Closed DaCHack closed 4 weeks ago
It needs a protocol to construct a valid link. If there would be two urls, how would it decide which one to use? If we'd remove the protocol and only use the domain, how should it figure out which protocol to use? (We can't rely on the current protocol because that won't work in async events like email sending)
And, why do you need to access Vikunja via http, when you have an https deployment set up as well?
Hmm tough one. I would need to understand these async events better. What is the actual issue here?
As for the use case: First and foremost changing from HTTP to HTTPS on a running system is a breaking change for all users this way. In my case I am not able to migrate all users in an automated way. There can also be other reasons to having both ways: Lets say you want to have a HTTP-fallback in case there is an issue with traefik. There could be other reasons as well...
Hmm tough one. I would need to understand these async events better. What is the actual issue here?
Async events happen without someone making an http request, so no protocol is available at that time.
As the use case seems to be quite an edge case, I don't think we can easily support multiple public urls.
Description
When running Vikunja in docker you need to specify the VIKUNJA_SERVICE_PUBLICURL environment variable.
I have a setup now that currently would allow HTTP (native) and HTTPS (via traefik) connections to Vikunja.
But it seems I can only set this environment variable to either one of these protocols which does not allow for a dual protocol setup.
Vikunja Version
0.24.3
Browser and version
Brave Version 1.71.114 Chromium: 130.0.6723.58
Can you reproduce the bug on the Vikunja demo site?
No
Screenshots
No response