Closed rursache closed 4 months ago
Thanks for your PR @rursache !
I appreciate your work, but it's worth saying that the infinite loop on the script doesn't apply only to docker, but also to other forms of execution.
The biggest problem with this is that it makes it incompatible with GitHub Actions. The secondary problem is that I don't personally like running it like this, as it doesn't deal well with failures: Imo starting 'from scratch' from time to time has it's advantages, when run outside a docker container.
I also think that the first two and last two bullet points would be better served with two separate PRs to be honest (among others because I'd merge that in right away).
while i understand what you say, i made these changes because first of all i want to run this script in a docker container, set and forget it. i see no benefit of running it manually. the loop "starts from scratch" if it crashes due to --restart unless-stopped
in docker run/compose param. also its not resource intensive so it can be ran locally on a machine instead of semi-abusing GH actions.
a new flag (--docker
) could be added to the python script to do this loop only when ran as a container but i don't plan to implement and maintain that. that would fix everything you mentioned imo.
good luck and all the best
@rursache Check our PR #102
config.json
or via the--sleep-interval
argumentconfig.json
. Cleaner and more elegant + fixeshttps://github.com/nanos/FediFetcher/issues/64