Open mohmad-null opened 5 months ago
It's parsel, not Scrapy.
Yes, I noticed that, but I figured Scrapy should catch it.
I don't think Scrapy should catch it, and Scrapy will only able to do that by overriding remove_namespaces()
.
But AttributeError: 'dict' object has no attribute 'iter'
is a meaningless error for a scrapy user.
If there's an exception (fine), it should be something meaningful "Cannot remove namespaces from JSON" or something - that's why I consider it a Scrapy issue.
The code is in parsel, not Scrapy. It should be done in parsel, not in Scrapy. Migrating.
Hello @wrar, I'm new to contributing to parsel and would be more than happy to take a shot at this. Can you elaborate a bit on what needs to be done?
@kanjikinsmoke we need to make the remove_namespaces()
call on JSON-type selectors not fail with an uncaught exception. I think it needs a self.type
check similar to e.g. Selector.css()
.
Also it would be nice to check if there are any other methods or properties that behave the same and do the same fix for them, e.g. attrib
also gives an unhandled exception (#284 is likely about that).
I saw
Regression. As part of my unit tests, I have a well-formed JSON file that I'm using in some tests as "bad" data (the script wants XML).
I do something like this:
This used to work (1.7.x), now (2.11.1) it asserts:
I appreciate JSON doesn't have namespaces, but it should fail gracefully, not raise an assertion.