From c905032a39db900f5b2522572dc8dcf6792f8e78 Mon Sep 17 00:00:00 2001
From: Ryan Malloy <ryan@supported.systems>
Date: Sat, 11 Jul 2026 18:46:29 -0600
Subject: [PATCH] Fix DSN tokenizer to accept KiCad net names like /*52
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Specctra's SpecCharASCII includes / and *, and an Identifier may start with /,
so KiCad emits hierarchical net names such as /*52 and /53. The tokenizer
treated any /* as a block-comment start and raised 'unterminated comment' when
no */ followed — rejecting real KiCad DSN. Match FreeRouting's JFlex rule-order
resolution: /* is a comment only when a closing */ exists; otherwise it is an
ordinary name run. Validated against a KiCad 10.0.4 pcbnew-exported DSN (78 nets
incl. /*52, /53).
---
 src/freeroute/dsn/tokenizer.py | 18 ++++++++++++------
 tests/dsn/test_tokenizer.py    | 13 ++++++++++---
 2 files changed, 22 insertions(+), 9 deletions(-)

diff --git a/src/freeroute/dsn/tokenizer.py b/src/freeroute/dsn/tokenizer.py
index 1238a53..a671dc2 100644
--- a/src/freeroute/dsn/tokenizer.py
+++ b/src/freeroute/dsn/tokenizer.py
@@ -126,13 +126,19 @@ def tokenize(text: str, quote_chars: str = _QUOTES) -> list[Token]:
             eol = text.find("\n", i)
             i = n if eol == -1 else eol
             continue
-        if ch == "/" and i + 1 < n and text[i + 1] == "*":  # block comment
+        # `/* ... */` is a block comment ONLY when a closing `*/` exists.
+        # SpecCharASCII includes `/` and `*`, and an Identifier may start with
+        # `/`, so KiCad emits net names like `/*52`. FreeRouting's JFlex scanner
+        # resolves the ambiguity by longest-match / rule-order: with no closing
+        # `*/`, the comment rule fails and the run is read as a name. Mirror
+        # that — an unclosed `/*` is an ordinary token, not an error.
+        if ch == "/" and i + 1 < n and text[i + 1] == "*":
             end = text.find("*/", i + 2)
-            if end == -1:
-                raise DsnSyntaxError(f"unterminated /* */ comment at line {line}")
-            line += text.count("\n", i, end)
-            i = end + 2
-            continue
+            if end != -1:
+                line += text.count("\n", i, end)
+                i = end + 2
+                continue
+            # no closing `*/` — fall through and read `/*...` as a normal token
 
         # Brackets -----------------------------------------------------------
         if ch == "(":
diff --git a/tests/dsn/test_tokenizer.py b/tests/dsn/test_tokenizer.py
index f6bb8f4..cb01314 100644
--- a/tests/dsn/test_tokenizer.py
+++ b/tests/dsn/test_tokenizer.py
@@ -117,9 +117,16 @@ def test_unterminated_string_raises():
         tokenize('(a "no end')
 
 
-def test_unterminated_block_comment_raises():
-    with pytest.raises(DsnSyntaxError):
-        tokenize("(a /* no end")
+def test_unclosed_block_comment_is_name_token():
+    # KiCad emits hierarchical net names like `/*52` (SpecCharASCII includes
+    # `/` and `*`). With no closing `*/`, `/*...` is a name run, not an
+    # unterminated comment — matching FreeRouting's JFlex rule-order resolution.
+    assert texts("(net /*52)") == ["(", "net", "/*52", ")"]
+
+
+def test_closed_block_comment_is_stripped():
+    # A properly closed `/* ... */` is still a comment and gets dropped.
+    assert texts("(a /* c */ b)") == ["(", "a", "b", ")"]
 
 
 def test_hyphenated_pin_ref_is_single_atom():