Guillaume Gelin

@tiangolo While it wasn't my first purpose, it also seems to make `jsonable_encoder` ~40% faster on the benchmark you used in #756: ```bash (.venv) $ # without #5718 (.venv) $...

I just updated the first message with a code example to make the problem clearer. @tiangolo Any hint on the direction that this PR should take?

Also, FWIW, Pydantic uses the same approach: https://github.com/pydantic/pydantic/blob/f55515820a407c5d31c45a7b0307925eecf21f34/pydantic/json.py#L82-L90

Thanks for the feedback @dkatzan! Yeah, I can definitely add a test or anything else that would help; I'd prefer a confirmation that it's the correct approach before putting more...

rebased

rebased

rebased

rebased

As a LangChain newcomer, I just got hit by this one as well. The tools documentation shows examples of multi-input tools without mentioning agent types ([e.g. here](https://python.langchain.com/docs/modules/agents/tools/how_to/custom_tools#using-the-decorator)).

No, it's not possible as of today. You can already add an instance of `libqtile.layout.Floating` in your configuration `layouts` variable, but as you noticed, any modification of a window's size...