You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Ref: #402
Ref: https://bugs.ruby-lang.org/issues/10222
Ref: ruby/ruby@5754f15
Ref: ruby/ruby@b6d3927
Up to Ruby 2.3, `require` would resolve symlinks, but `require_relative` wouldn't:
```ruby
require 'fileutils'
FileUtils.mkdir_p("realpath")
File.write("realpath/a.rb", "p :a_loaded")
File.symlink("realpath", "symlink") rescue nil
$LOAD_PATH.unshift(File.realpath(__dir__) + "/symlink")
require "a.rb" # load symlink/a.rb in 2.3 and older, load realpath/a.rb on 2.4 and newer
require_relative "realpath/a.rb" # noop on 2.4+
```
This would easily cause double loading issue when `require` and `require_relative`
were mixed, but was fixed in 2.4 (https://bugs.ruby-lang.org/issues/10222).
The problem is that `Bootsnap` kinda negated this fix, because `realpath()`
wouldn't be applied to absolute paths:
```ruby
require 'fileutils'
FileUtils.mkdir_p("realpath")
File.write("realpath/a.rb", "p :a_loaded")
File.symlink("realpath", "symlink") rescue nil
$LOAD_PATH.unshift(File.realpath(__dir__) + "/symlink")
require File.expand_path("symlink/a.rb") # load symlink/a.rb in 3.0 and older, load realpath/a.rb on 3.1 and newer
require_relative "realpath/a.rb" # noop on 3.1+
```
And for performance reasons, Bootsnap tried really hard not to call `realpath`,
as it's a syscall, instead it used `expand_path`, which is entirely in use
space and doesn't reach to the file system. So if you had a `symlink` in
`$LOAD_PATH`, `bootcsnap` would perpetuate this bug, which led to the
addition of #136.
This was ultimately fixed in Ruby 3.1 (https://bugs.ruby-lang.org/issues/17885),
now `realpath` is applied even on absolute paths.
While `realpath` is indeed expensive, I think the performance impact is ok if
we only call it for `$LOAD_PATH` members, rather than for all requirable files.
So if you have X gems, it's going to be more or less X `realpath` calls.
It would stay a problem if a gem actually contained symlinks and used
`require_relative`, but it's quite the stretch, and with 3.1 now
handling it, it's not worth keeping such workaround.
See: #402
0 commit comments