From 7e4a10160e250f5880016c92f35edbef08a78ed6 Mon Sep 17 00:00:00 2001
From: Raimund Merkert <raimund.merkert@str.us>
Date: Wed, 28 Sep 2022 18:29:34 -0400
Subject: [PATCH] Use regex cache in the synthesized C++ code

Update the synthesizer to use a cache for regexes.
One drawback is that we now need to #include <regex>,
but the structure of the synthesizer makes impossible
to generate the cache only when needed.
---
 src/include/souffle/CompiledSouffle.h  |  1 +
 src/include/souffle/SouffleInterface.h |  2 ++
 src/synthesiser/Synthesiser.cpp        | 10 ++++++++--
 3 files changed, 11 insertions(+), 2 deletions(-)

diff --git a/src/include/souffle/CompiledSouffle.h b/src/include/souffle/CompiledSouffle.h
index e13579b81e0..9183e9802e1 100644
--- a/src/include/souffle/CompiledSouffle.h
+++ b/src/include/souffle/CompiledSouffle.h
@@ -23,6 +23,7 @@
 #include "souffle/SymbolTable.h"
 #include "souffle/datastructure/BTreeDelete.h"
 #include "souffle/datastructure/Brie.h"
+#include "souffle/datastructure/ConcurrentCache.h"
 #include "souffle/datastructure/EqRel.h"
 #include "souffle/datastructure/Info.h"
 #include "souffle/datastructure/Nullaries.h"
diff --git a/src/include/souffle/SouffleInterface.h b/src/include/souffle/SouffleInterface.h
index 2087b38456f..be9a3f582a7 100644
--- a/src/include/souffle/SouffleInterface.h
+++ b/src/include/souffle/SouffleInterface.h
@@ -19,6 +19,7 @@
 #include "souffle/RamTypes.h"
 #include "souffle/RecordTable.h"
 #include "souffle/SymbolTable.h"
+#include "souffle/datastructure/ConcurrentCache.h"
 #include "souffle/utility/MiscUtil.h"
 #include <algorithm>
 #include <cassert>
@@ -29,6 +30,7 @@
 #include <map>
 #include <memory>
 #include <optional>
+#include <regex>
 #include <string>
 #include <tuple>
 #include <utility>
diff --git a/src/synthesiser/Synthesiser.cpp b/src/synthesiser/Synthesiser.cpp
index 6865fd5ddd5..07379fd02a3 100644
--- a/src/synthesiser/Synthesiser.cpp
+++ b/src/synthesiser/Synthesiser.cpp
@@ -2607,6 +2607,7 @@ void Synthesiser::generateCode(GenDb& db, const std::string& id, bool& withShare
         std::vector<std::tuple<Mode, std::string /*name*/, std::string /*type*/>> args;
         args.push_back(std::make_tuple(Reference, "symTable", "SymbolTable"));
         args.push_back(std::make_tuple(Reference, "recordTable", "RecordTable"));
+        args.push_back(std::make_tuple(Reference, "regexCache", "ConcurrentCache<std::string,std::regex>"));
         args.push_back(std::make_tuple(Reference, "pruneImdtRels", "bool"));
         args.push_back(std::make_tuple(Reference, "performIO", "bool"));
         args.push_back(std::make_tuple(Reference, "signalHandler", "SignalHandler*"));
@@ -2665,13 +2666,14 @@ void Synthesiser::generateCode(GenDb& db, const std::string& id, bool& withShare
         if (SubroutineUsingStdRegex) {
             // regex wrapper
             GenFunction& wrapper = gen.addFunction("regex_wrapper", Visibility::Private);
-            gen.addInclude("<regex>");
             wrapper.setRetType("inline bool");
             wrapper.setNextArg("const std::string&", "pattern");
             wrapper.setNextArg("const std::string&", "text");
             wrapper.body()
                     << "   bool result = false; \n"
-                    << "   try { result = std::regex_match(text, std::regex(pattern)); } catch(...) { \n"
+                    << "   try { result = std::regex_match(text, regexCache.getOrCreate(pattern)); } "
+                       "catch(...) { "
+                       "\n"
                     << "     std::cerr << \"warning: wrong pattern provided for match(\\\"\" << pattern << "
                        "\"\\\",\\\"\" "
                        "<< text << \"\\\").\\n\";\n}\n"
@@ -2728,6 +2730,9 @@ void Synthesiser::generateCode(GenDb& db, const std::string& id, bool& withShare
     mainClass.addField(rt.str(), "recordTable", Visibility::Private);
     constructor.setNextInitializer("recordTable", "");
 
+    mainClass.addField("ConcurrentCache<std::string,std::regex>", "regexCache", Visibility::Private);
+    constructor.setNextInitializer("regexCache", "");
+
     if (Global::config().has("profile")) {
         std::size_t numFreq = 0;
         visit(prog, [&](const Statement&) { numFreq++; });
@@ -3024,6 +3029,7 @@ void Synthesiser::generateCode(GenDb& db, const std::string& id, bool& withShare
     setNumThreads.body() << "SouffleProgram::setNumThreads(numThreadsValue);\n";
     setNumThreads.body() << "symTable.setNumLanes(getNumThreads());\n";
     setNumThreads.body() << "recordTable.setNumLanes(getNumThreads());\n";
+    setNumThreads.body() << "regexCache.setNumLanes(getNumThreads());\n";
 
     if (!prog.getSubroutines().empty()) {
         // generate subroutine adapter