log looks the same. I think that login check is issuing a command on the proxy to check if the proxy is working ("echo LOGIN OK"). And due to the nature of our gateway (see above) this fails, because it is an invalid command.
Unfortunately I don't really see if this assumption is correct because I have no access to the gateway logs and the x2goclient logs do not contain any information _why_ the login check failed. I have tried getting some gateway logs but I have not yet gotten anything.
Is there anything I can do to bypass that login check?
I have now done some tests:
- this also happens with the x2goclient 4.1.2.2 on Linux
- I can confirm that my assumption about the LOGIN OK check was correct. Inserting "return true;" just before the line " if ( ssh_channel_request_exec ( channel, "echo \"LOGIN OK\"" ) != SSH_OK ) " in sshmasterconnection.cpp and thus skipping the whole interactivity code makes it work again.
This check was introduced as a result of Bug #592 to enable the user changing an expired password interactively but I have not fully understood the whole idea of sending echo "LOGIN OK" and then checking if is NOT in the buffer. Isn't checking for pty sufficient?
QString inf=QByteArray ( buffer,nbytes );
x2goDebug<<"LOGIN CHECK:"<<inf;
if(inf.indexOf("LOGIN OK")!=-1)
{
x2goDebug<<"don't have interaction";
hasInterraction=false;
break;
}@Alex can you please explain?
Uli