Issue #16978 has been updated by jeremyevans0 (Jeremy Evans).


vo.x (Vit Ondruch) wrote in #note-13:
> Dan0042 (Daniel DeLorme) wrote in #note-12:
> > I think this is a bug and should be fixed, but IMO the proper fix is to use realpath for `__FILE__`
> 
> This is though one thinking about this again, but you are probably right. Maybe the biggest concern is that the behavior is inconsistent. If the `require` and `__FILE__` used `realpath`, I think the example in #16978-11 would work as you said.

It's not difficult to make `__FILE__` use `realpath`:

```diff
@@ -10307,6 +10307,14 @@ numparam_nested_p(struct parser_params *p)
     return 0;
 }

+VALUE rb_realpath_internal(VALUE basedir, VALUE path, int strict);
+
+static VALUE
+rb_realpath__FILE__(VALUE file)
+{
+    return rb_realpath_internal(Qnil, file, 1);
+}
+
 static NODE*
 gettable(struct parser_params *p, ID id, const YYLTYPE *loc)
 {
@@ -10326,8 +10334,19 @@ gettable(struct parser_params *p, ID id, const YYLTYPE *loc)
            VALUE file = p->ruby_sourcefile_string;
            if (NIL_P(file))
                file = rb_str_new(0, 0);
-           else
-               file = rb_str_dup(file);
+           else if (compile_for_eval) {
+                file = rb_str_dup(file);
+            }
+           else {
+                int state;
+               VALUE realpath_file = rb_protect(rb_realpath__FILE__, file, &state);
+                if (state) {
+                    file = rb_str_dup(file);
+                }
+                else {
+                    file = realpath_file;
+                }
+            }
            node = NEW_STR(file, loc);
             RB_OBJ_WRITTEN(p->ast, Qnil, file);
        }
```

Unfortunately, if `__FILE__` uses `realpath`, unless you also change the behavior of `$0` (which I'm fairly sure would not be acceptable), you break the common idiom for executing code if the file is executed instead of required:

```ruby
if __FILE__ == $0
  do_something
end
```



----------------------------------------
Feature #16978: Ruby should not use realpath for __FILE__
https://bugs.ruby-lang.org/issues/16978#change-92682

* Author: vo.x (Vit Ondruch)
* Status: Open
* Priority: Normal
* Assignee: nobu (Nobuyoshi Nakada)
----------------------------------------
This is the simplest test case:

~~~
$ mkdir a

$ echo "puts __FILE__" > a/test.rb

$ ln -s a b

$ ruby -Ib -e "require 'test'"
/builddir/a/test.rb
~~~

This behavior is problematic, because Ruby should not know nothing about the `a` directory. It was not instructed to use it. I should always refer to the file using the original path and do not dig into the underlying details, otherwise depending on file system setup, one might be forced to used `File.realpath` everywhere trying to use `__FILE__`.



-- 
https://bugs.ruby-lang.org/

Unsubscribe: <mailto:ruby-core-request / ruby-lang.org?subject=unsubscribe>
<http://lists.ruby-lang.org/cgi-bin/mailman/options/ruby-core>