Bug#948676: at-spi2-core: /etc/X11/Xsession.d/90qt-a11y does not enable Qt5 support as the Wiki indicates
On Sun, Jan 12, 2020 at 09:05:10PM +0100, Samuel Thibault wrote: > Control: forwarded -1 https://codereview.qt-project.org/c/qt/qtbase/+/286336 > > Dmitry Shachnev, le dim. 12 janv. 2020 22:31:14 +0300, a ecrit: > > Any chance you can submit the patch to Qt Gerrit [1]? > > Here it is. Thanks a lot for that! Let's wait for some review, and then I will include it in the packaging. -- Dmitry Shachnev signature.asc Description: PGP signature
Bug#948676: at-spi2-core: /etc/X11/Xsession.d/90qt-a11y does not enable Qt5 support as the Wiki indicates
Control: forwarded -1 https://codereview.qt-project.org/c/qt/qtbase/+/286336 Dmitry Shachnev, le dim. 12 janv. 2020 22:31:14 +0300, a ecrit: > On Sun, Jan 12, 2020 at 07:54:22PM +0100, Samuel Thibault wrote: > > Could we at least restore looking at IsEnabled in Qt, as the attached > > patch does? Actually, that is needed for proper focus tracking in Qt > > applications for zoom support in compiz when Orca is not used. The > > comment there says that IsEnabled is always true, but that's not what I > > am observing with e.g. Debian 8's gnome 3.14, Debian 9's gnome 3.22, and > > Debian 10's 3.30... Ideally, we'd backport this fix to buster too. > > Any chance you can submit the patch to Qt Gerrit [1]? Here it is. Samuel
Bug#948676: at-spi2-core: /etc/X11/Xsession.d/90qt-a11y does not enable Qt5 support as the Wiki indicates
Thorsten Glaser, le dim. 12 janv. 2020 19:37:03 +, a ecrit: > Samuel Thibault dixit: > > >I guess in your case what you expected was that at the very least > >applications started after orca got started should be accessible, even > >if orca is only started by hand. This would work if Qt was not only > >looking at org.a11y.Status ScreenReaderEnabled (the control panel > >configuration), but also at org.a11y.Status IsEnabled (set by Orca on > >startup). This would not be enough for making existing applications > >accessible when you first start Orca, Bug#874054 mentions that it's > >because Qt currently doesn't monitor the dbus IsEnabled variable. > > Ooh, interesting. From what the MuseScore people say, after start > of orca would be enough (also, AIUI, orca starts with the desktop > session or can be made to). Starting orca only wouldn't be enough. But enabling Orca startup in the control panel would be enough. My patch is for the former case. > >Could we at least restore looking at IsEnabled in Qt, as the attached > >patch does? Actually, that is needed for proper focus tracking in Qt > >applications for zoom support in compiz when Orca is not used. The > > Would this have the effect of losing keyboard input in MuseScore > when orca is used, which is what one tester reported? That shouldn't happen, but that could be an effect indeed. > I’m under plain X11 with evilwm, no desktop environment, and did > not have that issue. Getting Orca to access the content of the application would be enough to get the same situation as the users. Samuel
Bug#948676: at-spi2-core: /etc/X11/Xsession.d/90qt-a11y does not enable Qt5 support as the Wiki indicates
Dmitry Shachnev, le dim. 12 janv. 2020 22:31:14 +0300, a ecrit: > Any chance you can submit the patch to Qt Gerrit [1]? I'll have a look at it. Samuel
Bug#948676: at-spi2-core: /etc/X11/Xsession.d/90qt-a11y does not enable Qt5 support as the Wiki indicates
Samuel Thibault dixit: >I guess in your case what you expected was that at the very least >applications started after orca got started should be accessible, even >if orca is only started by hand. This would work if Qt was not only >looking at org.a11y.Status ScreenReaderEnabled (the control panel >configuration), but also at org.a11y.Status IsEnabled (set by Orca on >startup). This would not be enough for making existing applications >accessible when you first start Orca, Bug#874054 mentions that it's >because Qt currently doesn't monitor the dbus IsEnabled variable. Ooh, interesting. From what the MuseScore people say, after start of orca would be enough (also, AIUI, orca starts with the desktop session or can be made to). >Could we at least restore looking at IsEnabled in Qt, as the attached >patch does? Actually, that is needed for proper focus tracking in Qt >applications for zoom support in compiz when Orca is not used. The Would this have the effect of losing keyboard input in MuseScore when orca is used, which is what one tester reported? I’m under plain X11 with evilwm, no desktop environment, and did not have that issue. (On the other hand, this means I have no control panel to configure orca either. But I’ve only ran tests on Debian for the benefit of upstream anyway…) bye, //mirabilos -- >> Why don't you use JavaScript? I also don't like enabling JavaScript in > Because I use lynx as browser. +1 -- Octavio Alvarez, me and ⡍⠁⠗⠊⠕ (Mario Lang) on debian-devel
Bug#948676: at-spi2-core: /etc/X11/Xsession.d/90qt-a11y does not enable Qt5 support as the Wiki indicates
Hi Samuel! On Sun, Jan 12, 2020 at 07:54:22PM +0100, Samuel Thibault wrote: > [...] > > Could we at least restore looking at IsEnabled in Qt, as the attached > patch does? Actually, that is needed for proper focus tracking in Qt > applications for zoom support in compiz when Orca is not used. The > comment there says that IsEnabled is always true, but that's not what I > am observing with e.g. Debian 8's gnome 3.14, Debian 9's gnome 3.22, and > Debian 10's 3.30... Ideally, we'd backport this fix to buster too. Any chance you can submit the patch to Qt Gerrit [1]? I do not want to include patches not from upstream, and I cannot submit it myself because the patch author must agree with the CLA. [1]: https://codereview.qt-project.org/ -- Dmitry Shachnev signature.asc Description: PGP signature
Bug#948676: at-spi2-core: /etc/X11/Xsession.d/90qt-a11y does not enable Qt5 support as the Wiki indicates
Control: reassign -1 qtbase-opensource-src Hello, Thorsten Glaser, le sam. 11 janv. 2020 21:12:47 +0100, a ecrit: > https://wiki.debian.org/accessibility-devel says it should export > QT_LINUX_ACCESSIBILITY_ALWAYS_ON=1 but it doesn’t. > > Exporting that before starting MuseScore makes the latter show up > in “orca -l” output, so it’s probably needed. Yes and no. There is a permanent battle between enabling accessibility by default, and enabling it only when actually needed. See Bug#874054 ("Setting QT_LINUX_ACCESSIBILITY_ALWAYS_ON=1 has a huge negative performance impact, should not be always on") for more details. Currently Qt only enables accessibility when the orca screen reader is enabled in the control panel. I guess in your case what you expected was that at the very least applications started after orca got started should be accessible, even if orca is only started by hand. This would work if Qt was not only looking at org.a11y.Status ScreenReaderEnabled (the control panel configuration), but also at org.a11y.Status IsEnabled (set by Orca on startup). This would not be enough for making existing applications accessible when you first start Orca, Bug#874054 mentions that it's because Qt currently doesn't monitor the dbus IsEnabled variable. Could we at least restore looking at IsEnabled in Qt, as the attached patch does? Actually, that is needed for proper focus tracking in Qt applications for zoom support in compiz when Orca is not used. The comment there says that IsEnabled is always true, but that's not what I am observing with e.g. Debian 8's gnome 3.14, Debian 9's gnome 3.22, and Debian 10's 3.30... Ideally, we'd backport this fix to buster too. Samuel No, isEnabled does not always return true currently. We do need to test for isEnabled for the case of running Orca by hand. It's Orca (or compiz' focuspoll) which sets IsEnabled. --- src/platformsupport/linuxaccessibility/dbusconnection.cpp |3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) --- a/src/platformsupport/linuxaccessibility/dbusconnection.cpp +++ b/src/platformsupport/linuxaccessibility/dbusconnection.cpp @@ -120,8 +120,7 @@ void DBusConnection::serviceRegistered() //debugging. static const bool a11yAlwaysOn = qEnvironmentVariableIsSet("QT_LINUX_ACCESSIBILITY_ALWAYS_ON"); -// a11yStatus->isEnabled() returns always true (since Gnome 3.6) -bool enabled = a11yAlwaysOn || a11yStatus->screenReaderEnabled(); +bool enabled = a11yAlwaysOn || a11yStatus->screenReaderEnabled() || a11yStatus->isEnabled(); if (enabled != m_enabled) { m_enabled = enabled;
Bug#948676: at-spi2-core: /etc/X11/Xsession.d/90qt-a11y does not enable Qt5 support as the Wiki indicates
Package: at-spi2-core Version: 2.34.0-3 Severity: normal https://wiki.debian.org/accessibility-devel says it should export QT_LINUX_ACCESSIBILITY_ALWAYS_ON=1 but it doesn’t. Exporting that before starting MuseScore makes the latter show up in “orca -l” output, so it’s probably needed. -- System Information: Debian Release: bullseye/sid APT prefers buildd-unstable APT policy: (500, 'buildd-unstable'), (500, 'unstable'), (1, 'experimental') Architecture: amd64 (x86_64) Foreign Architectures: i386 Kernel: Linux 5.4.0-1-amd64 (SMP w/2 CPU cores) Locale: LANG=C, LC_CTYPE=C.UTF-8 (charmap=UTF-8), LANGUAGE=C (charmap=UTF-8) Shell: /bin/sh linked to /bin/lksh Init: sysvinit (via /sbin/init) Versions of packages at-spi2-core depends on: ii libatspi2.0-0 2.34.0-3 ii libc6 2.29-7 ii libdbus-1-31.12.16-2 ii libglib2.0-0 2.62.4-1 ii libx11-6 2:1.6.8-1 ii libxtst6 2:1.2.3-1 at-spi2-core recommends no packages. at-spi2-core suggests no packages. -- no debconf information