HTML: don't remove http(s) from links

[tdewolff]

HTML: don't remove http(s) from links

[stroborobo]

Ah, I see, you want people to use the new rel="external" attribute on anchor elements. IMHO all links containing a domain are to be considered external, what do you think?

[tillberg]

I too had this issue. I serve my sites only over https, and I already specify protocol for links only when sites do not support https, so for my own use I just removed the block of code that deals with this: tillberg@cfaa3ee (I'd be happy to turn that into a PR if desired, but it doesn't seem friendly to just go and create a PR for something that basically just nukes someone else's code :) )

I agree with @Knorkebrot above. I get that there might be some use case for trying to strip the protocol off of pages that include them in self-references (e.g. some heavy server-side application frameworks do this, I think), but that seems like a small benefit to get compared to requiring everyone to mark up their anchors with rel="external" whenever they need to force the link protocol. (I also had never heard of rel="external" before, so I might be biased by my own ignorance.)

[tdewolff]

Yes you are right, it's not ideal. I will make changes to keep the protocol. It's a pity because stripping http: can considerably reduce the size.

[tdewolff]

Fixed in 3b07a59 on the develop branch. It will be re-enabled in the future when passing options is possible, including the currently used protocol. Either http or https can then be omitted, but not both.

[tdewolff]

Please check the options branch, it contains an implementation of passing options to the minifier. HTML now has the option scheme which can be either http or https and it will remove only that protocol for links. Call m.Set("scheme", "http") to enable that.