Bug #15206: require_relative in std_lib

• Status: Closed
• Assignee: -
• Target version: -
• Backport: 2.3: UNKNOWN, 2.4: UNKNOWN, 2.5: UNKNOWN

[marcandre (Marc-Andre Lafortune)]

I noticed that there are some require for internal files in lib/. Ideally, they would be using require_relative instead. This is faster and more explicit.

Note: require_relative had a potential issue with symlinks that was fixed in 2.5, so libraries that might be loaded from earlier Ruby, like rubygems, can not yet be updated.

I've created a PR and would like to know if there are comments / objections / things I'm missing.

[MSP-Greg (Greg L)]

The Appveyor run on the PR passed.

Appveyor testing has been intermittently failing. With parallel/multi testing (test-all and/or test-spec), there can be a lot more files being checked and possibly loaded. Considering more than one CI VM may be exist on a system, and they may share drives, anything that lowers disk I/O is good.

[shyouhei (Shyouhei Urabe)]

• Is duplicate of Feature #8781: Use require_relative() instead of require() if possible added

[nobu (Nobuyoshi Nakada)]

• Status changed from Open to Closed

Applied in changeset trunk|r65511.

---

lib/cgi/util.rb: use require

• lib/cgi/util.rb: reverted r65505. require_relative does not
  work for cgi/escape.so, as extension libraries are placed in
  other directories than ruby libraries. [Bug #15206]

[stomar (Marcus Stollsteimer)]

For lib/racc/parser.rb I reverted this in r65514 for the require of racc/cparse.so (similar to r65511 for cgi/escape).

Possibly there might be other C extensions that are required from within /lib, where require_relative would not work?

(I only had a quick look into /lib and /ext, and racc was an obvious other case, besides cgi, where this problem might occur.)

PS. The LoadError for racc/cparse is not that obvious because it is rescued and there is a Ruby fallback.

[marcandre (Marc-Andre Lafortune)]

Oh, good catch, I should have thought of that...

I modified my script to check if the path with .rb actually exists and it came back with only these two exceptions.

[znz (Kazuhiro NISHIYAMA)]

I revert examples in comment and replace numeric features too at r65548.