Caught exception at Mir/EGL driver boundary: /build/buildd/mir-0.8.0+14.10.20141010/src/client/rpc/stream_socket_transport.cpp(164): Throw in function virtual void mir::client::rpc::StreamSocketTransport::send_data(const std::vector<unsigned char>&)

marking dup of bug 1368652

not a dup, in this instance we need to determine why the mir server is dying

is this launched from the command line? or simply launched from the dash?

also, is there a particular application this is happening with ?

@alf - seems like this one may also be related
bug 1382162

@kevin: I thought the log I attached would give that away :-)

So,

- launched from the dash
- here maps (but I'm not sure it's the only app that can cause this)

marking as high per RTM bug review

With rtm-proposed image r135 I haven't been able to reproduce this exact problem yet. However, occasionally when I tap the maps icon nothing happens and I get:

UbuntuClientIntegration: connection to Mir server failed. Check that a Mir server is
running, and the correct socket is being used and is accessible. The shell may have
rejected the incoming connection, so check its log file

and in unity8.log:

qtmir.mir: SessionAuthorizer::connection_is_allowed - this= SessionAuthorizer(0x191d5fc) pid= 31607
qtmir.applications: ApplicationManager::authorizeSession - pid= 31607
ApplicationManager REJECTED connection from app with pid 31607 as no desktop_file_hint specified
qtmir.applications: ApplicationManager::onProcessFailed - appId= "com.nokia.heremaps_here" duringStartup= true
ApplicationManager::onProcessFailed - upstart reports failure of application "com.nokia.heremaps_here" that AppManager is not managing
qtmir.applications: ApplicationManager::onProcessStopped - appId= "com.nokia.heremaps_here"
ApplicationManager::onProcessStopped reports stop of appId= "com.nokia.heremaps_here" which AppMan is not managing, ignoring the event

So it seems that under some circumstances the desktop_file_hint is not correctly set or read. Could be a different issue, or a different manifestation of the same issue this bug describes, not sure yet.

AFAIK, this error always gets spat out when the server vanishes and the client doesn't expect it. Are you sure unity8 hadn't crashed and quickly restarted?

Also, are we sure we need to keep this bug open when bug 1368652 and bug 1312099 already exist?

On miércoles, 11 de febrero de 2015 04h'20:12 BRST, Daniel van Vugt wrote:
> AFAIK, this error always gets spat out when the server vanishes and the
> client doesn't expect it. Are you sure unity8 hadn't crashed and quickly
> restarted?

You can't really expect me to look over a 6month old bug that seldom
happens?

> Also, are we sure we need to keep this bug open when bug 1368652 and bug
> 1312099 already exist?

Just dup or close it ;-)

>
> ** Changed in: mir
> Milestone: 0.12.0 => None
>
> ** Changed in: mir
> Status: New => Incomplete
>

--
Sent using Dekko from my Ubuntu device