xorg-server @1.17.2_0: XQuartz 2.7.7 on Snow Leopard, Mac OS X 10.6.8, produces defects

[ballapete (Peter "Pete" Dyballa)]

The screenshot shows some defects when drawing the GNU Emacs windows (frames in Emacs speak) and defects when restoring the original screen background. The right-most Emacs, 25.0.50, was widened to the left and then (kind of) restored to the original width. You can see some black on the left and the bottom in the white line that surround this Emacs. The left-most Emacs, 23.4x, was widened to the right. This led to the white holes in the black lines above and below the widow bar. The lower left shows a similar defect although I did not change the window's height. The middle Emacs "Live-Kastel 2015" shows an "unfinished" black line just below the title bar.

[ballapete (Peter "Pete" Dyballa)]

The mentioned screenshot

[ballapete (Peter "Pete" Dyballa)]

XQuartz 2.7.7 (xorg-server 1.17.99.901) shows the same behaviour.

[ballapete (Peter "Pete" Dyballa)]

XQuartz 2.7.8 (xorg-server 1.16.4) shows correct behaviour. Therefore it also draws black lines around the X window manager's desktop and GNU Emacs "25.X.50".

[ballapete (Peter "Pete" Dyballa)]

Replying to Peter_Dyballa@…:

XQuartz 2.7.7 (xorg-server 1.17.99.901) shows the same behaviour.

Ditto XQuartz 2.7.7 (xorg-server 1.17.99.902).

[ryandesign (Ryan Carsten Schmidt)]

When you file a ticket, remember to Cc the port maintainer so that they learn about the problem sooner.

[graziosi.angelo@…]

Hi Ryan, just for completeness...

Replying to ryandesign@…:

When you file a ticket, remember to Cc the port maintainer so that they learn about the problem sooner.

I have just upgraded to xorg-server to 1.17.3. I see that XQuartz 2.7.8 is out (​http://xquartz.macosforge.org/landing/), but when I start an X application, the "X11/Informazioni su X11" (i.e. 'About X11' or 'Information on X11'), says

XQuartz 2.7.7 (xorg-server 1.17.3)

I wonder if this should be expected...

Thanks, Angelo.

[jeremyhu (Jeremy Huddleston Sequoia)]

Angelo, you're running MacPorts' X11.app there, which is 1.17.3. It's newer than the XQuartz 2.7.8 server which contains 1.16.4

[jeremyhu (Jeremy Huddleston Sequoia)]

I wonder if this is the same as ​http://xquartz.macosforge.org/trac/ticket/290

I dropped that patch in 1.17 because that code had been refactored and the patch no longer applied.

[ballapete (Peter "Pete" Dyballa)]

Replying to jeremyhu@…:

I wonder if this is the same as ​http://xquartz.macosforge.org/trac/ticket/290

I dropped that patch in 1.17 because that code had been refactored and the patch no longer applied.

There seem to be similarities, indeed. Which is that dropped patch? I'd like to rebuild xorg-server-devel @1.17.99.902_0 with that patch on PPC Leopard to see whether the behaviour gets cured. And later on Snow Leopard. And how can I build the demo programme from border.c?

[ballapete (Peter "Pete" Dyballa)]

XQuartz 2.7.7 (xorg-server 1.17.4) on PPC Leopard, Mac OS X 10.5.8, also shows this problem with border lines.

[jeremyhu (Jeremy Huddleston Sequoia)]

Replying to Peter_Dyballa@…:

Replying to jeremyhu@…:

I wonder if this is the same as ​http://xquartz.macosforge.org/trac/ticket/290

I dropped that patch in 1.17 because that code had been refactored and the patch no longer applied.

There seem to be similarities, indeed. Which is that dropped patch?

​https://github.com/XQuartz/xorg-server/commit/596efccbd9fa0a4b6c5ea64fc1f2e11a42c17e01

[ballapete (Peter "Pete" Dyballa)]

Replying to jeremyhu@…:

Replying to Peter_Dyballa@…:

​https://github.com/XQuartz/xorg-server/commit/596efccbd9fa0a4b6c5ea64fc1f2e11a42c17e01

The code in a recent snapshot has changed a lot. The first four patches cannot be applied because there is no Rootless* and IsFramedWindow* code anymore in mi/miexpose.c. What's left from the set of patches is attached as miexpose.c-patch. I think the compilation will finish when I'll be in deep sleep…

[ballapete (Peter "Pete" Dyballa)]

What is left from the 596efccbd9fa0a4b6c5ea64fc1f2e11a42c17e01 patch set

[ballapete (Peter "Pete" Dyballa)]

Compilation fails early here:

/bin/sh ../libtool  --tag=CC   --mode=compile /usr/bin/gcc-4.2 -std=gnu99 -DHAVE_CONFIG_H -I. -I../include   -I/opt/local/include -I/usr/include -I/opt/mports/trunk/dports/x11/xorg-server-devel/files/dri -I/opt/mports/trunk/dports/x11/xorg-server-devel/files/include -DHAVE_DIX_CONFIG_H -Wall -Wpointer-arith -Wmissing-declarations -Wformat=2 -Wstrict-prototypes -Wmissing-prototypes -Wnested-externs -Wbad-function-cast -Wold-style-definition -Wdeclaration-after-statement -Wunused -Wuninitialized -Wshadow -Wmissing-noreturn -Wmissing-format-attribute -Wredundant-decls -Werror=implicit -Werror=nonnull -Werror=init-self -Werror=main -Werror=missing-braces -Werror=sequence-point -Werror=return-type -Werror=trigraphs -Werror=write-strings -Werror=address -Werror=int-to-pointer-cast -Werror=pointer-to-int-cast -fno-strict-aliasing -fno-strict-aliasing -D_DEFAULT_SOURCE -D_BSD_SOURCE -DHAS_FCHOWN -DHAS_STICKY_DIR_BIT -I/opt/local/include -I/opt/local/include/pixman-1 -I/opt/local/include -I/opt/local/include/freetype2 -I/opt/local/include -I/opt/local/include/libpng16 -I/opt/local/include  -I../include -I../include -I../Xext  -I../damageext -I../xfixes -I../Xi -I../mi -I../miext/sync -I../miext/shadow  -I../miext/damage -I../render -I../randr -I../fb -I../dbe -I../present -pipe -Os -arch ppc -DROOTLESS_WORKAROUND -DROOTLESS_SAFEALPHA -DNO_ALLOCA -MT libfb_la-fbpict.lo -MD -MP -MF .deps/libfb_la-fbpict.Tpo -c -o libfb_la-fbpict.lo `test -f 'fbpict.c' || echo './'`fbpict.c
libtool: compile:  /usr/bin/gcc-4.2 -std=gnu99 -DHAVE_CONFIG_H -I. -I../include -I/opt/local/include -I/usr/include -I/opt/mports/trunk/dports/x11/xorg-server-devel/files/dri -I/opt/mports/trunk/dports/x11/xorg-server-devel/files/include -DHAVE_DIX_CONFIG_H -Wall -Wpointer-arith -Wmissing-declarations -Wformat=2 -Wstrict-prototypes -Wmissing-prototypes -Wnested-externs -Wbad-function-cast -Wold-style-definition -Wdeclaration-after-statement -Wunused -Wuninitialized -Wshadow -Wmissing-noreturn -Wmissing-format-attribute -Wredundant-decls -Werror=implicit -Werror=nonnull -Werror=init-self -Werror=main -Werror=missing-braces -Werror=sequence-point -Werror=return-type -Werror=trigraphs -Werror=write-strings -Werror=address -Werror=int-to-pointer-cast -Werror=pointer-to-int-cast -fno-strict-aliasing -fno-strict-aliasing -D_DEFAULT_SOURCE -D_BSD_SOURCE -DHAS_FCHOWN -DHAS_STICKY_DIR_BIT -I/opt/local/include -I/opt/local/include/pixman-1 -I/opt/local/include -I/opt/local/include/freetype2 -I/opt/local/include -I/opt/local/include/libpng16 -I/opt/local/include -I../include -I../include -I../Xext -I../damageext -I../xfixes -I../Xi -I../mi -I../miext/sync -I../miext/shadow -I../miext/damage -I../render -I../randr -I../fb -I../dbe -I../present -pipe -Os -arch ppc -DROOTLESS_WORKAROUND -DROOTLESS_SAFEALPHA -DNO_ALLOCA -MT libfb_la-fbpict.lo -MD -MP -MF .deps/libfb_la-fbpict.Tpo -c fbpict.c  -fno-common -DPIC -o .libs/libfb_la-fbpict.o
fbpict.c: In function ‘fbGlyphs’:
fbpict.c:148: error: too many arguments to function ‘image_from_pict’
fbpict.c:174: error: too many arguments to function ‘image_from_pict’
fbpict.c:177: error: too many arguments to function ‘image_from_pict’
make[1]: *** [libfb_la-fbpict.lo] Error 1
make[1]: Leaving directory `/opt/local/var/macports/build/_opt_mports_trunk_dports_x11_xorg-server-devel/xorg-server-devel/work/xorg-server-devel-1.17.99.902/fb'

So I would have to try it better with xorg-server…

[ballapete (Peter "Pete" Dyballa)]

Replying to jeremyhu@…:

​https://github.com/XQuartz/xorg-server/commit/596efccbd9fa0a4b6c5ea64fc1f2e11a42c17e01

Something's not right with that file. Every time I try to apply it I get this report:

[g]patch: **** malformed patch at line 60: @@ -559,6 +574,18 @@ miPaintWindow(WindowPtr pWin, RegionPtr prgn, int what)

Could you send me that file by eMail or explain how I could download it? I copied from Safari since I don't know how to fetch the file contents properly.

Applying the patches manually seems to provoke errors…

[jeremyhu (Jeremy Huddleston Sequoia)]

You can apply a .diff to the end to get the raw patch:

​https://github.com/XQuartz/xorg-server/commit/596efccbd9fa0a4b6c5ea64fc1f2e11a42c17e01.diff

But that won't help much. The contents changed upstream, so the patch no longer applies (as you discovered). You'll need to rewrite the changes and generate a new patch. The exact contents that are there now should serve as a guide but the implementation will be different now.

[ballapete (Peter "Pete" Dyballa)]

Replying to jeremyhu@…:

You can apply a .diff to the end to get the raw patch:

​https://github.com/XQuartz/xorg-server/commit/596efccbd9fa0a4b6c5ea64fc1f2e11a42c17e01.diff

For hunk 3 I need some help. The old code could have been:

@@ -574,6 +589,18 @@
     else {
         PixmapPtr pixmap;
 
         tile_x_off = drawable->x;
         tile_y_off = drawable->y;

which is now:

    else {
        PixmapPtr pixmap;

        fill = pWin->border;
        solid = pWin->borderIsPixel;

        /* servers without pixmaps draw their own borders */
        if (!pScreen->GetWindowPixmap)
            return;
        pixmap = (*pScreen->GetWindowPixmap) ((WindowPtr) drawable);
        drawable = &pixmap->drawable;

        while (pWin->backgroundState == ParentRelative)
            pWin = pWin->parent;

        tile_x_off = pWin->drawable.x;
        tile_y_off = pWin->drawable.y;

Can the patch

+#ifdef ROOTLESS
+	if(IsFramedWindow(pWin)) {
+	    RootlessStartDrawing(pWin);
+	    RootlessDamageRegion(pWin, prgn);
+	    
+	    if(!pWin->borderIsPixel &&
+		pWin->backgroundState == ParentRelative) {
+		RootlessSetPixmapOfAncestors(pWin);
+	    }
+	}
+#endif
+

be installed before the newly added code, i.e., after the else clause? Similarly for hunk 4. The old code could have been:

@@ -582,6 +609,12 @@
             return;
         pixmap = (*pScreen->GetWindowPixmap) ((WindowPtr) drawable);
         drawable = &pixmap->drawable;
#ifdef COMPOSITE
         draw_x_off = pixmap->screen_x;
         draw_y_off = pixmap->screen_y;

which is now:

            return;
        pixmap = (*pScreen->GetWindowPixmap) ((WindowPtr) drawable);
        drawable = &pixmap->drawable;

        while (pWin->backgroundState == ParentRelative)
            pWin = pWin->parent;

        tile_x_off = pWin->drawable.x;
        tile_y_off = pWin->drawable.y;

#ifdef COMPOSITE
        draw_x_off = pixmap->screen_x;
        draw_y_off = pixmap->screen_y;

I think the patch

+
+#ifdef XQUARTZ_CLIP_DEBUG
+        ErrorF("     Draw: %d %d %d %d\n",
+               drawable->x, drawable->y, drawable->width, drawable->height);    
+#endif
+	

can safely be installed before the newly added code, i.e., the while loop, it's just for debugging means.

[ballapete (Peter "Pete" Dyballa)]

Replying to jeremyhu@…:

You can apply a .diff to the end to get the raw patch:

​https://github.com/XQuartz/xorg-server/commit/596efccbd9fa0a4b6c5ea64fc1f2e11a42c17e01.diff

The modification of this patch repair the defect.

[ballapete (Peter "Pete" Dyballa)]

The version that repaired the defect for xorg-server @1.17.4

[ballapete (Peter "Pete" Dyballa)]

Replying to Peter_Dyballa@…:

Replying to jeremyhu@…:

You can apply a .diff to the end to get the raw patch:

​https://github.com/XQuartz/xorg-server/commit/596efccbd9fa0a4b6c5ea64fc1f2e11a42c17e01.diff

The modification of this patch repair the defect.

On Snow Leopard, Mac OS X 10.6.8, I can still see the defect… I'll try the build once more.

[ballapete (Peter "Pete" Dyballa)]

Replying to Peter_Dyballa@…:

On Snow Leopard, Mac OS X 10.6.8, I can still see the defect… I'll try the build once more.

The cause is: port fetched for installation an archive from the internet!

I tested with port build xorg-server and invoked then port install xorg-server. Last night I performed a port archive xorg-server before installation, and this installed XQuartz 2.7.7 (xorg-server 1.17.4) without defects from my own archive.

[ballapete (Peter "Pete" Dyballa)]

What's puzzling me is why xorg-server @1.17.4 does not show this defect on PPC Tiger, Mac OS X 10.4.11. Just recently I built it and it works alright without patch.

[jeremyhu (Jeremy Huddleston Sequoia)]

Addressed in r143917.