gitlab_xorg.atom.xml - sfeed_tests - sfeed tests and RSS and Atom files
(HTM) git clone git://git.codemadness.org/sfeed_tests
(DIR) Log
(DIR) Files
(DIR) Refs
(DIR) README
(DIR) LICENSE
---
gitlab_xorg.atom.xml (50248B)
---
1 <?xml version="1.0" encoding="UTF-8"?>
2 <feed xmlns="http://www.w3.org/2005/Atom" xmlns:media="http://search.yahoo.com/mrss/">
3 <title>xorg activity</title>
4 <link href="https://gitlab.freedesktop.org/xorg.atom" rel="self" type="application/atom+xml"/>
5 <link href="https://gitlab.freedesktop.org/xorg" rel="alternate" type="text/html"/>
6 <id>https://gitlab.freedesktop.org/xorg</id>
7 <updated>2020-10-19T15:36:23Z</updated>
8 <entry>
9 <id>tag:gitlab.freedesktop.org,2020-10-19:701030</id>
10 <link href="https://gitlab.freedesktop.org/xorg/driver/xf86-input-libinput/-/issues/31#note_666793"/>
11 <title>Trolli Schmittlauch commented on issue #31 at xorg / driver / xf86-input-libi...</title>
12 <updated>2020-10-19T15:36:23Z</updated>
13 <media:thumbnail width="40" height="40" url="https://secure.gravatar.com/avatar/aa9e73587abe4bc0bbab8258d51ba0c2?s=80&d=identicon"/>
14 <author>
15 <username>schmittlauch</username>
16 <name>Trolli Schmittlauch</name>
17 <email></email>
18 </author>
19 <summary type="xhtml">
20 <div xmlns="http://www.w3.org/1999/xhtml">
21 <p data-sourcepos="1:1-2:129" dir="auto">Unfortunately NixOS unstable is still at 0.28.2.
22 Do you think I can just upgrade the libinput driver without updating the rest of the xserver, or are there any interdependencies?</p>
23 </div>
24 </summary>
25 </entry>
26 <entry>
27 <id>tag:gitlab.freedesktop.org,2020-10-19:701006</id>
28 <link href="https://gitlab.freedesktop.org/xorg/xserver/-/issues/1087#note_666778"/>
29 <title>Tudor Brindus commented on issue #1087 at xorg / xserver</title>
30 <updated>2020-10-19T15:10:28Z</updated>
31 <media:thumbnail width="40" height="40" url="https://secure.gravatar.com/avatar/29c12d5cd9e0e6a2737ebe699edb20e3?s=80&d=identicon"/>
32 <author>
33 <username>Xyene</username>
34 <name>Tudor Brindus</name>
35 <email></email>
36 </author>
37 <summary type="xhtml">
38 <div xmlns="http://www.w3.org/1999/xhtml">
39 <p data-sourcepos="1:1-1:63" dir="auto">Ah, raced loading the page and didn't see the latest message...</p>
40 <blockquote data-sourcepos="3:1-3:212" dir="auto">
41 <p data-sourcepos="3:3-3:212">This would probably be <code>wl_pointer.leave</code> here though, as it's the surface being left the one interested in undoing its own confinement (plus, there's no guarantees that the next surface entered is Xwayland's).</p>
42 </blockquote>
43 <p data-sourcepos="5:1-5:309" dir="auto">I'm not sure it's strictly necessary to destroy the confinement on leave. Unless it's a one-shot, it can become active again when the app window is refocused. Switching from X11 app → Wayland app → same X11 app wouldn't need to destroy it, though X11 app → ... → <em>different</em> X11 app would.</p>
44 <p data-sourcepos="8:1-8:147" dir="auto">Of course, Xwayland could just unconditionally destroy and recreate the confine anyway. Not sure if there'd be any unintended side-effects of that.</p>
45 </div>
46 </summary>
47 </entry>
48 <entry>
49 <id>tag:gitlab.freedesktop.org,2020-10-19:700996</id>
50 <link href="https://gitlab.freedesktop.org/xorg/xserver/-/issues/1087#note_666774"/>
51 <title>Tudor Brindus commented on issue #1087 at xorg / xserver</title>
52 <updated>2020-10-19T15:00:22Z</updated>
53 <media:thumbnail width="40" height="40" url="https://secure.gravatar.com/avatar/29c12d5cd9e0e6a2737ebe699edb20e3?s=80&d=identicon"/>
54 <author>
55 <username>Xyene</username>
56 <name>Tudor Brindus</name>
57 <email></email>
58 </author>
59 <summary type="xhtml">
60 <div xmlns="http://www.w3.org/1999/xhtml">
61 <blockquote data-sourcepos="1:1-1:104" dir="auto">
62 <p data-sourcepos="1:3-1:104">Weird, outputs do not come into play at all for pointer locking and confinement emulation in Xwayland.</p>
63 </blockquote>
64 <p data-sourcepos="3:1-3:376" dir="auto">I suspect that the warping to center of the inactive-focused container that Sway does when focusing another output is the trigger, as it results in a <code>wl_pointer.enter</code> with no <code>wl_pointer.motion</code>. I imagine Alt-Tabbing a surface such that it appears under the pointer would be an analog on non-tiling WMs, but I haven't tried this. The issue title is bad here, in retrospect.</p>
65 <blockquote data-sourcepos="5:1-5:142" dir="auto">
66 <p data-sourcepos="5:3-5:142">xterm is a bit of a special case as it hides the cursor when typing, which sometimes confuses the pointer locking and confinement emulation.</p>
67 </blockquote>
68 <p data-sourcepos="7:1-7:138" dir="auto">You're right, of course. I ought to have picked a better example, but for the sake of being explicit: s/xterm/<any X11 app> still holds.</p>
69 <blockquote data-sourcepos="9:1-9:156" dir="auto">
70 <p data-sourcepos="9:3-9:156">None of those are in the stable branch that you run, so it might be interesting to try with Xwayland from git master to see if that helps with your issue.</p>
71 </blockquote>
72 <p data-sourcepos="11:1-11:380" dir="auto">I can give that a try tonight, and report back. I've definitely encountered <a href="https://gitlab.freedesktop.org/xorg/xserver/-/commit/1345f804a88efc11c58f8388983d34445d3e5928" data-original="https://gitlab.freedesktop.org/xorg/xserver/-/commit/1345f804a88efc11c58f8388983d34445d3e5928" data-link="false" data-link-reference="true" data-project="371" data-commit="1345f804a88efc11c58f8388983d34445d3e5928" data-reference-type="commit" data-container="body" data-placement="top" title="xwayland: confine motion events to the confined window" class="gfm gfm-commit has-tooltip">1345f804</a> before, as once the pointer is in the "broken" state, the app owning the confine continues receiving mouse motion (even though it's unfocused / on a different workspace / the pointer is on a different X11 app).</p>
73 </div>
74 </summary>
75 </entry>
76 <entry>
77 <id>tag:gitlab.freedesktop.org,2020-10-19:700937</id>
78 <link href="https://gitlab.freedesktop.org/xorg/xserver/-/issues/1087#note_666716"/>
79 <title>Carlos Garnacho commented on issue #1087 at xorg / xserver</title>
80 <updated>2020-10-19T14:16:48Z</updated>
81 <media:thumbnail width="40" height="40" url="https://secure.gravatar.com/avatar/b12f1f530e642125bfb5192072dcec2a?s=80&d=identicon"/>
82 <author>
83 <username>carlosg</username>
84 <name>Carlos Garnacho</name>
85 <email></email>
86 </author>
87 <summary type="xhtml">
88 <div xmlns="http://www.w3.org/1999/xhtml">
89 <blockquote data-sourcepos="1:1-1:295" dir="auto">
90 <p data-sourcepos="1:3-1:295">I haven't taken a look at the Xwayland code, but it sounds like it may be ignoring the initial movement that's part of <code>wl_pointer.enter</code> event when considering whether to destroy confines. wlroots, at least, does not send a <code>wl_pointer.motion</code> event until there is actual movement post-enter.</p>
91 </blockquote>
92 <p data-sourcepos="3:1-3:248" dir="auto">It sounds plausible. The only places where we maybe destroy the warp emulator are 1) motion events, 2) window is detroyed and 2) sprite cursor visibility changes. It seems pointer focus changes are not taken into account for undoing warp emulation.</p>
93 <p data-sourcepos="5:1-5:277" dir="auto">This would probably be <code>wl_pointer.leave</code> here though, as it's the surface being left the one interested in undoing its own confinement (plus, there's no guarantees that the next surface entered is Xwayland's). It would have saved me a couple extra reads of this issue, too :p.</p>
94 </div>
95 </summary>
96 </entry>
97 <entry>
98 <id>tag:gitlab.freedesktop.org,2020-10-19:700793</id>
99 <link href="https://gitlab.freedesktop.org/xorg/lib/libxkbfile/-/issues/2#note_666576"/>
100 <title>Dmitry Maevsky commented on issue #2 at xorg / lib / libxkbfile</title>
101 <updated>2020-10-19T11:43:51Z</updated>
102 <media:thumbnail width="40" height="40" url="https://secure.gravatar.com/avatar/c3f452fb49d187e5fef856ad083c0c14?s=80&d=identicon"/>
103 <author>
104 <username>dmitry.maevsky</username>
105 <name>Dmitry Maevsky</name>
106 <email></email>
107 </author>
108 <summary type="xhtml">
109 <div xmlns="http://www.w3.org/1999/xhtml">
110 <p data-sourcepos="1:1-1:102" dir="auto">It's 2020 :) I'm using Kubuntu 20.04, and I have precisely the same symptoms as described in this bug:</p>
111 <ul data-sourcepos="2:1-2:39" dir="auto">
112 <li data-sourcepos="2:1-2:39">I have redefined keys in xkb symbols:</li>
113 </ul>
114 <pre class="code highlight js-syntax-highlight plaintext" lang="plaintext" v-pre="true" xml:lang="plaintext"><code><span id="LC1" class="line" lang="plaintext" xml:lang="plaintext">partial modifier_keys</span>
115 <span id="LC2" class="line" lang="plaintext" xml:lang="plaintext">xkb_symbols "super_arrows" {</span>
116 <span id="LC3" class="line" lang="plaintext" xml:lang="plaintext"> key <LEFT> {</span>
117 <span id="LC4" class="line" lang="plaintext" xml:lang="plaintext"> type = "SHIFT_SUPER_4LEVELS",</span>
118 <span id="LC5" class="line" lang="plaintext" xml:lang="plaintext"> symbols[Group1] = [ Left, Left, Home, Home ],</span>
119 <span id="LC6" class="line" lang="plaintext" xml:lang="plaintext"> actions[Group1] = [ NoAction(), NoAction(), RedirectKey(keycode=<HOME>, clearmods=Mod4), RedirectKey(keycode=<HOME>, clearmods=Mod4) ]</span>
120 <span id="LC7" class="line" lang="plaintext" xml:lang="plaintext"> };</span>
121 <span id="LC8" class="line" lang="plaintext" xml:lang="plaintext"> key <RGHT> {</span>
122 <span id="LC9" class="line" lang="plaintext" xml:lang="plaintext"> type = "SHIFT_SUPER_4LEVELS",</span>
123 <span id="LC10" class="line" lang="plaintext" xml:lang="plaintext"> symbols[Group1] = [ Right, Right, End, End ],</span>
124 <span id="LC11" class="line" lang="plaintext" xml:lang="plaintext"> actions[Group1] = [ NoAction(), NoAction(), RedirectKey(keycode=<END>, clearmods=Mod4), RedirectKey(keycode=<END>, clearmods=Mod4) ]</span>
125 <span id="LC12" class="line" lang="plaintext" xml:lang="plaintext"> };</span>
126 <span id="LC13" class="line" lang="plaintext" xml:lang="plaintext"> key <UP> {</span>
127 <span id="LC14" class="line" lang="plaintext" xml:lang="plaintext"> type = "SHIFT_SUPER_4LEVELS",</span>
128 <span id="LC15" class="line" lang="plaintext" xml:lang="plaintext"> symbols[Group1] = [ Up, Up, Prior, Prior ],</span>
129 <span id="LC16" class="line" lang="plaintext" xml:lang="plaintext"> actions[Group1] = [ NoAction(), NoAction(), RedirectKey(keycode=<PGUP>, clearmods=Mod4), RedirectKey(keycode=<PGUP>, clearmods=Mod4) ]</span>
130 <span id="LC17" class="line" lang="plaintext" xml:lang="plaintext"> };</span>
131 <span id="LC18" class="line" lang="plaintext" xml:lang="plaintext"> key <DOWN> {</span>
132 <span id="LC19" class="line" lang="plaintext" xml:lang="plaintext"> type = "SHIFT_SUPER_4LEVELS",</span>
133 <span id="LC20" class="line" lang="plaintext" xml:lang="plaintext"> symbols[Group1] = [ Down, Down, Next, Next ],</span>
134 <span id="LC21" class="line" lang="plaintext" xml:lang="plaintext"> actions[Group1] = [ NoAction(), NoAction(), RedirectKey(keycode=<PGDN>, clearmods=Mod4), RedirectKey(keycode=<PGDN>, clearmods=Mod4) ]</span>
135 <span id="LC22" class="line" lang="plaintext" xml:lang="plaintext"> };</span>
136 <span id="LC23" class="line" lang="plaintext" xml:lang="plaintext"> key <INS> {</span>
137 <span id="LC24" class="line" lang="plaintext" xml:lang="plaintext"> symbols[Group1] = [ Print ],</span>
138 <span id="LC25" class="line" lang="plaintext" xml:lang="plaintext"> actions[Group1] = [ RedirectKey(keycode=<PRSC>) ]</span>
139 <span id="LC26" class="line" lang="plaintext" xml:lang="plaintext"> };</span>
140 <span id="LC27" class="line" lang="plaintext" xml:lang="plaintext">};</span></code></pre>
141 <p data-sourcepos="32:1-32:18" dir="auto">and my types file:</p>
142 <pre class="code highlight js-syntax-highlight plaintext" lang="plaintext" v-pre="true" xml:lang="plaintext"><code><span id="LC1" class="line" lang="plaintext" xml:lang="plaintext">partial default xkb_types "default" {</span>
143 <span id="LC2" class="line" lang="plaintext" xml:lang="plaintext"> type "SHIFT_SUPER_4LEVELS" {</span>
144 <span id="LC3" class="line" lang="plaintext" xml:lang="plaintext"> modifiers = Shift+Mod4;</span>
145 <span id="LC4" class="line" lang="plaintext" xml:lang="plaintext"> map[None] = Level1;</span>
146 <span id="LC5" class="line" lang="plaintext" xml:lang="plaintext"> map[Shift] = Level2;</span>
147 <span id="LC6" class="line" lang="plaintext" xml:lang="plaintext"> map[Mod4] = Level3;</span>
148 <span id="LC7" class="line" lang="plaintext" xml:lang="plaintext"> map[Shift+Mod4] = Level4;</span>
149 <span id="LC8" class="line" lang="plaintext" xml:lang="plaintext"> level_name[Level1] = "Base";</span>
150 <span id="LC9" class="line" lang="plaintext" xml:lang="plaintext"> level_name[Level2] = "Shift";</span>
151 <span id="LC10" class="line" lang="plaintext" xml:lang="plaintext"> level_name[Level3] = "Super";</span>
152 <span id="LC11" class="line" lang="plaintext" xml:lang="plaintext"> level_name[Level4] = "Shift+Super";</span>
153 <span id="LC12" class="line" lang="plaintext" xml:lang="plaintext"> };</span>
154 <span id="LC13" class="line" lang="plaintext" xml:lang="plaintext">};</span></code></pre>
155 <p data-sourcepos="49:1-50:75" dir="auto">Works perfectly when I'm loading it manually via <code>xkbcomp -I$HOME/.xkb ~/.xkb/keymap/mykbd $DISPLAY</code>
156 However, when included from /usr/share/X11/xkb, the keys do not autorepeat.</p>
157 <p data-sourcepos="52:1-52:25" dir="auto">Has it never been solved?</p>
158 </div>
159 </summary>
160 </entry>
161 <entry>
162 <id>tag:gitlab.freedesktop.org,2020-10-19:700609</id>
163 <link href="https://gitlab.freedesktop.org/xorg/xserver/-/issues/1087#note_666379"/>
164 <title>Olivier Fourdan commented on issue #1087 at xorg / xserver</title>
165 <updated>2020-10-19T09:05:19Z</updated>
166 <media:thumbnail width="40" height="40" url="https://secure.gravatar.com/avatar/5d8d8a23f53dfcfa7ad01c4877b2fe75?s=80&d=identicon"/>
167 <author>
168 <username>ofourdan</username>
169 <name>Olivier Fourdan</name>
170 <email></email>
171 </author>
172 <summary type="xhtml">
173 <div xmlns="http://www.w3.org/1999/xhtml">
174 <blockquote data-sourcepos="1:1-1:202" dir="auto">
175 <p data-sourcepos="1:3-1:202">Sometimes, Xwayland (1.20.9) doesn't destroy the pointer confine region after focusing a different Xwayland window, if that window was on a different output than the one the constraint was created on.</p>
176 </blockquote>
177 <p data-sourcepos="3:1-3:102" dir="auto">Weird, outputs do not come into play at all for pointer locking and confinement emulation in Xwayland.</p>
178 <blockquote data-sourcepos="5:1-5:43" dir="auto">
179 <p data-sourcepos="5:3-5:43">To reproduce, 3 workspaces are necessary.</p>
180 </blockquote>
181 <p data-sourcepos="7:1-7:95" dir="auto">Same as workspaces, those are a window manager/compositor concept completely alien to Xwayland.</p>
182 <blockquote data-sourcepos="9:1-9:76" dir="auto">
183 <p data-sourcepos="9:3-9:76">Move [the cursor] over Xterm, and notice that the cursor disappears again.</p>
184 </blockquote>
185 <p data-sourcepos="11:1-11:140" dir="auto">xterm is a bit of a special case as it hides the cursor when typing, which sometimes confuses the pointer locking and confinement emulation.</p>
186 <blockquote data-sourcepos="13:1-13:295" dir="auto">
187 <p data-sourcepos="13:3-13:295">I haven't taken a look at the Xwayland code, but it sounds like it may be ignoring the initial movement that's part of <code>wl_pointer.enter</code> event when considering whether to destroy confines. wlroots, at least, does not send a <code>wl_pointer.motion</code> event until there is actual movement post-enter.</p>
188 </blockquote>
189 <p data-sourcepos="15:1-15:131" dir="auto">Beware, that code is a minefield! <gl-emoji title="bomb" data-name="bomb" data-unicode-version="6.0">💣</gl-emoji> It is designed to work with some very odd cases and support games running in wine/proton.</p>
190 <p data-sourcepos="17:1-17:92" dir="auto">That reminds me of <a href="https://gitlab.freedesktop.org/xorg/xserver/-/issues/962" data-original="#962" data-link="false" data-link-reference="false" data-project="371" data-issue="25196" data-reference-type="issue" data-container="body" data-placement="top" title="xwayland: Confining the cursor (cursor grab) fails if the cursor is not over the window when the grab happens" class="gfm gfm-issue has-tooltip">#962 (closed)</a>. There are a few fixes in master which may come into play with this:</p>
191 <ul data-sourcepos="19:2-24:0" dir="auto">
192 <li data-sourcepos="19:2-19:18">commit <a href="https://gitlab.freedesktop.org/xorg/xserver/-/commit/5929b789f9c6531ee257504a7be9c9e3a49b30eb" data-original="5929b789" data-link="false" data-link-reference="false" data-project="371" data-commit="5929b789f9c6531ee257504a7be9c9e3a49b30eb" data-reference-type="commit" data-container="body" data-placement="top" title="xwayland: Do not lock the pointer on the wrong window" class="gfm gfm-commit has-tooltip">5929b789</a>
193 </li>
194 <li data-sourcepos="20:2-20:18">commit <a href="https://gitlab.freedesktop.org/xorg/xserver/-/commit/1345f804a88efc11c58f8388983d34445d3e5928" data-original="1345f804" data-link="false" data-link-reference="false" data-project="371" data-commit="1345f804a88efc11c58f8388983d34445d3e5928" data-reference-type="commit" data-container="body" data-placement="top" title="xwayland: confine motion events to the confined window" class="gfm gfm-commit has-tooltip">1345f804</a>
195 </li>
196 <li data-sourcepos="21:2-21:18">commit <a href="https://gitlab.freedesktop.org/xorg/xserver/-/commit/baa8d12e464664b5ad3c591be05a0087482790ca" data-original="baa8d12e" data-link="false" data-link-reference="false" data-project="371" data-commit="baa8d12e464664b5ad3c591be05a0087482790ca" data-reference-type="commit" data-container="body" data-placement="top" title="xwayland: Lock on entering surface if needed" class="gfm gfm-commit has-tooltip">baa8d12e</a>
197 </li>
198 <li data-sourcepos="22:2-22:18">commit <a href="https://gitlab.freedesktop.org/xorg/xserver/-/commit/f486e2fdaa1b252405a3aee90bd495b8b4c851f2" data-original="f486e2fd" data-link="false" data-link-reference="false" data-project="371" data-commit="f486e2fdaa1b252405a3aee90bd495b8b4c851f2" data-reference-type="commit" data-container="body" data-placement="top" title="xwayland: Remove undeeded test" class="gfm gfm-commit has-tooltip">f486e2fd</a>
199 </li>
200 <li data-sourcepos="23:2-24:0">commit <a href="https://gitlab.freedesktop.org/xorg/xserver/-/commit/0777cf46d7260f3d0c7fe82af5c94137d1e4f3de" data-original="0777cf46" data-link="false" data-link-reference="false" data-project="371" data-commit="0777cf46d7260f3d0c7fe82af5c94137d1e4f3de" data-reference-type="commit" data-container="body" data-placement="top" title="xwayland: Improve checks for confined_to on InputOnly windows" class="gfm gfm-commit has-tooltip">0777cf46</a>
201 </li>
202 </ul>
203 <p data-sourcepos="25:1-25:154" dir="auto">None of those are in the stable branch that you run, so it might be interesting to try with Xwayland from git master to see if that helps with your issue.</p>
204 <p data-sourcepos="27:1-27:12" dir="auto">CC: <a href="https://gitlab.freedesktop.org/carlosg" data-user="263" data-reference-type="user" data-container="body" data-placement="top" class="gfm gfm-project_member js-user-link" title="Carlos Garnacho">@carlosg</a></p>
205 </div>
206 </summary>
207 </entry>
208 <entry>
209 <id>tag:gitlab.freedesktop.org,2020-10-19:700483</id>
210 <link href="https://gitlab.freedesktop.org/xorg/xserver/-/merge_requests/527#note_666266"/>
211 <title>Peter Hutterer commented on merge request !527 at xorg / xserver</title>
212 <updated>2020-10-19T06:49:29Z</updated>
213 <media:thumbnail width="40" height="40" url="https://secure.gravatar.com/avatar/f1981449a295357c313b9c7640cc89fa?s=80&d=identicon"/>
214 <author>
215 <username>whot</username>
216 <name>Peter Hutterer</name>
217 <email></email>
218 </author>
219 <summary type="xhtml">
220 <div xmlns="http://www.w3.org/1999/xhtml">
221 <p data-sourcepos="1:1-1:211" dir="auto">staring at this again: yes, probably not needed. there's a very niche case where you're writing events one-by-one and thus after each <code>read()</code> there's more data waiting again. Not sure it's worth worrying about.</p>
222 </div>
223 </summary>
224 </entry>
225 <entry>
226 <id>tag:gitlab.freedesktop.org,2020-10-19:700480</id>
227 <link href="https://gitlab.freedesktop.org/xorg/xserver/-/merge_requests/527#note_666263"/>
228 <title>Peter Hutterer commented on merge request !527 at xorg / xserver</title>
229 <updated>2020-10-19T06:45:11Z</updated>
230 <media:thumbnail width="40" height="40" url="https://secure.gravatar.com/avatar/f1981449a295357c313b9c7640cc89fa?s=80&d=identicon"/>
231 <author>
232 <username>whot</username>
233 <name>Peter Hutterer</name>
234 <email></email>
235 </author>
236 <summary type="xhtml">
237 <div xmlns="http://www.w3.org/1999/xhtml">
238 <p data-sourcepos="1:1-1:344" dir="auto">yes, but: <code>open()</code> cannot ever return -1 as valid fd. <code>close()</code> must be able to handle an invalid fd be(<code>EBADF</code> is an explicit error condition, we're not talking about undefined behaviour here). Which means any implementation that cannot handle <code>close(-1)</code> would struggle to pass the specs. Take this with my non-existing lawyer party hat on ;)</p>
239 </div>
240 </summary>
241 </entry>
242 <entry>
243 <id>tag:gitlab.freedesktop.org,2020-10-19:700460</id>
244 <link href="https://gitlab.freedesktop.org/xorg/driver/xf86-input-libinput/-/merge_requests/16#note_666251"/>
245 <title>Peter Hutterer commented on merge request !16 at xorg / driver / xf86-input-l...</title>
246 <updated>2020-10-19T05:28:42Z</updated>
247 <media:thumbnail width="40" height="40" url="https://secure.gravatar.com/avatar/f1981449a295357c313b9c7640cc89fa?s=80&d=identicon"/>
248 <author>
249 <username>whot</username>
250 <name>Peter Hutterer</name>
251 <email></email>
252 </author>
253 <summary type="xhtml">
254 <div xmlns="http://www.w3.org/1999/xhtml">
255 <p data-sourcepos="1:1-1:125" dir="auto">same seems to be the case for a few other <code>xf86libinput_init_</code> functions, can you fix those up in the same MR please? Thanks.</p>
256 </div>
257 </summary>
258 </entry>
259 <entry>
260 <id>tag:gitlab.freedesktop.org,2020-10-19:700443</id>
261 <link href="https://gitlab.freedesktop.org/xorg/driver/xf86-input-libinput/-/issues/31#note_666238"/>
262 <title>Peter Hutterer commented on issue #31 at xorg / driver / xf86-input-libinput</title>
263 <updated>2020-10-19T04:46:20Z</updated>
264 <media:thumbnail width="40" height="40" url="https://secure.gravatar.com/avatar/f1981449a295357c313b9c7640cc89fa?s=80&d=identicon"/>
265 <author>
266 <username>whot</username>
267 <name>Peter Hutterer</name>
268 <email></email>
269 </author>
270 <summary type="xhtml">
271 <div xmlns="http://www.w3.org/1999/xhtml">
272 <p data-sourcepos="1:1-1:52" dir="auto">looks like <a href="https://gitlab.freedesktop.org/xorg/driver/xf86-input-libinput/-/issues/25" data-original="#25" data-link="false" data-link-reference="false" data-project="577" data-issue="13493" data-reference-type="issue" data-container="body" data-placement="top" title="Wacom Eraser not recognised properly" class="gfm gfm-issue has-tooltip">#25 (closed)</a> which was fixed with 0.29.0 and later</p>
273 </div>
274 </summary>
275 </entry>
276 <entry>
277 <id>tag:gitlab.freedesktop.org,2020-10-18:700349</id>
278 <link href="https://gitlab.freedesktop.org/xorg/lib/libxt/-/merge_requests/45#note_666163"/>
279 <title>Thomas E. Dickey commented on merge request !45 at xorg / lib / libXt</title>
280 <updated>2020-10-18T17:55:34Z</updated>
281 <media:thumbnail width="40" height="40" url="https://gitlab.freedesktop.org/uploads/-/system/user/avatar/4137/avatar.png"/>
282 <author>
283 <username>dickey</username>
284 <name>Thomas E. Dickey</name>
285 <email></email>
286 </author>
287 <summary type="xhtml">
288 <div xmlns="http://www.w3.org/1999/xhtml">
289 <p data-sourcepos="1:1-1:207" dir="auto">From the way this is described, I would not expect this <a href="https://invisible-island.net/xterm/manpage/xterm.html#h3-Default-Key-Bindings" rel="nofollow noreferrer noopener" target="_blank">default-binding</a> to work (since <em>Ctrl</em> is not mentioned in the change):</p>
290 <pre class="code highlight js-syntax-highlight plaintext" lang="plaintext" v-pre="true" xml:lang="plaintext"><code><span id="LC1" class="line" lang="plaintext" xml:lang="plaintext"> Shift~Ctrl <KeyPress> KP_Add:larger-vt-font() \n\</span>
291 <span id="LC2" class="line" lang="plaintext" xml:lang="plaintext"> Shift Ctrl <KeyPress> KP_Add:smaller-vt-font() \n\</span>
292 <span id="LC3" class="line" lang="plaintext" xml:lang="plaintext"> Shift <KeyPress> KP_Subtract:smaller-vt-font() \n\</span></code></pre>
293 <p data-sourcepos="7:1-7:181" dir="auto">However, in testing the change, that appears to work (and there is no difference between the old/new values of <code>useful_mods</code>). But the lengthy comment for the change is misleading.</p>
294 </div>
295 </summary>
296 </entry>
297 <entry>
298 <id>tag:gitlab.freedesktop.org,2020-10-18:700164</id>
299 <link href="https://gitlab.freedesktop.org/xorg/driver/xf86-video-amdgpu/-/issues/17#note_656539"/>
300 <title>Julian Hille commented on issue #17 at xorg / driver / xf86-video-amdgpu</title>
301 <updated>2020-10-18T10:02:43Z</updated>
302 <media:thumbnail width="40" height="40" url="https://secure.gravatar.com/avatar/f772af92c9358221d666941ccf8373ec?s=80&d=identicon"/>
303 <author>
304 <username>julianhille</username>
305 <name>Julian Hille</name>
306 <email></email>
307 </author>
308 <summary type="xhtml">
309 <div xmlns="http://www.w3.org/1999/xhtml">
310 <p data-sourcepos="1:1-1:41" dir="auto">I've seen this changelog on kernel 5.9.1:</p>
311 <blockquote data-sourcepos="2:1-3:71" dir="auto">
312 <p data-sourcepos="2:3-3:71">Alex Deucher (1):
313 Revert "drm/amdgpu: Fix NULL dereference in dpm sysfs handlers"</p>
314 </blockquote>
315 <p data-sourcepos="5:1-5:26" dir="auto">Will test and report back.</p>
316 </div>
317 </summary>
318 </entry>
319 <entry>
320 <id>tag:gitlab.freedesktop.org,2020-10-17:700089</id>
321 <link href="https://gitlab.freedesktop.org/xorg/lib/libxfont/-/issues/10"/>
322 <title>Ulrich Sibiller opened issue #10: memory leak in FreeTypeGetInfoScalable/Free...</title>
323 <updated>2020-10-17T23:28:44Z</updated>
324 <media:thumbnail width="40" height="40" url="https://secure.gravatar.com/avatar/9317eef2e7cdd7b3d0cd287a3e63a153?s=80&d=identicon"/>
325 <author>
326 <username>uli42</username>
327 <name>Ulrich Sibiller</name>
328 <email></email>
329 </author>
330 <summary type="xhtml">
331 <div xmlns="http://www.w3.org/1999/xhtml">
332 <p dir="auto">FreeTypeLoadXFont()'s third parameter is a FontPtr called "xf" and it is used to store the allocated FTFontPtr:</p>
<p dir="auto"><a href="https://gitlab.freedesktop.org/xorg/lib/libxfont/-/blob/master/src/FreeType/ftfuncs.c#L3615">https://gitlab.freedesktop.org/xorg/lib/libxfont/-/blob/master/src/FreeType/ftfuncs.c#L3615</a></p>
<p dir="auto">FreeTypeGetInfoScalable() calls FreeTypeLoadXFont(..,..,0,....). This results in the allocated FTFontPtr not being stored anywhere within FreeTypeLoadXFont() and therefore being lost.</p>
333 </div>
334 </summary>
335 </entry>
336 <entry>
337 <id>tag:gitlab.freedesktop.org,2020-10-17:700079</id>
338 <link href="https://gitlab.freedesktop.org/xorg/xserver/-/issues/1087"/>
339 <title>Tudor Brindus opened issue #1087: Xwayland pointer constraints get confused b...</title>
340 <updated>2020-10-17T22:36:41Z</updated>
341 <media:thumbnail width="40" height="40" url="https://secure.gravatar.com/avatar/29c12d5cd9e0e6a2737ebe699edb20e3?s=80&d=identicon"/>
342 <author>
343 <username>Xyene</username>
344 <name>Tudor Brindus</name>
345 <email></email>
346 </author>
347 <summary type="xhtml">
348 <div xmlns="http://www.w3.org/1999/xhtml">
349 <p dir="auto">Hi,</p>
<p dir="auto">Sometimes, Xwayland (1.20.9) doesn't destroy the pointer confine region after focusing a different Xwayland window, if that window was on a different output than the one the constraint was created on. Reproduction steps are a bit icky, but should hopefully make clear what goes wrong.</p>
<p dir="auto">I'm using sway/wlroots master, but from the logs I <em>believe</em> this would happen on other compositors too.</p>
<p dir="auto">To reproduce, 3 workspaces are necessary.</p>
<p dir="auto">Workspace 1 contains an instance of Xterm, and for convenience an instance of another terminal tailing the Sway log file (<code>-d</code> turned on for debug output; I tail and <code>| grep --line-buffered -E '(pointer|cursor)'</code>). <strong>The Xterm window should be the one focused here.</strong></p>
<p dir="auto"><a class="no-attachment-icon gfm" href="https://gitlab.freedesktop.org/xorg/xserver/uploads/34639adf52e298ed67d81ba47d2ea427/image.png" target="_blank" rel="noopener noreferrer" data-link="true"><img src="data:image/gif;base64,R0lGODlhAQABAAAAACH5BAEKAAEALAAAAAABAAEAAAICTAEAOw==" alt="image" class="lazy gfm" data-src="https://gitlab.freedesktop.org/xorg/xserver/uploads/34639adf52e298ed67d81ba47d2ea427/image.png" /></a></p>
<p dir="auto">Workspace 2 should be <strong>on another output</strong> and contain a single Xwayland application that captures the mouse. I use <a href="https://github.com/gnif/LookingGlass" rel="nofollow noreferrer noopener" target="_blank">https://github.com/gnif/LookingGlass</a> for this (it's a VM utility, but the bug manifests even if there is no VM running, so it seems like a convenient reproduction).</p>
<p dir="auto">Start LookingGlass with <code>looking-glass-client spice:captureOnStart=true</code>.</p>
<p dir="auto"><a class="no-attachment-icon gfm" href="https://gitlab.freedesktop.org/xorg/xserver/uploads/c1c711344b797c018da13e1809a8eaf9/image.png" target="_blank" rel="noopener noreferrer" data-link="true"><img src="data:image/gif;base64,R0lGODlhAQABAAAAACH5BAEKAAEALAAAAAABAAEAAAICTAEAOw==" alt="image" class="lazy gfm" data-src="https://gitlab.freedesktop.org/xorg/xserver/uploads/c1c711344b797c018da13e1809a8eaf9/image.png" /></a></p>
<p dir="auto">Workspace 3 should be <strong>on the same output as workspace 2</strong>, and contain two applications: one Wayland-native, and one Xwayland. I used Kitty and Xterm again. <strong>Focus the Wayland-native app, not Xterm.</strong></p>
<p dir="auto"><a class="no-attachment-icon gfm" href="https://gitlab.freedesktop.org/xorg/xserver/uploads/2c5df03d0ee5158eefc8171c9447a3bf/image.png" target="_blank" rel="noopener noreferrer" data-link="true"><img src="data:image/gif;base64,R0lGODlhAQABAAAAACH5BAEKAAEALAAAAAABAAEAAAICTAEAOw==" alt="image" class="lazy gfm" data-src="https://gitlab.freedesktop.org/xorg/xserver/uploads/2c5df03d0ee5158eefc8171c9447a3bf/image.png" /></a></p>
<p dir="auto">Now, to reproduce:</p>
<ol dir="auto">
<li>Focus LookingGlass in workspace 2. Verify that the pointer is captured.</li>
</ol>
<pre class="code highlight js-syntax-highlight plaintext" lang="plaintext" v-pre="true" xml:lang="plaintext"><code><span id="LC1" class="line" lang="plaintext" xml:lang="plaintext">02:08:58.563 [DEBUG] [types/wlr_pointer_constraints_v1.c:244] new locked_pointer 0x557285200a40 (res 0x5572850577b0)</span>
<span id="LC2" class="line" lang="plaintext" xml:lang="plaintext">02:08:58.563 [DEBUG] [types/wlr_pointer_constraints_v1.c:343] constrained 0x557285200a40</span></code></pre>
<ol start="2" dir="auto">
<li>Use keybinding to switch to workspace 1. <strong>Do not move the mouse.</strong> Observe that Xterm is focused, but that the cursor has not appeared.</li>
</ol>
<pre class="code highlight js-syntax-highlight plaintext" lang="plaintext" v-pre="true" xml:lang="plaintext"><code><span id="LC1" class="line" lang="plaintext" xml:lang="plaintext">02:09:05.027 [DEBUG] [types/wlr_pointer_constraints_v1.c:353] unconstrained 0x557285200a40</span></code></pre>
<ol start="3" dir="auto">
<li>Use keybinding to switch to workspace 3. The cursor will have warped to the middle of the Wayland-native application. Move it over Xterm, and notice that the cursor disappears again.</li>
<li>Click on Xterm. The cursor re-appears, and the pointer confine is finally destroyed.</li>
</ol>
<pre class="code highlight js-syntax-highlight plaintext" lang="plaintext" v-pre="true" xml:lang="plaintext"><code><span id="LC1" class="line" lang="plaintext" xml:lang="plaintext">02:12:10.674 [DEBUG] [types/seat/wlr_seat_pointer.c:384] button_count=1 grab_serial=40811 serial=40838</span>
<span id="LC2" class="line" lang="plaintext" xml:lang="plaintext">02:12:10.674 [DEBUG] [types/wlr_pointer_constraints_v1.c:47] destroying constraint 0x5572852236f0</span></code></pre>
<p dir="auto">This is in contrast to "normal" behavior, where Xwayland destroys the confine during the first motion event on another Xwayland surface. Indeed, during step 2), if the mouse is moved then the confine is indeed destroyed and the bugs in 3) no longer happen. However, moving the mouse over an Xterm in 3) no longer triggers the confine destruction.</p>
<p dir="auto">I haven't taken a look at the Xwayland code, but it sounds like it may be ignoring the initial movement that's part of <code>wl_pointer.enter</code> event when considering whether to destroy confines. wlroots, at least, does not send a <code>wl_pointer.motion</code> event until there is actual movement post-enter.</p>
350 </div>
351 </summary>
352 </entry>
353 <entry>
354 <id>tag:gitlab.freedesktop.org,2020-10-16:699700</id>
355 <link href="https://gitlab.freedesktop.org/xorg/xserver/-/issues/1086"/>
356 <title>Hoernchen opened issue #1086: CorePointer &amp; CoreKeyboard ignored in ServerLay...</title>
357 <updated>2020-10-16T22:18:50Z</updated>
358 <media:thumbnail width="40" height="40" url="https://secure.gravatar.com/avatar/09fb001e9444e0f1d15ca0cf3ed599c5?s=80&d=identicon"/>
359 <author>
360 <username>Hoernchen</username>
361 <name>Hoernchen</name>
362 <email></email>
363 </author>
364 <summary type="xhtml">
365 <div xmlns="http://www.w3.org/1999/xhtml">
366 <p dir="auto"><strong>(known workaround at the end of this issue description)</strong></p>
<p dir="auto">I'm trying to use xrdp on ubuntu 20.04 with xorg-xserver 1.20.8-2ubuntu2.4 - but the supplied config file results in no mouse or keyboard in the session (both provided by xorgxrdp drivers), even though the config appears to match what the xorg.conf manpage says:</p>
<p dir="auto">relevant excerpt from <a href="https://github.com/neutrinolabs/xorgxrdp/blob/devel/xrdpdev/xorg.conf" rel="nofollow noreferrer noopener" target="_blank">https://github.com/neutrinolabs/xorgxrdp/blob/devel/xrdpdev/xorg.conf</a> as distributed by ubuntu:</p>
<pre class="code highlight js-syntax-highlight plaintext" lang="plaintext" v-pre="true" xml:lang="plaintext"><code><span id="LC1" class="line" lang="plaintext" xml:lang="plaintext">Section "ServerLayout"</span>
<span id="LC2" class="line" lang="plaintext" xml:lang="plaintext"> Identifier "X11 Server"</span>
<span id="LC3" class="line" lang="plaintext" xml:lang="plaintext"> Screen "Screen (xrdpdev)"</span>
<span id="LC4" class="line" lang="plaintext" xml:lang="plaintext"> InputDevice "xrdpMouse" "CorePointer" <------------------------------</span>
<span id="LC5" class="line" lang="plaintext" xml:lang="plaintext"> InputDevice "xrdpKeyboard" "CoreKeyboard" <--------------------------</span>
<span id="LC6" class="line" lang="plaintext" xml:lang="plaintext">EndSection</span>
<span id="LC7" class="line" lang="plaintext" xml:lang="plaintext"></span>
<span id="LC8" class="line" lang="plaintext" xml:lang="plaintext">Section "ServerFlags"</span>
<span id="LC9" class="line" lang="plaintext" xml:lang="plaintext"> Option "DontVTSwitch" "on"</span>
<span id="LC10" class="line" lang="plaintext" xml:lang="plaintext"> Option "AutoAddDevices" "off"</span>
<span id="LC11" class="line" lang="plaintext" xml:lang="plaintext">EndSection</span>
<span id="LC12" class="line" lang="plaintext" xml:lang="plaintext"></span>
<span id="LC13" class="line" lang="plaintext" xml:lang="plaintext">Section "Module"</span>
<span id="LC14" class="line" lang="plaintext" xml:lang="plaintext"> Load "dbe"</span>
<span id="LC15" class="line" lang="plaintext" xml:lang="plaintext"> Load "ddc"</span>
<span id="LC16" class="line" lang="plaintext" xml:lang="plaintext"> Load "extmod"</span>
<span id="LC17" class="line" lang="plaintext" xml:lang="plaintext"> Load "glx"</span>
<span id="LC18" class="line" lang="plaintext" xml:lang="plaintext"> Load "int10"</span>
<span id="LC19" class="line" lang="plaintext" xml:lang="plaintext"> Load "record"</span>
<span id="LC20" class="line" lang="plaintext" xml:lang="plaintext"> Load "vbe"</span>
<span id="LC21" class="line" lang="plaintext" xml:lang="plaintext"> Load "xorgxrdp"</span>
<span id="LC22" class="line" lang="plaintext" xml:lang="plaintext"> Load "fb"</span>
<span id="LC23" class="line" lang="plaintext" xml:lang="plaintext">EndSection</span>
<span id="LC24" class="line" lang="plaintext" xml:lang="plaintext"></span>
<span id="LC25" class="line" lang="plaintext" xml:lang="plaintext">Section "InputDevice"</span>
<span id="LC26" class="line" lang="plaintext" xml:lang="plaintext"> Identifier "xrdpKeyboard"</span>
<span id="LC27" class="line" lang="plaintext" xml:lang="plaintext"> Driver "xrdpkeyb"</span>
<span id="LC28" class="line" lang="plaintext" xml:lang="plaintext">EndSection</span>
<span id="LC29" class="line" lang="plaintext" xml:lang="plaintext"></span>
<span id="LC30" class="line" lang="plaintext" xml:lang="plaintext">Section "InputDevice"</span>
<span id="LC31" class="line" lang="plaintext" xml:lang="plaintext"> Identifier "xrdpMouse"</span>
<span id="LC32" class="line" lang="plaintext" xml:lang="plaintext"> Driver "xrdpmouse"</span>
<span id="LC33" class="line" lang="plaintext" xml:lang="plaintext">EndSection</span>
<span id="LC34" class="line" lang="plaintext" xml:lang="plaintext"></span></code></pre>
<p dir="auto">The <a href="https://www.x.org/releases/current/doc/man/man5/xorg.conf.5.xhtml" rel="nofollow noreferrer noopener" target="_blank">manpage </a> states :</p>
<pre class="code highlight js-syntax-highlight plaintext" lang="plaintext" v-pre="true" xml:lang="plaintext"><code><span id="LC1" class="line" lang="plaintext" xml:lang="plaintext">SERVERLAYOUT SECTION</span>
<span id="LC2" class="line" lang="plaintext" xml:lang="plaintext"></span>
<span id="LC3" class="line" lang="plaintext" xml:lang="plaintext">-- snip 8< ----</span>
<span id="LC4" class="line" lang="plaintext" xml:lang="plaintext"></span>
<span id="LC5" class="line" lang="plaintext" xml:lang="plaintext">InputDevice "idev−id" "option" ...</span>
<span id="LC6" class="line" lang="plaintext" xml:lang="plaintext"></span>
<span id="LC7" class="line" lang="plaintext" xml:lang="plaintext">-- snip 8< ----</span>
<span id="LC8" class="line" lang="plaintext" xml:lang="plaintext">The options permitted here are any that may also be given in the InputDevice sections. Normally only session−specific input device options would be used here. The most commonly used options are:</span>
<span id="LC9" class="line" lang="plaintext" xml:lang="plaintext"></span>
<span id="LC10" class="line" lang="plaintext" xml:lang="plaintext">"CorePointer"</span>
<span id="LC11" class="line" lang="plaintext" xml:lang="plaintext">"CoreKeyboard"</span>
<span id="LC12" class="line" lang="plaintext" xml:lang="plaintext">"SendCoreEvents"</span>
<span id="LC13" class="line" lang="plaintext" xml:lang="plaintext"></span>
<span id="LC14" class="line" lang="plaintext" xml:lang="plaintext">-- snip 8< ---</span>
<span id="LC15" class="line" lang="plaintext" xml:lang="plaintext"></span>
<span id="LC16" class="line" lang="plaintext" xml:lang="plaintext">Here is an example of a ServerLayout section for a dual headed configuration with two mice:</span>
<span id="LC17" class="line" lang="plaintext" xml:lang="plaintext"></span>
<span id="LC18" class="line" lang="plaintext" xml:lang="plaintext">Section "ServerLayout"</span>
<span id="LC19" class="line" lang="plaintext" xml:lang="plaintext">Identifier "Layout 1"</span>
<span id="LC20" class="line" lang="plaintext" xml:lang="plaintext">Screen "MGA 1"</span>
<span id="LC21" class="line" lang="plaintext" xml:lang="plaintext">Screen "MGA 2" RightOf "MGA 1"</span>
<span id="LC22" class="line" lang="plaintext" xml:lang="plaintext">InputDevice "Keyboard 1" "CoreKeyboard" <--------------------------</span>
<span id="LC23" class="line" lang="plaintext" xml:lang="plaintext">InputDevice "Mouse 1" "CorePointer" <---------------------------</span>
<span id="LC24" class="line" lang="plaintext" xml:lang="plaintext">InputDevice "Mouse 2" "SendCoreEvents"</span>
<span id="LC25" class="line" lang="plaintext" xml:lang="plaintext">Option "BlankTime" "5"</span>
<span id="LC26" class="line" lang="plaintext" xml:lang="plaintext">EndSection</span>
<span id="LC27" class="line" lang="plaintext" xml:lang="plaintext"></span></code></pre>
<p dir="auto">So the config file agrees with the manpage - but this leads to <code><default pointer></code> and <code><default keyboard></code> being used, since no CorePointer and CoreKeyboard is found and therefore a unusable rdp session, <strong>unless</strong> I repeat the CorePointer and CoreKeyboard options in the InputDevice sections like this, which fixes the issue and makes it work as expected:</p>
<pre class="code highlight js-syntax-highlight plaintext" lang="plaintext" v-pre="true" xml:lang="plaintext"><code><span id="LC1" class="line" lang="plaintext" xml:lang="plaintext">Section "InputDevice"</span>
<span id="LC2" class="line" lang="plaintext" xml:lang="plaintext"> Identifier "xrdpKeyboard"</span>
<span id="LC3" class="line" lang="plaintext" xml:lang="plaintext"> Driver "xrdpkeyb"</span>
<span id="LC4" class="line" lang="plaintext" xml:lang="plaintext"> Option "CoreKeyboard"</span>
<span id="LC5" class="line" lang="plaintext" xml:lang="plaintext">EndSection</span>
<span id="LC6" class="line" lang="plaintext" xml:lang="plaintext"></span>
<span id="LC7" class="line" lang="plaintext" xml:lang="plaintext">Section "InputDevice"</span>
<span id="LC8" class="line" lang="plaintext" xml:lang="plaintext"> Identifier "xrdpMouse"</span>
<span id="LC9" class="line" lang="plaintext" xml:lang="plaintext"> Driver "xrdpmouse"</span>
<span id="LC10" class="line" lang="plaintext" xml:lang="plaintext"> Option "CorePointer"</span>
<span id="LC11" class="line" lang="plaintext" xml:lang="plaintext">EndSection</span></code></pre>
<p dir="auto">Is this a bug? Is the config file parsing broken? Is the manpage wrong?
The config file is quite old and has been distributed like that for years, so.. what exactly is broken here?
I've also had a look at xf86Config.c but the code is years old as well and appears to agree with what the manpage and the original config contains.</p>
367 </div>
368 </summary>
369 </entry>
370 <entry>
371 <id>tag:gitlab.freedesktop.org,2020-10-16:699659</id>
372 <link href="https://gitlab.freedesktop.org/xorg/driver/xf86-video-amdgpu/-/issues/17#note_656102"/>
373 <title>Julian Hille commented on issue #17 at xorg / driver / xf86-video-amdgpu</title>
374 <updated>2020-10-16T20:55:10Z</updated>
375 <media:thumbnail width="40" height="40" url="https://secure.gravatar.com/avatar/f772af92c9358221d666941ccf8373ec?s=80&d=identicon"/>
376 <author>
377 <username>julianhille</username>
378 <name>Julian Hille</name>
379 <email></email>
380 </author>
381 <summary type="xhtml">
382 <div xmlns="http://www.w3.org/1999/xhtml">
383 <p data-sourcepos="1:1-1:124" dir="auto">I'm on a lenovo T14 and undocking leads to the same issue, let me know if it would be of any help if i attach similar infos.</p>
384 </div>
385 </summary>
386 </entry>
387 <entry>
388 <id>tag:gitlab.freedesktop.org,2020-10-16:699278</id>
389 <link href="https://gitlab.freedesktop.org/xorg/lib/libxcb/-/issues/38#note_655730"/>
390 <title>Remi Denis-Courmont commented on issue #38 at xorg / lib / libxcb</title>
391 <updated>2020-10-16T15:36:32Z</updated>
392 <media:thumbnail width="40" height="40" url="https://secure.gravatar.com/avatar/10e418e48350379fc7f4113116872943?s=80&d=identicon"/>
393 <author>
394 <username>RemiDenis</username>
395 <name>Remi Denis-Courmont</name>
396 <email></email>
397 </author>
398 <summary type="xhtml">
399 <div xmlns="http://www.w3.org/1999/xhtml">
400 <p data-sourcepos="1:1-1:338" dir="auto">The client (in this case, Mesa) should subscribe to, and look out for the CloseNotify event (from the base X11 protocol) on the window. Timing out instead seems likely to break other use cases and cause weird side effects, such as causing spurious failures when a process is stopped for debugging, power management, I/O spike or whatever.</p>
401 <p data-sourcepos="3:1-3:108" dir="auto">That should work fine if the XCB connection is only used by Mesa, and does not require any change to libxcb.</p>
402 <p data-sourcepos="5:1-5:538" dir="auto">If it is shared by multiple threads, then there is a problem that libxcb has no race-free means to wait for one of several event (at least that I'd know). If the XCB connection is shared by the creator of the window, then an extra problem is that the window destroy event will be sent only once per connection, and there would be competition between Mesa and the GUI framework to handle it. (In my experience, sharing the X client connection with the GL driver is just a recipe for trouble; that's why VLC creates a dedicated connection.)</p>
403 </div>
404 </summary>
405 </entry>
406 <entry>
407 <id>tag:gitlab.freedesktop.org,2020-10-16:698977</id>
408 <link href="https://gitlab.freedesktop.org/xorg/driver/xf86-video-intel/-/commit/67f15b36faf8c4aa902d953067df0faf43387208"/>
409 <title>Chris Wilson pushed to project branch master at xorg / driver / xf86-video-intel</title>
410 <updated>2020-10-16T12:14:56Z</updated>
411 <media:thumbnail width="40" height="40" url="https://secure.gravatar.com/avatar/b370d944f8518e3d9778130fcfd11f5d?s=80&d=identicon"/>
412 <author>
413 <username>ickle</username>
414 <name>Chris Wilson</name>
415 <email></email>
416 </author>
417 <summary type="xhtml">
418 <div xmlns="http://www.w3.org/1999/xhtml">
419 <p>
420 <strong>Chris Wilson</strong>
421 <a href="https://gitlab.freedesktop.org/xorg/driver/xf86-video-intel/-/commit/67f15b36faf8c4aa902d953067df0faf43387208">(67f15b36)</a>
422 <i>
423 at
424 16 Oct 12:14
425 </i>
426 </p>
427 <div class="blockquote"><p dir="auto">sna/gen7: Prefer blitter for plain copies on Haswell</p></div>
428 </div>
429 </summary>
430 </entry>
431 <entry>
432 <id>tag:gitlab.freedesktop.org,2020-10-16:698853</id>
433 <link href="https://gitlab.freedesktop.org/xorg/lib/libxcb/-/merge_requests/9#note_655369"/>
434 <title>Thomas Guillem commented on merge request !9 at xorg / lib / libxcb</title>
435 <updated>2020-10-16T10:12:01Z</updated>
436 <media:thumbnail width="40" height="40" url="https://secure.gravatar.com/avatar/4116f2a3dd20ed95c09bd89eeff2ddc1?s=80&d=identicon"/>
437 <author>
438 <username>tguillem</username>
439 <name>Thomas Guillem</name>
440 <email></email>
441 </author>
442 <summary type="xhtml">
443 <div xmlns="http://www.w3.org/1999/xhtml">
444 <p data-sourcepos="1:1-1:6" dir="auto">Hello,</p>
445 <p data-sourcepos="3:1-3:36" dir="auto">Thanks Michel and Uli for your work.</p>
446 <p data-sourcepos="5:1-5:68" dir="auto">VLC developer here. I tried to fix the comments addressed by Michel.</p>
447 <p data-sourcepos="7:1-7:120" dir="auto">Here is my fixup patch: <a href="https://gitlab.freedesktop.org/tguillem/libxcb/-/commit/d78820c7b1963dd07c79008a7b3155dfbf92ca6f" data-original="https://gitlab.freedesktop.org/tguillem/libxcb/-/commit/d78820c7b1963dd07c79008a7b3155dfbf92ca6f" data-link="false" data-link-reference="true" data-project="7854" data-commit="d78820c7b1963dd07c79008a7b3155dfbf92ca6f" data-reference-type="commit" data-container="body" data-placement="top" title="fixup! WIP: xcb_wait_for_special_event_with_timeout()" class="gfm gfm-commit has-tooltip">tguillem/libxcb@d78820c7</a></p>
448 <p data-sourcepos="9:1-9:146" dir="auto">I could test it, using the following patch on mesa: <a href="https://gitlab.freedesktop.org/tguillem/mesa/-/commit/ea93da11943a73dd58c001cb266dc1c03c700dc7" data-original="https://gitlab.freedesktop.org/tguillem/mesa/-/commit/ea93da11943a73dd58c001cb266dc1c03c700dc7" data-link="false" data-link-reference="true" data-project="7855" data-commit="ea93da11943a73dd58c001cb266dc1c03c700dc7" data-reference-type="commit" data-container="body" data-placement="top" title="glx/dri3: work arround deadlock with xcb_wait_for_special_event()" class="gfm gfm-commit has-tooltip">tguillem/mesa@ea93da11</a></p>
449 <p data-sourcepos="11:1-11:55" dir="auto">It fixes VLC deadlocks with Qt when closing the window.</p>
450 </div>
451 </summary>
452 </entry>
453 <entry>
454 <id>tag:gitlab.freedesktop.org,2020-10-16:698794</id>
455 <link href="https://gitlab.freedesktop.org/xorg/xserver/-/merge_requests/401#note_655316"/>
456 <title>Konstantin commented on merge request !401 at xorg / xserver</title>
457 <updated>2020-10-16T09:22:10Z</updated>
458 <media:thumbnail width="40" height="40" url="https://gitlab.freedesktop.org/uploads/-/system/user/avatar/1697/avatar.png"/>
459 <author>
460 <username>rilian-la-te</username>
461 <name>Konstantin</name>
462 <email></email>
463 </author>
464 <summary type="xhtml">
465 <div xmlns="http://www.w3.org/1999/xhtml">
466 <p data-sourcepos="1:1-1:33" dir="auto"><a href="https://gitlab.freedesktop.org/daenzer" data-user="80" data-reference-type="user" data-container="body" data-placement="top" class="gfm gfm-project_member js-user-link" title="Michel Dänzer">@daenzer</a> will this be upstreamed?</p>
467 </div>
468 </summary>
469 </entry>
470 </feed>