From 3482a0846577ffdfdefcd2e9431a455ca5cd7a40 Mon Sep 17 00:00:00 2001
From: yhirose <yuji.hirose.bug@gmail.com>
Date: Tue, 27 Aug 2019 20:08:48 -0400
Subject: [PATCH] Fix #67

---
 peglib.h     | 42 ++++++++++++++++++++++++++++--------------
 test/test.cc | 45 +++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 73 insertions(+), 14 deletions(-)

diff --git a/peglib.h b/peglib.h
index 7192834..1ee75d2 100644
--- a/peglib.h
+++ b/peglib.h
@@ -432,6 +432,23 @@ inline std::pair<size_t, size_t> line_info(const char* start, const char* cur) {
     return std::make_pair(no, col);
 }
 
+/*
+* String tag
+*/
+#ifndef PEGLIB_NO_CONSTEXPR_SUPPORT
+inline constexpr unsigned int str2tag(const char* str, int h = 0) {
+    return !str[h] ? 5381 : (str2tag(str, h + 1) * 33) ^ static_cast<unsigned char>(str[h]);
+}
+
+namespace udl {
+
+inline constexpr unsigned int operator "" _(const char* s, size_t) {
+    return str2tag(s);
+}
+
+}
+#endif
+
 /*
 * Semantic values
 */
@@ -452,6 +469,10 @@ struct SemanticValues : protected std::vector<any>
     // Definition name
     const std::string& name() const { return name_; }
 
+#ifndef PEGLIB_NO_CONSTEXPR_SUPPORT
+    std::vector<unsigned int> tags;
+#endif
+
     // Line number and column at which the matched string is
     std::pair<size_t, size_t> line_info() const {
         return peg::line_info(ss, s_);
@@ -860,6 +881,7 @@ public:
         auto& sv = *value_stack[value_stack_size++];
         if (!sv.empty()) {
             sv.clear();
+            sv.tags.clear();
         }
         sv.reset();
         sv.path = path;
@@ -961,6 +983,7 @@ public:
             i += len;
         }
         sv.insert(sv.end(), chldsv.begin(), chldsv.end());
+        sv.tags.insert(sv.tags.end(), chldsv.tags.begin(), chldsv.tags.end());
         sv.s_ = chldsv.c_str();
         sv.n_ = chldsv.length();
         sv.tokens.insert(sv.tokens.end(), chldsv.tokens.begin(), chldsv.tokens.end());
@@ -1010,6 +1033,7 @@ public:
             auto len = rule.parse(s, n, chldsv, c, dt);
             if (success(len)) {
                 sv.insert(sv.end(), chldsv.begin(), chldsv.end());
+                sv.tags.insert(sv.tags.end(), chldsv.tags.begin(), chldsv.tags.end());
                 sv.s_ = chldsv.c_str();
                 sv.n_ = chldsv.length();
                 sv.choice_count_ = opes_.size();
@@ -1056,6 +1080,7 @@ public:
             } else {
                 if (sv.size() != save_sv_size) {
                     sv.erase(sv.begin() + static_cast<std::ptrdiff_t>(save_sv_size));
+                    sv.tags.erase(sv.tags.begin() + static_cast<std::ptrdiff_t>(save_sv_size));
                 }
                 if (sv.tokens.size() != save_tok_size) {
                     sv.tokens.erase(sv.tokens.begin() + static_cast<std::ptrdiff_t>(save_tok_size));
@@ -1114,6 +1139,7 @@ public:
             } else {
                 if (sv.size() != save_sv_size) {
                     sv.erase(sv.begin() + static_cast<std::ptrdiff_t>(save_sv_size));
+                    sv.tags.erase(sv.tags.begin() + static_cast<std::ptrdiff_t>(save_sv_size));
                 }
                 if (sv.tokens.size() != save_tok_size) {
                     sv.tokens.erase(sv.tokens.begin() + static_cast<std::ptrdiff_t>(save_tok_size));
@@ -1155,6 +1181,7 @@ public:
         } else {
             if (sv.size() != save_sv_size) {
                 sv.erase(sv.begin() + static_cast<std::ptrdiff_t>(save_sv_size));
+                sv.tags.erase(sv.tags.begin() + static_cast<std::ptrdiff_t>(save_sv_size));
             }
             if (sv.tokens.size() != save_tok_size) {
                 sv.tokens.erase(sv.tokens.begin() + static_cast<std::ptrdiff_t>(save_tok_size));
@@ -2278,6 +2305,7 @@ inline size_t Holder::parse(const char* s, size_t n, SemanticValues& sv, Context
     if (success(len)) {
         if (!outer_->ignoreSemanticValue) {
             sv.emplace_back(val);
+            sv.tags.emplace_back(str2tag(outer_->name.c_str()));
         }
     } else {
         if (outer_->error_message) {
@@ -2957,20 +2985,6 @@ private:
  *  AST
  *---------------------------------------------------------------------------*/
 
-const int AstDefaultTag = -1;
-
-#ifndef PEGLIB_NO_CONSTEXPR_SUPPORT
-inline constexpr unsigned int str2tag(const char* str, int h = 0) {
-    return !str[h] ? 5381 : (str2tag(str, h + 1) * 33) ^ static_cast<unsigned char>(str[h]);
-}
-
-namespace udl {
-inline constexpr unsigned int operator "" _(const char* s, size_t) {
-    return str2tag(s);
-}
-}
-#endif
-
 template <typename Annotation>
 struct AstBase : public Annotation
 {
diff --git a/test/test.cc b/test/test.cc
index f30323e..2ceb4a1 100644
--- a/test/test.cc
+++ b/test/test.cc
@@ -842,6 +842,51 @@ TEST_CASE("Ordered choice count 2", "[general]")
     parser.parse("b");
 }
 
+TEST_CASE("Semantic value tag", "[general]")
+{
+    parser parser(R"(
+        S <- A? B* C?
+        A <- 'a'
+        B <- 'b'
+        C <- 'c'
+    )");
+
+    {
+        using namespace udl;
+        parser["S"] = [](const SemanticValues& sv) {
+            REQUIRE(sv.size() == 1);
+            REQUIRE(sv.tags.size() == 1);
+            REQUIRE(sv.tags[0] == "C"_);
+        };
+        auto ret = parser.parse("c");
+        REQUIRE(ret == true);
+    }
+
+    {
+        using namespace udl;
+        parser["S"] = [](const SemanticValues& sv) {
+            REQUIRE(sv.size() == 2);
+            REQUIRE(sv.tags.size() == 2);
+            REQUIRE(sv.tags[0] == "B"_);
+            REQUIRE(sv.tags[1] == "B"_);
+        };
+        auto ret = parser.parse("bb");
+        REQUIRE(ret == true);
+    }
+
+    {
+        using namespace udl;
+        parser["S"] = [](const SemanticValues& sv) {
+            REQUIRE(sv.size() == 2);
+            REQUIRE(sv.tags.size() == 2);
+            REQUIRE(sv.tags[0] == "A"_);
+            REQUIRE(sv.tags[1] == "C"_);
+        };
+        auto ret = parser.parse("ac");
+        REQUIRE(ret == true);
+    }
+}
+
 TEST_CASE("Packrat parser test with %whitespace%", "[packrat]")
 {
     peg::parser parser(R"(