Multi-generators cannot be used without 'build_type' setting

[jgsogo]

Changelog: Fix: Multi-generators cannot be used without build_type setting. A failure is forced to cmake_find_package_multi and visual_studio_multi as it was in cmake_multi.
Docs: omit

closes #6135

This PR forces in all the multi-generators the same failure that we were triggering for the cmake_multi one.

Currently, the full error shown to the user that runs the conan install is:

[...]

conanfile.py: ERROR: Generator cmake_multi(file:None) failed
'settings.build_type' doesn't exist
'settings' possible configurations are ['arch', 'compiler', 'os']
ERROR: 'settings.build_type' doesn't exist
'settings' possible configurations are ['arch', 'compiler', 'os']

Let me know if we want to improve it to add a more meaningful message.

#tags: slow

[jgsogo]

@lasote I feel like it makes no sense to use a multi generator in a recipe without build_type. Existing generator cmake_multi fails in this scenario, with this PR generators cmake_find_package_multi and visual_studio_multi will fail too (we may write a better error).

The advantage of the multi generators is to have the files for all the build_types in the same directory, but if the recipe doesn't have this setting, the build type information won't reach the generators and we can't know which file to create.... also, we can't just create a XXXTarget-AllBuildType.cmake file because it will contain only the information corresponding to one build type. So, IMO the best alternative is to fail with an error in the recipe.

This fixes #6135 in the sense that the error now says something about the build_type and the generator.

[memsharded]

This works both for:

• conanfile.py with build_type setting.
• conanfile.txt (because this one doesn't remove/blank the settings, it keeps the ones defined in the profile)

If we wanted to build & test a header-only library it is necessary to define the build_type setting if you want to use this generator, but I am fine with that.

I was about to merge it, but I saw comment from @lasote here #6228 (comment), so I might be missing some previous discussion. Leaving it open and assigning to him.

[lasote]

I haven't commented anything about this. Go ahead.