rpms/xorg-x11-server/devel xserver-1.5.0-xkb-core-kbd-map-fix.patch, NONE, 1.1 xorg-x11-server.spec, 1.357, 1.358
Peter Hutterer
whot at fedoraproject.org
Tue Oct 7 04:06:10 UTC 2008
Author: whot
Update of /cvs/pkgs/rpms/xorg-x11-server/devel
In directory cvs1.fedora.phx.redhat.com:/tmp/cvs-serv22893
Modified Files:
xorg-x11-server.spec
Added Files:
xserver-1.5.0-xkb-core-kbd-map-fix.patch
Log Message:
* Tue Oct 7 2008 Peter Hutterer <peter.hutterer at redhat.com> 1.5.1-6
- xserver-1.5.0-xkb-core-kbd-map-fix.patch: don't invent groups when mapping
from xkb to core and back, and squash canonical types into explicit ones on
core reconstruction (2 patches). #460545
xserver-1.5.0-xkb-core-kbd-map-fix.patch:
--- NEW FILE xserver-1.5.0-xkb-core-kbd-map-fix.patch ---
>From 707503f8455da9cbbb97a030f242a7edf75526b6 Mon Sep 17 00:00:00 2001
From: Peter Hutterer <peter.hutterer at redhat.com>
Date: Fri, 19 Sep 2008 18:27:24 +0930
Subject: [PATCH] xkb: fix core keyboard map generation. #14373
According to Section 12.4 of the XKB Protocol Spec, if a key only has a single
group but the keyboard has multiple groups defined, the core description of
the key is a duplication of the single group across all symbols. i.e.
G1L1 G1L2 G1L1 G1L2 G1L3 G1L4 G1L3 G1L4
The previous code generated G1L1 G1L2 G1L3 G1L4 G1L3 G1L4, leading to
"invented" groups when the process is reversed.
Note that this creates wrong key types on reconstruction from core to xkb,
i.e. any single-group key with a key type that is not one of the canonical
four (Sec 12.2.3), will get the assigned type on group 1, and a canonical type
for the other gruops.
X.Org Bug 14373 <http://bugs.freedesktop.org/show_bug.cgi?id=14373>
---
xkb/xkbUtils.c | 39 ++++++++++++++++++++++++++++++++++-----
1 files changed, 34 insertions(+), 5 deletions(-)
diff --git a/xkb/xkbUtils.c b/xkb/xkbUtils.c
index 8339cef..b5c0ac2 100644
--- a/xkb/xkbUtils.c
+++ b/xkb/xkbUtils.c
@@ -486,6 +486,40 @@ CARD8 keysPerMod[XkbNumModifiers];
if (groupWidth>2)
nOut= groupWidth;
}
+
+ /* See XKB Protocol Sec, Section 12.4.
+ A 1-group key with ABCDE on a 2 group keyboard must be
+ duplicated across all groups as ABABCDECDE.
+ */
+ if (nGroups == 1)
+ {
+ int idx;
+
+ groupWidth = XkbKeyGroupWidth(xkb, key, XkbGroup1Index);
+
+ /* AB..CDE... -> ABABCDE... */
+ if (groupWidth > 0 && maxSymsPerKey >= 3)
+ pCore[2] = pCore[0];
+ if (groupWidth > 1 && maxSymsPerKey >= 4)
+ pCore[3] = pCore[1];
+
+ /* ABABCDE... -> ABABCDECDE */
+ idx = 2 + groupWidth;
+ while (groupWidth > 2 &&
+ idx < maxSymsPerKey &&
+ idx < groupWidth * 2)
+ {
+ pCore[idx] = pCore[idx - groupWidth + 2];
+ idx++;
+ }
+ idx = 2 * groupWidth;
+ if (idx < 4)
+ idx = 4;
+ /* 3 or more groups: ABABCDECDEABCDEABCDE */
+ for (n = 0; n < groupWidth && idx < maxSymsPerKey; n++)
+ pCore[idx++] = pXKB[n];
+ }
+
pXKB+= XkbKeyGroupsWidth(xkb,key);
nOut+= 2;
if (nGroups>1) {
@@ -507,11 +541,6 @@ CARD8 keysPerMod[XkbNumModifiers];
}
pXKB+= XkbKeyGroupsWidth(xkb,key);
}
- if (!pCore[2] && !pCore[3] && maxSymsPerKey >= 6 &&
- (pCore[4] || pCore[5])) {
- pCore[2] = pCore[4];
- pCore[3] = pCore[5];
- }
}
if (keyc->modifierMap[key]!=0) {
register unsigned bit,i,mask;
--
1.5.5.2
>From cd6b5a8da00219a1f14a3ab498af416296398483 Mon Sep 17 00:00:00 2001
From: Peter Hutterer <peter.hutterer at redhat.com>
Date: Mon, 22 Sep 2008 11:10:46 +0930
Subject: [PATCH] xkb: squash canonical types into explicit ones on core reconstruction.
If we update key types from core, and groups 2 - n have a canonical type but
the same symbols as the explicit type of group 1, assume that it was a core
sym duplication according to Section 12.4 of the XKB Protocol Spec.
Ignore the canonical types and pretend there's only one group for the key -
with the explicit key type.
---
xkb/XKBMisc.c | 13 ++++++++++---
1 files changed, 10 insertions(+), 3 deletions(-)
diff --git a/xkb/XKBMisc.c b/xkb/XKBMisc.c
index 85415a4..ac81395 100644
--- a/xkb/XKBMisc.c
+++ b/xkb/XKBMisc.c
@@ -178,16 +178,23 @@ int nGroups,tmp,groupsWidth;
}
}
- /* step 7: check for all groups identical or all width 1 */
+ /* step 7: check for all groups identical or all width 1
+ *
+ * Special feature: if group 1 has an explicit type and all other groups
+ * have canonical types with same symbols, we assume it's info lost from
+ * the core replication.
+ */
if (nGroups>1) {
- Bool sameType,allOneLevel;
+ Bool sameType,allOneLevel, canonical = True;
allOneLevel= (xkb->map->types[types_inout[0]].num_levels==1);
for (i=1,sameType=True;(allOneLevel||sameType)&&(i<nGroups);i++) {
sameType=(sameType&&(types_inout[i]==types_inout[XkbGroup1Index]));
if (allOneLevel)
allOneLevel= (xkb->map->types[types_inout[i]].num_levels==1);
+ if (types_inout[i] > XkbLastRequiredType)
+ canonical = False;
}
- if ((sameType)&&
+ if (((sameType) || canonical)&&
(!(protected&(XkbExplicitKeyTypesMask&~XkbExplicitKeyType1Mask)))){
register int s;
Bool identical;
--
1.5.5.2
Index: xorg-x11-server.spec
===================================================================
RCS file: /cvs/pkgs/rpms/xorg-x11-server/devel/xorg-x11-server.spec,v
retrieving revision 1.357
retrieving revision 1.358
diff -u -r1.357 -r1.358
--- xorg-x11-server.spec 6 Oct 2008 01:49:31 -0000 1.357
+++ xorg-x11-server.spec 7 Oct 2008 04:05:40 -0000 1.358
@@ -19,7 +19,7 @@
Summary: X.Org X11 X server
Name: xorg-x11-server
Version: 1.5.1
-Release: 5%{?dist}
+Release: 6%{?dist}
URL: http://www.x.org
License: MIT
Group: User Interface/X
@@ -79,6 +79,9 @@
Patch6000: xserver-1.5.0-hide-cursor.patch
Patch6001: xserver-1.5.0-edid-backport.patch
+# FDO bug 14373 (FIXED), RH bug #460545
+Patch6002: xserver-1.5.0-xkb-core-kbd-map-fix.patch
+
%define moduledir %{_libdir}/xorg/modules
%define drimoduledir %{_libdir}/dri
%define sdkdir %{_includedir}/xorg
@@ -509,6 +512,11 @@
%changelog
+* Tue Oct 7 2008 Peter Hutterer <peter.hutterer at redhat.com> 1.5.1-6
+- xserver-1.5.0-xkb-core-kbd-map-fix.patch: don't invent groups when mapping
+ from xkb to core and back, and squash canonical types into explicit ones on
+ core reconstruction (2 patches). #460545
+
* Mon Oct 06 2008 Dave Airlie <airlied at redhat.com> 1.5.1-5
- xserver-1.5.1-exa-fix-glyph-segfault.patch - fix EXA rects crash (462447)
More information about the fedora-extras-commits
mailing list