Empty new lines written to docker log file about fhem-docker HOT 7 CLOSED

What do you mean with "only with one single newline"? Are you now able to recreate the issue or not?

I can't reproduce the error, with sudo docker logs <name of container> -f
Bit in the logfile sometimes newlines are added
{"log":"\n","stream":"stdout","time":"2023-09-12T15:06:34.882448783Z"}

And sorry to bring this up again, but: What check is in place so that printf is executed only if there is something in $logArray to be printed?

Currently, none is implemented, i'll fix this

from fhem-docker.

I saw this bug before, but not in this Image version.

Also i am currently not able to reproduce this

from fhem-docker.

Okay, interesting.
I just tested it with a fresh service, i.e. docker-compose.yml:

version: "3.3"
services:
  fhemtest:
    image: ghcr.io/fhem/fhem-docker:3-bullseye
    container_name: fhemtest
    restart: unless-stopped
    environment:
      TZ: Europe/Berlin

and can confirm that the bug (still) exists by running sudo docker logs fhemtest -f.

Edit: The logs command is not even needed. The same happens if you simply run sudo docker compose -f docker-compose.yml up.

from fhem-docker.

Can you check that you are really running container version 3.2.3?

Can you run a inspect, on the started container like this:
docker inspect <containerID>| grep version

from fhem-docker.

Yes, I am running 3.2.3:

                "com.docker.compose.version": "2.21.0",
                "org.fhem.version": "",
                "org.opencontainers.image.version": "3.2.3-bullseye"

Even if I pin the version in the docker-compose.yml to ghcr.io/fhem/fhem-docker:3.2.3-bullseye it's the same. I also just tested it on another system with the compose-file posted above (also Debian bullseye) and the bug is still present.

Just from a logical point of view:
There is no check in place whether $logArray is actually populated. So printf '%s\n' "${logArray[@]}" is executed every time PrintNewLines runs, independent of what $logArray contains (or doesn't contain). There is simply nothing there preventing it from printing \n every time it runs.

On another note: My "fix" from above, moving printf … into the else-branch works flawlessly, even after two nightly logrotates now.

from fhem-docker.

I see a similar problem only with one single newline.
For this i have already prepared a fix in the bookworm branch

from fhem-docker.

What do you mean with "only with one single newline"? Are you now able to recreate the issue or not?

And sorry to bring this up again, but: What check is in place so that printf is executed only if there is something in $logArray to be printed?

from fhem-docker.