diff options
| author | Bundlerbot <bot@bundler.io> | 2019-11-11 16:51:13 +0000 |
|---|---|---|
| committer | David RodrÃguez <deivid.rodriguez@riseup.net> | 2019-11-11 23:17:08 +0100 |
| commit | db79526704d76fc69027f94332a811387cf483e2 (patch) | |
| tree | 5bce1d9ad0baa4a93ed4df7d60a7c47c963fde08 /lib | |
| parent | 27ae25003bb410c42d313a6e2b216b94495bebeb (diff) | |
| download | bundler-db79526704d76fc69027f94332a811387cf483e2.tar.gz | |
Merge #7434
7434: More lazy loading and relative requires r=deivid-rodriguez a=deivid-rodriguez
### What was the end-user problem that led to this PR?
The problem was that testing the latest version of bundler as a vendored copy inside rubygems surfaced some more issues about requiring the wrong code, or activating default gems too early.
### What was your diagnosis of the problem?
My diagnosis was that we should follow similar techniques I've used in other PRs: delay as much as possible requiring default gems, and don't rely on the LOAD_PATH for internal requires.
### What is your fix for the problem, implemented in this PR?
My fix does just that, and also removes a bunch of `require`'s that were not necessary at all.
Co-authored-by: David RodrÃguez <deivid.rodriguez@riseup.net>
(cherry picked from commit 147b9aa079f531ba5bf216d6d694fce5d36cfbfc)
Diffstat (limited to 'lib')
| -rw-r--r-- | lib/bundler/gem_helper.rb | 1 |
1 files changed, 0 insertions, 1 deletions
diff --git a/lib/bundler/gem_helper.rb b/lib/bundler/gem_helper.rb index 81872b9429..46a6643233 100644 --- a/lib/bundler/gem_helper.rb +++ b/lib/bundler/gem_helper.rb @@ -1,6 +1,5 @@ # frozen_string_literal: true -require_relative "vendored_thor" unless defined?(Thor) require_relative "../bundler" require "shellwords" |