diff --git a/vit/application.py b/vit/application.py
index 52870cb..93f7e2a 100644
--- a/vit/application.py
+++ b/vit/application.py
@@ -38,6 +38,7 @@
 from vit.action_manager import ActionManagerRegistry
 from vit.denotation import DenotationPopupLauncher
 from vit.pid_manager import PidManager
+from vit.markup import markup_to_str
 
 # NOTE: This entire class is a workaround for the fact that urwid catches the
 # 'ctrl l' keypress in its unhandled_input code, and prevents that from being
@@ -571,9 +572,7 @@ def search_loop_warning(self, hit, reverse=False):
         self.search_display_message(reverse)
 
     def reconstitute_markup_element_as_string(self, accum, markup):
-        if isinstance(markup, tuple):
-            _, markup = markup
-        return accum + markup
+        return accum + markup_to_str(markup)
 
     def reconstitute_markup_as_string(self, markup):
         if isinstance(markup, list):
diff --git a/vit/markup.py b/vit/markup.py
new file mode 100644
index 0000000..8ecd1bc
--- /dev/null
+++ b/vit/markup.py
@@ -0,0 +1,74 @@
+from __future__ import annotations
+
+from dataclasses import dataclass
+
+from urwid.util import calc_width
+
+
+@dataclass(frozen=True)
+class RawText:
+    data: bytes
+
+
+def markup_contains_raw(markup) -> bool:
+    if isinstance(markup, RawText):
+        return True
+    if isinstance(markup, list):
+        return any(markup_contains_raw(item) for item in markup)
+    if isinstance(markup, tuple):
+        _, inner = markup
+        return markup_contains_raw(inner)
+    return False
+
+
+def markup_to_bytes(markup, encoding: str = "utf-8"):
+    if isinstance(markup, RawText):
+        return markup.data
+    if isinstance(markup, bytes):
+        return markup
+    if isinstance(markup, str):
+        return markup.encode(encoding)
+    if isinstance(markup, list):
+        return [markup_to_bytes(item, encoding=encoding) for item in markup]
+    if isinstance(markup, tuple):
+        attr, inner = markup
+        return (attr, markup_to_bytes(inner, encoding=encoding))
+    return markup
+
+
+def normalize_markup(markup, encoding: str = "utf-8"):
+    if markup_contains_raw(markup):
+        return markup_to_bytes(markup, encoding=encoding)
+    return markup
+
+
+def markup_display_width(markup) -> int:
+    if markup is None:
+        return 0
+    if isinstance(markup, RawText):
+        return 0
+    if isinstance(markup, list):
+        return sum(markup_display_width(item) for item in markup)
+    if isinstance(markup, tuple):
+        _, inner = markup
+        return markup_display_width(inner)
+    if isinstance(markup, (str, bytes)):
+        return calc_width(markup, 0, len(markup))
+    return 0
+
+
+def markup_to_str(markup, encoding: str = "utf-8") -> str:
+    if markup is None:
+        return ""
+    if isinstance(markup, RawText):
+        return ""
+    if isinstance(markup, bytes):
+        return markup.decode(encoding, errors="ignore")
+    if isinstance(markup, str):
+        return markup
+    if isinstance(markup, list):
+        return "".join(markup_to_str(item, encoding=encoding) for item in markup)
+    if isinstance(markup, tuple):
+        _, inner = markup
+        return markup_to_str(inner, encoding=encoding)
+    return ""
diff --git a/vit/task_list.py b/vit/task_list.py
index 34d0d0a..df971a7 100644
--- a/vit/task_list.py
+++ b/vit/task_list.py
@@ -14,6 +14,7 @@
 from vit.list_batcher import ListBatcher
 from vit.formatter.project import Project as ProjectFormatter
 from vit.util import unicode_len
+from vit.markup import normalize_markup, markup_display_width
 
 
 REDUCE_COLUMN_WIDTH_LIMIT = 20
@@ -266,10 +267,12 @@ def update_column_width(self, idx, current_width, new_width):
 
     def build_row_column(self, formatted_value):
         if isinstance(formatted_value, tuple):
-            return formatted_value
+            width, text_markup = formatted_value
+            return width, normalize_markup(text_markup)
         else:
-            width = unicode_len(formatted_value) if formatted_value else 0
-            return width, formatted_value
+            text_markup = normalize_markup(formatted_value)
+            width = markup_display_width(text_markup) if formatted_value else 0
+            return width, text_markup
 
     def subproject_indentable(self):
         return self.config.subproject_indentable and self.report['subproject_indentable']