Revert "Fixed #20296 -- Allowed SafeData and EscapeData to be lazy"

This reverts commit 2ee447fb5f.

That commit introduced a regression (#21882) and didn't really
do what it was supposed to: while it did delay the evaluation
of lazy objects passed to mark_safe(), they weren't actually
marked as such so they could end up being escaped twice.

Refs #21882.

django/utils/safestring.py

@@ -4,7 +4,7 @@ without further escaping in HTML. Marking something as a "safe string" means
 that the producer of the string has already turned characters that should not
 be interpreted by the HTML engine (e.g. '<') into the appropriate entities.
 """
-from django.utils.functional import curry, Promise, allow_lazy
+from django.utils.functional import curry, Promise
 from django.utils import six
 
 
@@ -16,14 +16,14 @@ class EscapeBytes(bytes, EscapeData):
     """
     A byte string that should be HTML-escaped when output.
     """
-    __new__ = allow_lazy(bytes.__new__, bytes)
+    pass
 
 
 class EscapeText(six.text_type, EscapeData):
     """
     A unicode string object that should be HTML-escaped when output.
     """
-    __new__ = allow_lazy(six.text_type.__new__, six.text_type)
+    pass
 
 if six.PY3:
     EscapeString = EscapeText
@@ -48,8 +48,6 @@ class SafeBytes(bytes, SafeData):
     A bytes subclass that has been specifically marked as "safe" (requires no
     further escaping) for HTML output purposes.
     """
-    __new__ = allow_lazy(bytes.__new__, bytes)
-
     def __add__(self, rhs):
         """
         Concatenating a safe byte string with another safe byte string or safe
@@ -83,8 +81,6 @@ class SafeText(six.text_type, SafeData):
     A unicode (Python 2) / str (Python 3) subclass that has been specifically
     marked as "safe" for HTML output purposes.
     """
-    __new__ = allow_lazy(six.text_type.__new__, six.text_type)
-
     def __add__(self, rhs):
         """
         Concatenating a safe unicode string with another safe byte string or