Kevin

Hi @ritwickdey is there any way to get this working for `.html.liquid` files?

Thanks @palmieric would you be able to resolve the conflicts?

In general I think this is fine and adds a useful option. We could probably look at extending [error_handler](https://github.com/3scale/APIcast/blob/e7f6ebbf15d1b053283f0d77ae7176e1c7bfcb44/gateway/src/apicast/policy/rate_limit/rate_limit.lua#L44-L53) in the future to map the same auth caching modes we...

@pimg apologies for the huge delay in reviewing this but are you able to rebase onto master so we can mere this finally?

@mikz correct, this might be considered an enhancement rather than a bug of course. The RFC states the port is accepted although in my experience it's not that common.

oh I see why it might not have been clear now, the markdown is not rendering the port correctly in my original description

@mikz that's interesting because on the previous version I can modify the `Host` header. Image: **registry.access.redhat.com/3scale-amp22/apicast-gateway:1.8-12** Timestamp: **2018-08-17T08:52:30.421763Z** With this image the `Host` header is modified using the same policy...

@mikz it's not possible to add the port in the **host header rewrite** field until [THREESCALE-1408](https://issues.jboss.org/browse/THREESCALE-1408) is fixed. Otherwise yeah that was my first solution.

@mikz so what will the decision be here? To fix the hostname rewrite field and keep the current behaviour? Just wondering if the change in behaviour on APIcast between versions...

@mikz what will be the decision here then? To fix the hostname rewrite field and use that or change the current behaviour in APIcast? Wondering if the change in behaviour...