If by WSGI you actually mean Apache/mod_wsgi, then although mounted WSGI applications normally get run in their own sub interpreters, the 80/443 split is a special case and even though in different VirtualHost so long as mount point for WSGIScriptAlias, and the ServerName are the same, they will be merged.
<VirtualHost *:80>
ServerName www.example.com
WSGIScriptAlias / /some/path/django.wsgi.
</VirtualHost>
<VirtualHost *:443>
ServerName www.example.com
WSGIScriptAlias / /some/path/django.wsgi.
</VirtualHost>
This will happen for daemon mode as well, but with daemon mode you need to define only a single daemon process group in first VirtualHost definition and then just refer to that from both with WSGIProcessGroup.
<VirtualHost *:80>
ServerName www.example.com
WSGIDaemonProcess mydjangosite ...
WSGIProcessGroup mydjangosite
WSGIScriptAlias / /some/path/django.wsgi.
</VirtualHost>
<VirtualHost *:444>
ServerName www.example.com
WSGIProcessGroup mydjangosite
WSGIScriptAlias / /some/path/django.wsgi.
</VirtualHost>
The WSGIProcessGroup can only reach across like to that VirtualHost for same ServerName.
Django provides a is_secure() method for determining when request came via HTTPS which derives from WSGI variable with request called 'wsgi.url_scheme' which is set by mod_wsgi.
So, you would have one single Django WSGI script file and settings file. You just need to duplicate application mounting as decsribed in Apache/mod_wsgi configuration.
与恶龙缠斗过久,自身亦成为恶龙;凝视深渊过久,深渊将回以凝视…