Articles duplicates when URI contains non-ascii characters

[Rurik19]

Issue details

Articles duplicates when URI contains non-ascii characters

Environment

• wallabag version 2.0.4

Steps to reproduce/test case

add param with nonascii chars to url and save -> new article

[j0k3r]

Could you give us urls example?

[Rurik19]

Voila

this one saves only once
https://habrahabr.ru/company/intel/blog/283066/?aa=asciitext

but this one - as many times as i save it
https://habrahabr.ru/company/intel/blog/283066/?aa=русский

[j0k3r]

This will be a bit hard to solve.
Could you give me an example where the parameter really mater on the content? I mean, like if the parameter is broken the content will be 404 or completely different?

Thanks

[Rurik19]

Could you give me an example where the parameter really mater on the content?

No, I cannot. Parameters don't affect the content, but only influence the search of existing article with same URI. The links above are good example.

[MatElGran]

http://www.aritylabs.com/post/106091708292/des-contrôleurs-optionnels-grâce-à-constmissing poses problem too. There's a new item created each time I save it.