Add IFromBytes

As a counterpoint to IFromUnicode.

This is useful when you know the type of data you're dealing with, and conversions are either trivially non-lossy (or will fail anyway) or are not necessary.

The numeric fields as well as Bytes/Line and URI, DottedName and Id implement this, the last three because they are native string lines that want to accept bytes on Python 2 anyway.

On Python 2, lxml likes to produce byte strings "to save memory" even when parsing unicode and given proper XML encoding information, so this can be very helpful there.

1 files changed, 13 insertions, 1 deletions

diff --git a/src/zope/schema/_bootstrapinterfaces.py b/src/zope/schema/_bootstrapinterfaces.py
index 110aa4f..4f0419e 100644
--- a/src/zope/schema/_bootstrapinterfaces.py
+++ b/src/zope/schema/_bootstrapinterfaces.py
@@ -282,11 +282,23 @@ class IFromUnicode(zope.interface.Interface):
     values.
     """
 
-    def fromUnicode(str):
+    def fromUnicode(value):
         """Convert a unicode string to a value.
         """
 
 
+class IFromBytes(zope.interface.Interface):
+    """
+    Parse a byte string to a value.
+
+    If the string needs to be decoded, decoding is done using UTF-8.
+    """
+
+    def fromBytes(value):
+        """Convert a byte string to a value.
+        """
+
+
 class IContextAwareDefaultFactory(zope.interface.Interface):
     """A default factory that requires a context.