Mikko Koivunalho
Mikko Koivunalho
> If I understand correctly, the way you see it is: > > The App now uses Role::Config > The role Role::Config has config_readers and config_files > Each of the...
If we ditch `config_files` attribute, we create a non-compatible change. The name is bad but I like the idea that the `Config` object would actually know where the configs come...
Attribute `config_files` is part of the public interface, isn't it? If it isn't, then of course, it's not a compatibility issue. Just thinking that I like the idea that the...
I have changed all the small things and I removed the attribute`Dancer2::Core::Role::Config::config_files`.
@xsawyerx Did you really mean we should make so big change? To abandon App having Config as a role, and instead instantiate ConfigReader? It means we would need to implement...
To be on the safe side, I rebased the branch from master. So I had to force push. The new commits are: * [Create Role::HasEnvironment and tests](https://github.com/PerlDancer/Dancer2/pull/1637/commits/839c0be65c5f1792c6edee38e9edebdf729f6a7b) [839c0be](https://github.com/PerlDancer/Dancer2/pull/1637/commits/839c0be65c5f1792c6edee38e9edebdf729f6a7b) * [Add...
The test `t/issues/gh-634.t` is failing because it uses `config_location` from `Role::Config`. The test needs to be thought out again.
The tests now run. @xsawyerx Do we need more fine tuning before merging?
@DavsX - Any progress on this issue?
A script which creates config_local.yaml on the fly and then starts Dancer2 app would otherwise work but it means that container won't be immutable. Immutable is one of the architectural...