From 6044e19e7eb4ff978539b579b3bf5d522c105b60 Mon Sep 17 00:00:00 2001
From: Andrew Hundt <ATHundt@gmail.com>
Date: Wed, 1 Jul 2026 10:39:06 -0400
Subject: [PATCH] fix(extract): restore HDL call edges

Add a shared Verilog/SystemVerilog leaf-identifier resolver for wrapped subroutine and tf_call nodes, and include SystemVerilog subroutine_call nodes exposed by the vendored grammar.

Evidence: AST probes confirmed the wrapper shapes, failed-first canaries produced 252 passed / 2 failed before the fix, and the focused extraction suite now passes 254/254.
Signed-off-by: Andrew Hundt <ATHundt@gmail.com>
---
 internal/cbm/extract_calls.c | 31 +++++++++++++++++++++++++++++++
 internal/cbm/lang_specs.c    |  2 +-
 tests/test_extraction.c      | 36 ++++++++++++++++++++++++++++++++++++
 3 files changed, 68 insertions(+), 1 deletion(-)

diff --git a/internal/cbm/extract_calls.c b/internal/cbm/extract_calls.c
index caa41dc43..72dd1494f 100644
--- a/internal/cbm/extract_calls.c
+++ b/internal/cbm/extract_calls.c
@@ -834,6 +834,33 @@ static char *extract_vhdl_callee(CBMArena *a, TSNode node, const char *source, c
     return NULL;
 }
 
+static char *extract_first_leaf_identifier(CBMArena *a, TSNode node, const char *source) {
+    TSNode cur = node;
+    enum { CALLEE_LEAF_DESCENT_LIMIT = 8 };
+    for (int depth = 0; depth < CALLEE_LEAF_DESCENT_LIMIT && !ts_node_is_null(cur); depth++) {
+        const char *k = ts_node_type(cur);
+        if (strcmp(k, "simple_identifier") == 0 || strcmp(k, "identifier") == 0 ||
+            strcmp(k, "word") == 0 || strcmp(k, "name") == 0 || strcmp(k, "qid") == 0) {
+            char *t = cbm_node_text(a, cur, source);
+            return (t && t[0]) ? t : NULL;
+        }
+        if (ts_node_named_child_count(cur) == 0) {
+            return NULL;
+        }
+        cur = ts_node_named_child(cur, 0);
+    }
+    return NULL;
+}
+
+// HDL callees are often wrapped by subroutine/tf/hierarchical identifier nodes.
+static char *extract_hdl_callee(CBMArena *a, TSNode node, const char *source, const char *nk) {
+    if (strcmp(nk, "function_subroutine_call") != 0 && strcmp(nk, "subroutine_call") != 0 &&
+        strcmp(nk, "tf_call") != 0 && strcmp(nk, "system_tf_call") != 0) {
+        return NULL;
+    }
+    return extract_first_leaf_identifier(a, node, source);
+}
+
 // Make builtins are represented as function_call nodes; $(shell ...) is shell_function.
 static char *extract_make_callee(CBMArena *a, TSNode node, const char *source, const char *nk) {
     if (strcmp(nk, "shell_function") == 0) {
@@ -943,6 +970,10 @@ static char *extract_callee_lang_specific(CBMArena *a, TSNode node, const char *
         char *c = extract_vhdl_callee(a, node, source, nk);
         return c ? c : extract_scripting_callee(a, node, source, lang, nk);
     }
+    if (lang == CBM_LANG_VERILOG || lang == CBM_LANG_SYSTEMVERILOG) {
+        char *c = extract_hdl_callee(a, node, source, nk);
+        return c ? c : extract_scripting_callee(a, node, source, lang, nk);
+    }
 
     if (lang == CBM_LANG_CLOJURE || lang == CBM_LANG_COMMONLISP || lang == CBM_LANG_SCHEME ||
         lang == CBM_LANG_FENNEL || lang == CBM_LANG_RACKET || lang == CBM_LANG_EMACSLISP) {
diff --git a/internal/cbm/lang_specs.c b/internal/cbm/lang_specs.c
index 28c3d5c1e..120618816 100644
--- a/internal/cbm/lang_specs.c
+++ b/internal/cbm/lang_specs.c
@@ -1384,7 +1384,7 @@ static const char *systemverilog_class_types[] = {"class_declaration",
                                                   "type_declaration",
                                                   NULL};
 static const char *systemverilog_call_types[] = {"function_subroutine_call", "system_tf_call",
-                                                 "method_call", NULL};
+                                                 "subroutine_call", "method_call", NULL};
 static const char *systemverilog_import_types[] = {
     "package_import_declaration", "extends",  "import",     "include",
     "include_statement",          "instance", "use_clause", NULL};
diff --git a/tests/test_extraction.c b/tests/test_extraction.c
index 4e6fd809e..d71639a52 100644
--- a/tests/test_extraction.c
+++ b/tests/test_extraction.c
@@ -1260,6 +1260,40 @@ TEST(vhdl_function_call_edge) {
     PASS();
 }
 
+TEST(verilog_function_call_edge) {
+    CBMFileResult *r = extract("module m;\n"
+                               "  function integer foo;\n"
+                               "    input integer x;\n"
+                               "    begin foo = x; end\n"
+                               "  endfunction\n"
+                               "  integer y;\n"
+                               "  initial begin\n"
+                               "    y = foo(1);\n"
+                               "  end\n"
+                               "endmodule\n",
+                               CBM_LANG_VERILOG, "t", "mod.v");
+    ASSERT_NOT_NULL(r);
+    ASSERT_FALSE(r->has_error);
+    ASSERT(has_call_exact(r, "foo"));
+    cbm_free_result(r);
+    PASS();
+}
+
+TEST(systemverilog_subroutine_call_edge) {
+    CBMFileResult *r = extract("module m;\n"
+                               "  function int foo(); return 1; endfunction\n"
+                               "  initial begin\n"
+                               "    foo();\n"
+                               "  end\n"
+                               "endmodule\n",
+                               CBM_LANG_SYSTEMVERILOG, "t", "mod.sv");
+    ASSERT_NOT_NULL(r);
+    ASSERT_FALSE(r->has_error);
+    ASSERT(has_call_exact(r, "foo"));
+    cbm_free_result(r);
+    PASS();
+}
+
 /* --- Fortran --- */
 TEST(fortran_function) {
     /* Fortran subroutine name extraction is incomplete — just verify no crash */
@@ -3456,6 +3490,8 @@ SUITE(extraction) {
     RUN_TEST(nickel_function_application_edge);
     RUN_TEST(func_function_application_edge);
     RUN_TEST(vhdl_function_call_edge);
+    RUN_TEST(verilog_function_call_edge);
+    RUN_TEST(systemverilog_subroutine_call_edge);
     RUN_TEST(fortran_function);
 
     /* OOP/Systems variants */