Vít Ondruch

@eregon I don't disagree. I am just saying that the default should be changed and there should not be this "fallback" magic. It won't help to anybody. It just going...

> @voxik my understanding is that you're suggesting to just default to `--user-install` all the time. Is that correct? Right > It looks like [Fedora sets up `operating_system.rb` so root...

1. Bundler should always respect to RubyGems settings. 2. If there was need to configured anything, then operating_system.rb should be the place. This is the place where distributions modifies the...

Unfortunately, I am not convinced that #2909 would help here. Although it would possibly allow to have the bundled gems in the RubyGems directory structure instead of being scattered all...

Actually, you are right. The "executable" confused me, but the "executable" itself in `bin` dir would be still the same and the collision is on the .gemspec level.

My biggest concern with this PR is actually naming, because naming is hard. I think this demonstrates the issue a bit: https://github.com/rubygems/rubygems/pull/2909/commits/9c315143ea178a0626532883b9e8b6166e9fd257#diff-0c4b3e9996ebec38f191e68264665984R322 The commit 9c315143ea178a0626532883b9e8b6166e9fd257 moves the default gems into...

Chm, linting errors introduced long time ago preventing me from seeing CI results? That is disappointing :disappointed:

> so I'll remove the "except for alignment" thing on a separate PR. This makes my day much better, thx :wink:

@deivid-rodriguez I have removed the spaces, although I would reject such commit if it was upon me :blush: Thx for your help.

Testing this with Ruby, it seems I'll need to submit as PR against Ruby repository, or at least convince rbinstall to use `Gem::Installer` instead of the custom code used ATM...