fix _utf8, introduce _unicode (#23)

assetman/compile.py

@@ -302,7 +302,7 @@ def _build_manifest_helper(static_dir, src_paths, static_url_prefix, manifest):
     for src_path in src_paths:
         # Make sure every source path at least has the skeleton entry
         rel_src_path = make_relative_static_path(static_dir, src_path)
-        logging.info('_build_manifest_helper %s (crrent %s)', src_path, manifest.assets.get(rel_src_path))
+        logging.info('_build_manifest_helper %s (current %s)', src_path, manifest.assets.get(rel_src_path))
         manifest.assets.setdefault(rel_src_path, empty_asset_entry())
         for dep_path in iter_deps(static_dir, src_path, static_url_prefix):
             logging.info('%s > dependency %s', src_path, dep_path)

assetman/managers.py

@@ -5,7 +5,7 @@
 import functools
 import hashlib
 
-from assetman.tools import get_shard_from_list, _utf8
+from assetman.tools import get_shard_from_list, _unicode
 from assetman.manifest import Manifest
 
 class AssetManager(object):
@@ -46,7 +46,7 @@ def __init__(self, rel_url_text, local=False, include_tag=True, src_path=None, s
         Any extra kwargs will be interpreted as extra HTML params to include
         on the rendered element.
         """
-        self.rel_urls = [_f for _f in _utf8(rel_url_text).split() if _f]
+        self.rel_urls = [_f for _f in _unicode(rel_url_text).split() if _f]
         self.local = local
         self.include_tag = include_tag
         self.src_path = src_path
@@ -100,7 +100,7 @@ def render_attrs(self):
         """Returns this asset block's attrs as an HTML string. Includes a
         leading space.
         """
-        attrs = ' '.join('%s=%r' % (attr, _utf8(val))
+        attrs = ' '.join('%s=%r' % (attr, _unicode(val))
                          for attr, val in self.attrs.items())
         return ' ' + attrs if attrs else ''
 

assetman/tools.py

@@ -19,14 +19,27 @@ def _crc(key):
     return binascii.crc32(_utf8(key)) & 0xffffffff
 
 def _utf8(s):
-    """encode a unicode string as utf-8"""
+    """encode a string as utf-8, returning bytes"""
     if isinstance(s, str):
-        return s.encode("utf-8").decode()
+        return s.encode("utf-8")
     if isinstance(s, bytes):
-        return s.decode("utf-8")
-    assert isinstance(s, str), "_utf8 expected a str, not %r" % type(s)
+        try:
+            s.decode("utf-8")
+        except UnicodeDecodeError:
+            raise AssertionError("Invalid encoding. _utf8 expected a str or utf-8 encoded bytes")
+    assert isinstance(s, bytes), "_utf8 expected a str or utf-8 encoded bytes, not %r" % type(s)
     return s
 
+def _unicode(value):
+    """decode utf-8 bytes, returning string"""
+    if isinstance(value, bytes):
+        try:
+            return value.decode("utf-8")
+        except UnicodeDecodeError:
+            raise AssertionError("Invalid encoding. _unicode expected a str or utf-8 encoded bytes")
+    assert isinstance(value, str), "_unicode expected a str or utf-8 encoded bytes, not %r" % type(value)
+    return value
+
 def iter_template_paths(template_dirs, template_ext):
     """Walks each directory in the given list of template directories,
     yielding the path to each template found.

pyproject.toml

@@ -5,7 +5,7 @@ build-backend = "setuptools.build_meta"
 [project]
 name = "pyassetman"
 description = "assetman assetmanager"
-version = "0.3.0rc2"
+version = "0.3.0rc3"
 authors = [
   { name="Will McCutchen", email="wm@bit.ly" },
 ]