Should recv_timeout option be ignored on pooled connections?

[tomekowal]

Hi!

I've hit an issue where I am trying to set recv_timeout option when the connection comes from a pool.

I've made a test reproducing the issue in a fork:

tomekowal#1

Is it by design and pooled connections should have the recv_timeout set in stone or is it a bug introduced when changing the architecture?

[benoitc]

fixed in latest main branch. Thanks for the report!