Sam
I still believe something is wrong here; this example shows the inconsistency even more clearly: https://phpstan.org/r/b927e9f4-b350-4e7b-9dc7-8597669a6c48 - it only uses interfaces, so no actual implementation - it passes when the...
> please steer away from modifying proprietary software. No worries, I'm not modifying it, just researching it so that I can make a proper issue. Furthermore I don't think posting...
The problem is that phpstan does not realize it's running over TLS and it cannot find this out server side... So in the frontend it tries to connect over `ws:`...
> Sorry, I need to understand what problem are we fixing instead of blindly accepting a code change. I agree, no worries, if it's unclear I'm happy to elaborate. >...
Unfortunately this is not supported out of the box by the ReactPHP Http server, specifically it's `RequestHeaderParser`. A clean solution would be a middleware that parses the header. This luckily...
Can we mark this as a bug instead of a feature? I get that from a PHPStan perspective it might never have been supported, however from the perspective of set...
I think the issue is the same for the findroid client. The server (in 10.9) requires some extra attributes as part of the login message / communication..
How about instead labelling it as @ blocker for 10.9 stable release? Since you can't do a stable release if the apps don't work on it yet? 😕
> @rtribotte For the Nomad I think the main use case @shoenig was referring to would be to use the Task API's Unix socket as the Traefik provider address like...