SUMMARY STEPS TO REPRODUCE 1. open Kate instance 2. type 'echo $KATE_PID in terminal window 3. OBSERVED RESULT nothing EXPECTED RESULT pid of current active Kate instance SOFTWARE/OS VERSIONS Windows: macOS: Linux/KDE Plasma: Linux (x86_64) release 5.4.0-47-generic Ubuntu 18.04.5 LTS (Xubuntu) KDE Plasma Version: 5.44.0-0ubuntu3 amd64 KDE Frameworks Version: 5.44.0 Qt Version: Qt 5.9.5 (built against 5.9.5) ADDITIONAL INFORMATION Kate version: 17.12.3 According to apt: kate/bionic,now 4:17.12.3-0ubuntu1 amd64 [installed] kate5-data/bionic,bionic,now 4:17.12.3-0ubuntu1 all [installed,automatic] ktexteditor-katepart/bionic,now 5.44.0-0ubuntu2 amd64 [installed,automatic] libkate1/bionic,now 0.4.1-7build1 amd64 [installed,automatic] Also fails when seeking envar from same terminal window: $ kate --new & $ echo $KATE_PID Though the command does return the pid for the launched session, I have just seen. Expectation created by Kate documentation: https://docs.kde.org/stable5/en/applications/kate/fundamentals-using-sessions.html $KATE_PID would be most convenient and would save time, if for no other reason than not following the docs up blind alleys, with all due respect for a great editor, for the benefits of which I am humbly grateful. Worth considering also, the likelihood of multiple Kate instances in multiple workspaces, which is quite normal. If active session is in one workspace, and terminal or script running in another workspace wants active kate session, will it get last active session?
Does it return the pid or just print it? Can't seem to pick it up with Python subprocess.check_output
Ah okay: pid = subprocess.Popen( [ "/usr/bin/kate", '--new' ], universal_newlines=True, close_fds=True, cwd=os.getcwd() ).pid
You mean: Since Kate 2.5.1 the PID of the current instance is exported to the environment variable KATE_PID. When opening files from the built in terminal Kate will automatically select the current instance if nothing else is indicated on the command line. => This means, Kate exports that to the terminal views it provides, not the shell that was used to Kate (I would not even know how the sub-process should modify the parent process env). That works as intended, but I think doesn't help you. But yes, you can get the pid yourself. Sorry if the documentation was confusing, but I think the sentences is ok as is.