Mikhail Korobov

+1 to the issue. I like the idea of having 2 separate functions. It'd be nice to see how they simplify (or not) the code in Scrapy and in third-party...

A PR with socks proxy support is welcome! AFAIK nobody is working on it now.

Some random, unrelated thoughts, with no concrete suggestions :) 1. Another aproach could be to allow multiple download handlers for each schema, and make handlers themselves decide if they want...

I think the decision to use `idna` package instead of built-in `idna` is made by hyperlink, see https://github.com/python-hyper/hyperlink/pull/56 and https://github.com/python-hyper/hyperlink/issues/19 for motivation.

Me and @dangra wanted to create this package quite a long time ago; the idea was indeed to move scrapy deprecation utilities to a separate package, likely with a different...

I think we should reduce the scope of it a bit, and create a package just with `create_deprecated_class`. What do you think @Gallaecio? You had to copy this code to...

Won't it make it more confusing to understand what's the interface one have to implement to build a custom dupefilter?

I think this change might be backwards incompatible, because subclasses of BaseDupeFilter may implement .log() method which calls `super().log()`.

If I'm not mistaken, this parameter is deprecated in https://github.com/scrapy/scrapy/pull/5090; it was released with Scrapy 2.6.0 (https://docs.scrapy.org/en/latest/news.html#scrapy-2-6-0-2022-03-01), which happened 2022-03-01. According to https://docs.scrapy.org/en/latest/versioning.html#deprecation-policy, we promise at least one year before...

See also: https://github.com/scrapy/scrapy/pull/605