svgLib: handle arc commands with bool flags not separated by space/comma

Some SVG authoring tool write arc commands without any space or comma around the boolean 'large-arc' and 'sweep' flags, leading our svg path parser to choke.
This makes the path parser smarter by special-casing arc command parsing so that it only consumes one character ('0' or '1') for these special boolean flags.

Lib/fontTools/svgLib/path/parser.py

@@ -13,18 +13,81 @@ import re
 
 
 COMMANDS = set('MmZzLlHhVvCcSsQqTtAa')
+ARC_COMMANDS = set("Aa")
 UPPERCASE = set('MZLHVCSQTA')
 
 COMMAND_RE = re.compile("([MmZzLlHhVvCcSsQqTtAa])")
-FLOAT_RE = re.compile(r"[-+]?[0-9]*\.?[0-9]+(?:[eE][-+]?[0-9]+)?")
+FLOAT_RE = re.compile(
+    r"[-+]?"  # optional sign
+    r"(?:"
+    r"(?:0|[1-9][0-9]*)(?:\.[0-9]+(?:[eE][-+]?[0-9]+)?)?"  # int/float
+    r"|"
+    r"(?:\.[0-9]+(?:[eE][-+]?[0-9]+)?)"  # float with leading dot (e.g. '.42')
+    r")"
+)
+BOOL_RE = re.compile("^[01]")
+SEPARATOR_RE = re.compile(f"[, \t]")
 
 
 def _tokenize_path(pathdef):
+    arc_cmd = None
     for x in COMMAND_RE.split(pathdef):
         if x in COMMANDS:
+            arc_cmd = x if x in ARC_COMMANDS else None
             yield x
-        for token in FLOAT_RE.findall(x):
-            yield token
+            continue
+
+        if arc_cmd:
+            try:
+                yield from _tokenize_arc_arguments(x)
+            except ValueError as e:
+                raise ValueError(f"Invalid arc command: '{arc_cmd}{x}'") from e
+        else:
+            for token in FLOAT_RE.findall(x):
+                yield token
+
+
+ARC_ARGUMENT_TYPES = (
+    ("rx", FLOAT_RE),
+    ("ry", FLOAT_RE),
+    ("x-axis-rotation", FLOAT_RE),
+    ("large-arc-flag", BOOL_RE),
+    ("sweep-flag", BOOL_RE),
+    ("x", FLOAT_RE),
+    ("y", FLOAT_RE),
+)
+
+
+def _tokenize_arc_arguments(arcdef):
+    raw_args = [s for s in SEPARATOR_RE.split(arcdef) if s]
+    if not raw_args:
+        raise ValueError(f"Not enough arguments: '{arcdef}'")
+    raw_args.reverse()
+
+    i = 0
+    while raw_args:
+        arg = raw_args.pop()
+
+        name, pattern = ARC_ARGUMENT_TYPES[i]
+        match = pattern.search(arg)
+        if not match:
+            raise ValueError(f"Invalid argument for '{name}' parameter: {arg!r}")
+
+        j, k = match.span()
+        yield arg[j:k]
+        arg = arg[k:]
+
+        if arg:
+            raw_args.append(arg)
+
+        # wrap around every 7 consecutive arguments
+        if i == 6:
+            i = 0
+        else:
+            i += 1
+
+    if i != 0:
+        raise ValueError(f"Not enough arguments: '{arcdef}'")
 
 
 def parse_path(pathdef, pen, current_pos=(0, 0), arc_class=EllipticalArc):

Tests/svgLib/path/parser_test.py

@@ -353,3 +353,86 @@ def test_arc_pen_with_arcTo():
     ]
 
     assert pen.value == expected
+
+
+@pytest.mark.parametrize(
+    "path, expected",
+    [
+        (
+            "M1-2A3-4-1.0 01.5.7",
+            [
+                ("moveTo", ((1.0, -2.0),)),
+                ("arcTo", (3.0, -4.0, -1.0, False, True, (0.5, 0.7))),
+                ("endPath", ()),
+            ],
+        ),
+        (
+            "M21.58 7.19a2.51 2.51 0 10-1.77-1.77",
+            [
+                ("moveTo", ((21.58, 7.19),)),
+                ("arcTo", (2.51, 2.51, 0.0, True, False, (19.81, 5.42))),
+                ("endPath", ()),
+            ],
+        ),
+        (
+            "M22 12a25.87 25.87 0 00-.42-4.81",
+            [
+                ("moveTo", ((22.0, 12.0),)),
+                ("arcTo", (25.87, 25.87, 0.0, False, False, (21.58, 7.19))),
+                ("endPath", ()),
+            ],
+        ),
+        (
+            "M0,0 A1.2 1.2 0 012 15.8",
+            [
+                ("moveTo", ((0.0, 0.0),)),
+                ("arcTo", (1.2, 1.2, 0.0, False, True, (2.0, 15.8))),
+                ("endPath", ()),
+            ],
+        ),
+        (
+            "M12 7a5 5 0 105 5 5 5 0 00-5-5",
+            [
+
+                ("moveTo", ((12.0, 7.0),)),
+                ("arcTo", (5.0, 5.0, 0.0, True, False, (17.0, 12.0))),
+                ("arcTo", (5.0, 5.0, 0.0, False, False, (12.0, 7.0))),
+                ("endPath", ()),
+            ],
+        )
+    ],
+)
+def test_arc_flags_without_spaces(path, expected):
+    pen = ArcRecordingPen()
+    parse_path(path, pen)
+    assert pen.value == expected
+
+
+@pytest.mark.parametrize(
+    "path", ["A", "A0,0,0,0,0,0", "A 0 0 0 0 0 0 0 0 0 0 0 0 0"]
+)
+def test_invalid_arc_not_enough_args(path):
+    pen = ArcRecordingPen()
+    with pytest.raises(ValueError, match="Invalid arc command") as e:
+        parse_path(path, pen)
+
+    assert isinstance(e.value.__cause__, ValueError)
+    assert "Not enough arguments" in str(e.value.__cause__)
+
+
+def test_invalid_arc_argument_value():
+    pen = ArcRecordingPen()
+    with pytest.raises(ValueError, match="Invalid arc command") as e:
+        parse_path("M0,0 A0,0,0,2,0,0,0", pen)
+
+    cause = e.value.__cause__
+    assert isinstance(cause, ValueError)
+    assert "Invalid argument for 'large-arc-flag' parameter: '2'" in str(cause)
+
+    pen = ArcRecordingPen()
+    with pytest.raises(ValueError, match="Invalid arc command") as e:
+        parse_path("M0,0 A0,0,0,0,-2.0,0,0", pen)
+
+    cause = e.value.__cause__
+    assert isinstance(cause, ValueError)
+    assert "Invalid argument for 'sweep-flag' parameter: '-2.0'" in str(cause)