From 003207361a9ee39f886554da2d1ef66ac96d99ad Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Jonathan=20M=C3=BCller?= <git@foonathan.net>
Date: Mon, 19 Mar 2018 21:12:40 +0100
Subject: [PATCH] Properly handle unnecessary semicolons

Now fixes #45.
---
 src/libclang/cxtokenizer.cpp     |  3 +++
 src/libclang/parse_functions.cpp | 21 +++++++++++++--------
 test/cpp_namespace.cpp           |  3 +++
 3 files changed, 19 insertions(+), 8 deletions(-)

diff --git a/src/libclang/cxtokenizer.cpp b/src/libclang/cxtokenizer.cpp
index 35f075b..b502fdb 100644
--- a/src/libclang/cxtokenizer.cpp
+++ b/src/libclang/cxtokenizer.cpp
@@ -78,6 +78,9 @@ namespace
 
         simple_tokenizer tokenizer(tu, inc > 0 ? clang_getRange(loc, loc_after) :
                                                  clang_getRange(loc_after, loc));
+        if (tokenizer.size() == 0u)
+            return false;
+
         detail::cxstring spelling(clang_getTokenSpelling(tu, tokenizer[0u]));
         return spelling == token_str;
     }
diff --git a/src/libclang/parse_functions.cpp b/src/libclang/parse_functions.cpp
index 1a93792..72aee15 100644
--- a/src/libclang/parse_functions.cpp
+++ b/src/libclang/parse_functions.cpp
@@ -212,18 +212,15 @@ std::unique_ptr<cpp_entity> detail::parse_entity(const detail::parse_context& co
 
     if (!clang_isAttribute(clang_getCursorKind(cur)))
     {
-        auto msg = detail::format("unhandled cursor of kind '",
-                                  detail::get_cursor_kind_spelling(cur).c_str(), "'");
-        context.logger->log("libclang parser",
-                            format_diagnostic(severity::warning, detail::make_location(cur),
-                                              "unhandled cursor of kind '",
-                                              detail::get_cursor_kind_spelling(cur).c_str(), "'"));
-
         // build unexposed entity
-        auto                   name = detail::get_cursor_name(cur);
         detail::cxtokenizer    tokenizer(context.tu, context.file, cur);
         detail::cxtoken_stream stream(tokenizer, cur);
         auto                   spelling = detail::to_string(stream, stream.end());
+        if (spelling.begin() + 1 == spelling.end() && spelling.front().spelling == ";")
+            // unnecessary semicolon
+            return nullptr;
+
+        auto name = detail::get_cursor_name(cur);
 
         std::unique_ptr<cppast::cpp_entity> entity;
         if (name.empty())
@@ -233,6 +230,14 @@ std::unique_ptr<cpp_entity> detail::parse_entity(const detail::parse_context& co
                                                  name.c_str(), std::move(spelling));
 
         context.comments.match(*entity, cur);
+
+        auto msg = detail::format("unhandled cursor of kind '",
+                                  detail::get_cursor_kind_spelling(cur).c_str(), "'");
+        context.logger->log("libclang parser",
+                            format_diagnostic(severity::warning, detail::make_location(cur),
+                                              "unhandled cursor of kind '",
+                                              detail::get_cursor_kind_spelling(cur).c_str(), "'"));
+
         return entity;
     }
     else
diff --git a/test/cpp_namespace.cpp b/test/cpp_namespace.cpp
index e659e85..93a4313 100644
--- a/test/cpp_namespace.cpp
+++ b/test/cpp_namespace.cpp
@@ -37,6 +37,9 @@ namespace {}
 /// namespace f{
 /// }
 namespace e::f {}
+
+// unnecessary semicolon at end of file
+;
 )";
 
     auto file  = parse({}, "cpp_namespace.cpp", code);