Andrey Rakhmatullin

I was talking about the code you changed in https://github.com/scrapy/scrapy/commit/313f9de28dc39acc5941d5a089930326c3981fdb#diff-1d71030dd60272d3753d83a2dc43e6e24bda769fb4e9385c8723e06efe3f3087 :P As it looks like the comment doesn't match the behavior and if the behavior, not the comment, is correct,...

If you mean whether it should replace the current default ones then it shouldn't, unless we see reasons for that.

#1455 :)

@PaulyBearCoding as you can see there is already a PR in progress

The direct problem here is the exception bubbling up to `_start_request_processing()`, exceptions in which aren't handled. OTOH I'm not sure if such an exception should crash the spider or be...

> I'm also unable to make this scrapy-poet test pass with older Scrapy for some reason (it could be a change in caplog) It looks like the reason for that...

Yeah. If we split mock* into another library it will be more straightforward to add tests there :)

Thanks! The nntplib part looks fine to me and matches [the Debian patch for this issue](https://sources.debian.org/src/python-pyzipper/0.3.6-5/debian/patches/0003.removed-nntplib-deprecated.patch/).

@JoseG2003-g I don't think it will be possible for you, especially considering that it's at a discussion stage.

After a recent short discussion on this we decided that as the first step we want a way to handle specific responses (mainly the 429 ones but it's likely that...