Attribute handle_httpstatus_list not working for codes 301 and 302

[zxcfer]

When set handle_httpstatus_list = [301, 302], the spider doesn't execute parse. Howeber, it executes for other codes like 404.

[jdemaeyer]

The 3xx HTTP code range is for redirects, and those are handled by the Redirect Downloader Middleware. If you don't want that middleware to automatically follow redirects, but instead handle them in your spider, you have two options:

1. Completely disable the RedirectMiddlware by setting REDIRECT_ENABLED = False in your settings,
2. Or, more versatile, decide whether you want the redirect middleware to automatically follow on a per-request basis. It looks for a dont_redirect key set to True in the Request.meta dictionary. So if you instantiate a request like Request("http://some.url", meta={'dont_redirect': True}), the redirect middleware would be disabled for this request only.

[jdemaeyer]

maybe the RedirectMiddleware should honour handle_httpstatus_list as well?

[zxcfer]

I think RedirectMiddleware should honour handle_httpstatus_list as well.

I tried making REDIRECT_ENABLED and dont_redirect equals False. Both worked, however returned response with 200 status code.

[jdemaeyer]

You need to set dont_redirect to True (see the 'dont'? ;)) if you want to use that way. What version of Scrapy are you using and how do you update the settings? I ask b/c I just accidentally tried using Spider.custom_settings with a Scrapy version that didn't support it yet.

Here's a Spider that has REDIRECT_ENABLED = False and does what you want with Scrapy 1.0.1:

# ~/playground/spidy.py

import scrapy

class Spidy(scrapy.Spider):
    name = "Spidy the Spider"
    start_urls = ["https://jigsaw.w3.org/HTTP/300/301.html"]
    custom_settings = {'REDIRECT_ENABLED': False}
    handle_httpstatus_list = [301]

    def parse(self, response):
        print "Got this:", response.status

jakob@MosEisley ~/playground % scrapy runspider spidy.py 2> /dev/null
Got this: 301

[barraponto]

In the code above, dont_redirect is set to True from within scrapy.Spider.make_requests_from_url.

[jdemaeyer]

Hm, don't you mean dont_filter (which is irrelevant for RedirectMiddleware)? Unless the Request class has some default meta dict I don't think dont_redirect is messed with anywhere in stock Scrapy

[barraponto]

Yeah, sorry. So in your example above, dont_redirect is not set to True.
UPDATE: oh, it's either REDIRECT_ENABLED set to False or dont_redirect set to True. Not both.

[zxcfer]

It was right jdemaeyer, all worked as expected it was an issue from some pages I was crawling. If nobody else is interested in RedirectMiddleware follow handle_httpstatus_list rule, we can close this issue.

[kmike]

I think that's a good idea for RedirectMiddleware to respect handle_httpstatus_list.

[nramirezuy]

+1

[dangra]

fixed by #1364