Redirect Rules don't work on news.ycombinator.com

cvzakharchenko

Noticed it a long time ago, when Redirect Rules were just added.

These redirect rules don't ever work:

^{https://news.ycombinator.com/item?id=|https://www.hackernewz.com/item/}
^{https://news.ycombinator.com/item?id=(\d+)|https://www.hackernewz.com/item/$1}
^{https://news.ycombinator.com/item?id=([0-9]+)|https://www.hackernewz.com/item/$1}

After writing the rules, always tried a new search request to make sure, I don't hit any caches.
Similar rules work for other websites.

Link to news.ycombinator.com should have been replaced with a link to hackernewz.com.

gateway

I can get them to work by escaping the ?

https://news.ycombinator.com/item\?id=(\d+)|https://www.hackernewz.com/item/$1

This works, although the rewritten URL still has the parameter appended for some reason, so it looks like:
/item/31232722?id=31232722
Don't know why that happens but the URL still works fine (extra GET parameters are typically ignored).

cvzakharchenko

gateway Damn, I completely overlooked the question mark.
But the extra GET parameter is also a bug, though a minor one.

z64

Hi cvzakharchenko ,

But the extra GET parameter is also a bug, though a minor one.

This behavior is intentional, I gave an explanation here that goes over how it works & the motivation:

https://kagifeedback.org/d/2478-redirect-rules-dont-handle-url-query-string-correctly/7

Also see this section from our docs.

Our enactment of this behavior could be more sophisticated so that it doesn't show up in use cases such as your own - such as trying to detect that you "moved" the parameter - but we opted to not make it any more complicated since this is usually harmless, and this behavior otherwise has value as described in my other post / docs.

Hopefully that's clear, happy to hear other suggestions or improvements to our docs are welcome; they're open source on our GitHub.