timmarinin / discuss-on-hn Goto Github PK

gus_massa wrote:

An "advanced" feature request: Some sites are rewritten, for example https://example.blogspot.com.ar is transformed to by HN to https://example.blogspot.com and BlogSpot will transform it to whatever country you are, so it would be nice to add in version 2 some canonizations to make it more compatible with HN. Also, medium is a nightmare, because it adds a random mark to the url and the HN post may have a different random mark. And the nytimes adds something like &r_1=0 or something. (Probably medium is the more important use case.)

So there are few special cases that I'd like to be added

• medium.com, strip part of url
• blogspot.com, if there is country tld, drop it
• maybe, clear querystring nytimes

I don't want to just drop query string, because then I'll loose cases like http://c2.com/cgi/wiki?GreencoddsTenthRuleOfProgramming, so there should be caution.

It seems to make sense to add library to properly parse URI instead of ugly regexps at the same time.