Add an alter hook for typed config definitions

Good work @webflo. Few minor code review stuffs below. Overall, looks OK to me.

1. +++ b/core/lib/Drupal/Core/Config/TypedConfigManager.php
   @@ -40,6 +41,18 @@ class TypedConfigManager extends PluginManagerBase implements TypedConfigManager
   +  /**
   
   @@ -47,12 +60,26 @@ class TypedConfigManager extends PluginManagerBase implements TypedConfigManager
   +  protected function alterInfo(ModuleHandlerInterface $module_handler, $alter_hook) {
   

   Can we add the comment for this property please?

2. +++ b/core/lib/Drupal/Core/Config/TypedConfigManager.php
   @@ -47,12 +60,26 @@ class TypedConfigManager extends PluginManagerBase implements TypedConfigManager
   +    $this->alterInfo($module_handler, 'config_typed_info');
   

   isn't it hook_config_typed_data_info_alter() ? Also wondering, would this protected method add any value? can't we do the assignment in construction itself?

Thanks, looks good.

#4: add-hook_config_typed_info_alter-2100223-4.patch queued for re-testing.

Back to RTBC since core was failing with a random fail.

After discussing this @webflo in Prague I don't think that we should be providing a method to alter config typed data metadata as I don't think we should allow modules to be altering such fundamental metadata just to enhance it. Apparently the config translation ui is already extending the locale type data manager so we could decorate the data here instead. That way we won't have yet another alter that can completely pollute / change the meaning of stuff from under our feet.

Sure.

I would like to get some more reasoning for this. Specifically, I don't understand how this is not viable, if hook_data_type_info_alter() is. Thanks!

Seems to me like the last two sentences from @alexpott were reason enough? :)

Well, that could still use some clarification for noobs like me :-)
I was in the middle of posting a more detailed answer, but I ended quoting the entire last two sentences in #11. So in short: I don't understand at all, what is meant.

Thanks, guys!

Added _no_

I'll close this as won't fix, this seems to be the consensus.