Adrián Chaves

Are the latest test failures related to these changes?

I’m closing and reopening the pull request to trigger new CI tests. The last run had all jobs failing but 4, with seemingly unrelated issues..

> Should I add tests for `_guess_response_type` and `_guess_content_type`? I don’t think it is necessary as long as all their logic gets tested indirectly through other tests. It is hard...

Oh, I think I know [what’s happening](https://github.com/scrapy/xtractmime/pull/7/files#diff-8fca5263bd4283a3c3dcac7f148a871ea862fe9c50f10e81b95c0ec1241ef45c). @akshaysharmajs Maybe you can push a commit to xtractmime’s `main` branch removing that `print`? (no PR needed)

I’ve just run into something that may be worth addressing as part of these changes: https://github.com/scrapy/scrapy/blob/624a1ff3e97e693e85546a54a7abba3d94bbbebb/scrapy/downloadermiddlewares/httpcompression.py#L71-L74

My main concern is the amount of backward incompatible changes, if we consider different parameter types backward incompatible (which I think we should, and now I’m really glad we’ve started...

Could you also address the 1st bullet point of https://github.com/scrapy/scrapy/issues/4697? I imagine we should create a boolean `self._can_decode_br` in `__init__`, and then in `process_request` checking if the `Accept-Encoding` header exists...

-0.5, I cannot reproduce https://github.com/scrapy/scrapy/pull/5450#issuecomment-1189509615.

I think 1 is fine as it is: - I don’t think an exception during start request iteration should trigger the crawl to stop, just as exceptions during callbacks, and...

@lavoie005 Please, use [StackOverflow](https://stackoverflow.com/tags/scrapy) to ask questions, not existing, unrelated tickets. See [Getting Help](https://docs.scrapy.org/en/latest/#getting-help).