</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>1105371</commentid>
    <comment_count>7</comment_count>
    <who name="nick">unixman</who>
    <bug_when>2015-06-29 04:43:04 -0700</bug_when>
    <thetext>forgot: wayland-1.8.1  weston-1.8.0

weston is running on pure drm-backend even i deleted xwayland.so manually.</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>1106063</commentid>
    <comment_count>8</comment_count>
    <who name="Michael Catanzaro">mcatanzaro</who>
    <bug_when>2015-07-01 08:57:31 -0700</bug_when>
    <thetext>(In reply to comment #4) 
&gt; As far as I&apos;ve looked into this, the issue is that even under Wayland
&gt; WebKitGTK+ is entering accelerated compositing mode, for which there&apos;s no
&gt; support for just yet. This results in content not being displayed at all.

I tested this again today. When built without ENABLE_WAYLAND_TARGET, pages that trigger accelerated compositing do not load at all when running under XWayland -- that is just totally broken. But when built with ENABLE_WAYLAND_TARGET, accelerated compositing is not used and the pages work fine. So once 2.10 is out, distros can finally set ENABLE_WAYLAND_TARGET without having to disable ENABLE_X11_TARGET, thanks to Carlos Lopez, and we will be good. That was a very major problem and I&apos;m glad it appears to be solved.

I think we can simply say that accelerated compositing is not supported without ENABLE_WAYLAND_TARGET, so I will say RESOLVED FIXED. Although if it&apos;s possible to detect that we&apos;re running under XWayland and disable accelerated compositing in that case, that would be even better.

(In reply to comment #5)
&gt; gtk-3.16.4   webkit-2.8.3  epiphany-3.16.1 mesa-10-6.0

To be clear: if you upgrade to unstable WebKitGTK+ 2.9.3 and pass -DENABLE_WAYLAND_TARGET=ON to CMake, then this should work. Once the stable release 2.10.0 is out, please ask the Arch folks to use ENABLE_WAYLAND_TARGET to update and use ENABLE_WAYLAND_TARGET, then this should work.</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>1106066</commentid>
    <comment_count>9</comment_count>
    <who name="Michael Catanzaro">mcatanzaro</who>
    <bug_when>2015-07-01 09:00:05 -0700</bug_when>
    <thetext>(In reply to comment #5)
&gt; it still do not render sites like vimeo, youtube.
&gt; 
&gt; i am able to play youtube videos but only sound, video part not seen.

I can&apos;t test this since our video playback is totally broken due to bug #135972, but if it&apos;s still an issue with 2.9.3 built with ENABLE_WAYLAND_TARGET, please file a new bug.</thetext>
  </long_desc>
      
          <attachment
              isobsolete="0"
              ispatch="0"
              isprivate="0"
          >
            <attachid>255743</attachid>
            <date>2015-06-29 04:36:27 -0700</date>
            <delta_ts>2015-06-29 04:36:27 -0700</delta_ts>
            <desc>playing video ended.</desc>
            <filename>video_play.log</filename>
            <type>application/octet-stream</type>
            <size>885</size>
            <attacher name="nick">unixman</attacher>
            
              <data encoding="base64">KGVwaXBoYW55OjcyNik6IEdMaWItR09iamVjdC1XQVJOSU5HICoqOiBUaGUgcHJvcGVydHkgR3Rr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</data>

          </attachment>
      

    </bug>

</bugzilla>