I have a newly installed Flarum server. The server is installed on a web-hosting site, thus contacted via https. The server - as such - seems to be running fine.
My problem is that some users have access problems from time to time, where they either cannot log in at all, or
they appear to be logged in in a kind of half-way (but not able to access any internal data).
I am trying to fully comprehend why this happens (when it happens), therefore this request.
I do understand that different browsers might prefer http or https differently as first request when trying to access, say
myserver.mysite.com
If the user fully writes
https://myserver.mysite.com
then there is no ambiguity, of course.
However, if the browser (silently) attempts http://myserver.mysite.com when the user just writes myserver.mysite.com
then the user will usually get
Oops! Something went wrong during a cross-origin request.
Writing the full URL with the https protocol field will remedy this.
Now, I know that the server side will redirect a http://myserver.mysite.com request to https if set up to do so.
However, the problem - as far as I can understand - is the url entry in the config.php server configuration file
'url' => 'https://myserver.mysite.com'
Apparently the Flarum server spots the inconsistency between 'https://myserver.mysite.com' in the config.php file
and the user having (unknowingly) sent http://myserver.mysite.com when the protocol field is omitted.
My question the therefore, how can one avoid this problem from occurring, why should the user really have to
"force" the browser to use https? The user should really one care about writing
myserver.mysite.com
I would believe?
I realize that my 'real' problem might be my own lack of understanding something, in which can I only apologize.
I would really like to hear how others are addressing this issue, as I cannot believe I am the only one to experience this.