Markus Heiser

> FWIW this patch makes me have a fully functioning package. But I understand if you want to close it, if so I would think the best way forward would...

@SuperSandro2000 thanks a lot for your detailed feedback, given me some insights of NixOS packager :+1: > Just take away the ``-e`` and the package is done. :) > I...

@jwijenbergh @SuperSandro2000 sorry for the very long delay .. I have lost sight of the PR / sorry. I force pushed [1fc1d5f](https://github.com/searxng/searxng/pull/3045/commits/1fc1d5fd76647dc0cc0ea10a71d2a991e6c04e96) .. can you please test if it works...

> Drop support for Python 3.8 and earlier As long we support py3.8 we can't upgrade this package / py3.8 EOL is in 2024-10

> there is no memory leak on the master branch, only in some PR Thanks for reminder .. I modified the initial comment :+1: Even if we do not have...

> I wonder if the workers should really be restarted: > - If the worker are restarted, we won't notice the initial memory leaks: it will be under radar. So...

Should be fixed by .. can you verify? - https://github.com/searxng/searxng/pull/3480

Some thoughts of mine: - Do we need backward compatibility for a while? .. reading `hostname_replace` from settings.yml and prompt a deprecation message? - When we are implementing a replacement...

> The memory footprint is different: an long running instance Yes, this is known .. I assumed uWSGI worker processes do not live long .. but TBH I don't really...

> As far I understand, this require a change of the uwsgi configuration in all instances to make this PR relevant. I would have expected the processes to be restarted...