lucy24 - 11:50 am on Mar 24, 2013 (gmt 0)
If you have links between http pages and https pages, these are the options I can think of:
--use links with leading / and let users stay with https even when it isn't needed. This means the googlebot will eventually have duplicate versions of all your http pages. (Not of https pages, because I assume you don't let http get to those.)
--use links with / and then redirect internal traffic to http as needed. (Generally bad, because you shouldn't be redirecting your own links.)
--use complete links, including protocol, anywhere you link between http and https pages.
Oh, wait, there's one more possibility.
--change the whole site over to https. It will solve one group of problems, but may create others. (I assume there are drawbacks to using https, otherwise all websites everywhere would do it all the time.)
robots.txt is tricky because according to google-- I just found this yesterday while looking up something else--
The directives listed in the robots.txt file apply only to the host, protocol and port number where the file is hosted.
That makes it sound as if they expect separate robots.txt files for http and https, even if they belong to the same domain. But what are you supposed to do, rewrite to different robots.txt files depending on the robot's protocol? You might be able to do this on a brand-new site: Here are the rules for http, here are the ones for https. You can't really change at this point, though.