fix(util): split_into_lines merges a word split across 2+ colored spans

beets/util/layout.py

@@ -70,6 +70,15 @@ def split_into_lines(string: str, first_width: int, width: int) -> list[str]:
         words = string.split()
     else:
         # Use a regex to find escapes and the text within them.
+        # `pretext` is empty for every match after the first, since the
+        # previous match's `posttext` (being a greedy [^ESC]* group) already
+        # consumed all plain text up to this escape sequence. Whether that
+        # represents a real word boundary depends on whether the text
+        # preceding this escape actually ended in whitespace, so we track it
+        # explicitly across iterations instead of assuming pretext-empty
+        # always means "boundary" (which is only true for the very first
+        # match, i.e. the start of the string).
+        prev_ends_with_space = True
         for m in ESC_TEXT_REGEX.finditer(string):
             # m contains four groups:
             # pretext - any text before escape sequence
@@ -88,8 +97,7 @@ def split_into_lines(string: str, first_width: int, width: int) -> list[str]:
                     # Pretext ended mid-word, ensure next word
                     pass
             else:
-                # pretext empty, treat as if there is a space before
-                space_before_text = True
+                space_before_text = prev_ends_with_space
             if m.group("text")[0] == " ":
                 # First character of the text is a space
                 space_before_text = True
@@ -128,6 +136,15 @@ def split_into_lines(string: str, first_width: int, width: int) -> list[str]:
             else:
                 # Add any words after escape sequence
                 words += m.group("posttext").split()
+            # Determine the boundary state for the next match's pretext-empty
+            # case: look at whatever content immediately precedes the next
+            # escape sequence -- posttext if there is any, otherwise text
+            # itself (posttext being empty means reset is followed directly
+            # by the next escape, with nothing in between).
+            if m.group("posttext"):
+                prev_ends_with_space = m.group("posttext")[-1] == " "
+            else:
+                prev_ends_with_space = m.group("text")[-1] == " "
     result: list[str] = []
     next_substr = ""
     # Iterate over all words.

docs/changelog.rst

@@ -95,6 +95,9 @@ Bug fixes
   valid date/time string" error instead of crashing with an uncaught
   ``KeyError``. A ``|`` was being accepted as a relative-date unit due to a
   regular expression character-class typo.
+- Fix word wrapping of colored diff output for a word containing two or more
+  separately-highlighted spans, which was incorrectly split into two words at
+  the second highlighted span.
 
 ..
     For plugin developers

test/util/test_layout.py

@@ -1,5 +1,6 @@
 from unittest import TestCase
 
+from beets.util.color import uncolorize
 from beets.util.layout import split_into_lines
 
 
@@ -28,3 +29,14 @@ def test_split_into_lines(self):
         split_txt = ["\x1b[31mtest\x1b[39;49;00mt", "est", "test", "test"]
         txt = split_into_lines(colored_text, 5, 5)
         assert txt == split_txt
+
+    def test_split_into_lines_two_spans_in_one_word(self):
+        # A single word containing two separately-colored, non-adjacent
+        # spans (e.g. two highlighted typo fixes within one word) must stay
+        # one word, not get split at the second colored span.
+        red, reset = "\x1b[31m", "\x1b[39;49;00m"
+        colored_text = f"extra{red}A{reset}ordin{red}B{reset}ary"
+        txt = split_into_lines(colored_text, 100, 100)
+        assert len(txt) == 1
+        assert " " not in uncolorize(txt[0])
+        assert uncolorize(txt[0]) == "extraAordinBary"