Improve error handling #1427
Improve error handling #1427
Conversation
| @@ -68,12 +68,19 @@ namespace SDDM { | |||
| if (env.value(QStringLiteral("XDG_SESSION_TYPE")) == QLatin1String("wayland") && env.value(QStringLiteral("XDG_SESSION_CLASS")) == QLatin1String("greeter")) { | |||
| m_wayland = new WaylandHelper(this); | |||
| m_wayland->setEnvironment(env); | |||
| if (!m_wayland->startCompositor(m_displayServerCmd)) | |||
| connect(m_wayland, &WaylandHelper::failed, this, &UserSession::stop); | |||
There was a problem hiding this comment.
Why do we use a different path for late failure compared to failing to start?
There was a problem hiding this comment.
But you're right that no-one seems to be calling stop()
Should this be called from:
Backend::closeSession
?
There was a problem hiding this comment.
Why do we use a different path for late failure compared to failing to start?
Which path are you referring to? ::setup calls exit() upon error
Should this be called from
Backend::closeSession?
This will have closeSession be called because connect(m_session, &UserSession::finished, this, &HelperApp::sessionFinished);
There was a problem hiding this comment.
I was referring to:
if (!m_wayland->startCompositor(m_displayServerCmd)) {
Q_EMIT finished(1);
and asking why this isn't
connect(m_wayland, &WaylandHelper::failed, this,[this]() {
Q_EMIT finished(1);
});
There was a problem hiding this comment.
Addressed by having stop() emit finished.
I don't think it makes sense to stop when we failed to start.
Be it because the x11 user session could not be started or because the wayland compositor can't.
The alternative is to stay stuck in a weird limbo
aleixpol
force-pushed
the
work/error-handling
branch
from
e522744
to
c8789d2
Compare
Now that we are not running sessions with almighty processes they're bound to fail.
This patch tries to increase the visibility of these problems and offers some mechanisms so that it doesn't leave the system stuck.