symbol concatenation with binary + (#2431)

* dos2unix populate-deps.yml

* Introduce binary `+` symbol concatenation

.github/workflows/populate-deps.yml

@@ -1,40 +1,40 @@
-name: Populate dependencies
-
-on:
-    push:
-        branches: [ master ]
-    workflow_dispatch:
-
-permissions:
-    contents: write
-
-env:
-    GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-    VCPKG_FEATURE_FLAGS: dependencygraph
-
-jobs:
-  build:
-
-    runs-on: ubuntu-latest
-
-    steps:
-    - uses: actions/checkout@v3
-      with:
-        submodules: 'recursive'
-
-    - name: Checkout vcpkg
-      run: |
-            git clone https://github.com/Microsoft/vcpkg.git
-            cd vcpkg
-            git checkout 56954f1db97f38635782d5ad7cdfd45d2731c854
-
-    - name: Bootstrap vcpkg
-      working-directory: vcpkg
-      run: ./bootstrap-vcpkg.sh
-
-    # This will execute a dry-run, meaning that libraries will not be built and
-    # installed, but they will still be reported to the GitHub dependency graph.
-    # This step assumes `vcpkg` has been bootstrapped (run `./vcpkg/bootstrap-vcpkg`)
-    - name: Run vcpkg
-      run: ${{ github.workspace }}/vcpkg/vcpkg install --dry-run --debug
-
+name: Populate dependencies
+
+on:
+    push:
+        branches: [ master ]
+    workflow_dispatch:
+
+permissions:
+    contents: write
+
+env:
+    GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+    VCPKG_FEATURE_FLAGS: dependencygraph
+
+jobs:
+  build:
+
+    runs-on: ubuntu-latest
+
+    steps:
+    - uses: actions/checkout@v3
+      with:
+        submodules: 'recursive'
+
+    - name: Checkout vcpkg
+      run: |
+            git clone https://github.com/Microsoft/vcpkg.git
+            cd vcpkg
+            git checkout 56954f1db97f38635782d5ad7cdfd45d2731c854
+
+    - name: Bootstrap vcpkg
+      working-directory: vcpkg
+      run: ./bootstrap-vcpkg.sh
+
+    # This will execute a dry-run, meaning that libraries will not be built and
+    # installed, but they will still be reported to the GitHub dependency graph.
+    # This step assumes `vcpkg` has been bootstrapped (run `./vcpkg/bootstrap-vcpkg`)
+    - name: Run vcpkg
+      run: ${{ github.workspace }}/vcpkg/vcpkg install --dry-run --debug
+

src/FunctorOps.cpp

@@ -48,6 +48,7 @@ char const* functorOpNameLegacy(FunctorOp op) {
         /** Binary Functor Operators */
         case FunctorOp::ADD:
         case FunctorOp::FADD:
+        case FunctorOp::SSADD:
         case FunctorOp::UADD: return "+";
         case FunctorOp::SUB:
         case FunctorOp::USUB:
@@ -223,6 +224,9 @@ const std::vector<IntrinsicFunctorInfo> FUNCTOR_INTRINSICS = {
         VARIADIC(CAT, Symbol),
         OP_1(STRLEN, Symbol, Signed),
         OP_3(SUBSTR, Symbol, Signed, Signed, Symbol, false),
+
+        {functorOpNameSymbol(FOp::SSADD), {TAttr::Symbol, TAttr::Symbol}, TAttr::Symbol, FOp::SSADD, false,
+                false},
 };
 
 template <typename F>

src/FunctorOps.h

@@ -101,6 +101,7 @@ enum class FunctorOp {
     FMIN,                // min of two floats
     SMAX,                // max of two symbols
     SMIN,                // min of two symbols
+    SSADD,               // string-string concatenation
 
     // Produces values within a numeric range. Format is `range(bgn, endExcl, step = 1)`.
     // e.g. `range(0, 5)` produces the sequence `0, 1, 2, 3, 4`.

src/interpreter/Engine.cpp

@@ -784,6 +784,14 @@ RamDomain Engine::execute(const Node* node, Context& ctxt) {
                 case FunctorOp::URANGE:
                 case FunctorOp::FRANGE:
                     fatal("ICE: functor `%s` must map onto `NestedIntrinsicOperator`", cur.getOperator());
+
+                case FunctorOp::SSADD: {
+                    auto sleft = execute(shadow.getChild(0), ctxt);
+                    auto sright = execute(shadow.getChild(1), ctxt);
+                    const std::string& strleft = getSymbolTable().decode(sleft);
+                    const std::string& strright = getSymbolTable().decode(sright);
+                    return getSymbolTable().encode(strleft + strright);
+                }
             }
 
         {UNREACHABLE_BAD_CASE_ANALYSIS}

src/synthesiser/Synthesiser.cpp

@@ -1864,7 +1864,7 @@ void Synthesiser::emitCode(std::ostream& out, const Statement& stmt) {
 
                 // strings
                 case BinaryConstraintOp::MATCH: {
-                    if (const StringConstant* str = dynamic_cast<const StringConstant*>(&rel.getLHS()); str) {
+                    if (const StringConstant* str = as<StringConstant>(&rel.getLHS()); str) {
                         const auto& regex = synthesiser.compileRegex(str->getConstant());
                         if (regex) {
                             out << "std::regex_match(symTable.decode(";
@@ -1884,7 +1884,7 @@ void Synthesiser::emitCode(std::ostream& out, const Statement& stmt) {
                     break;
                 }
                 case BinaryConstraintOp::NOT_MATCH: {
-                    if (const StringConstant* str = dynamic_cast<const StringConstant*>(&rel.getLHS()); str) {
+                    if (const StringConstant* str = as<StringConstant>(&rel.getLHS()); str) {
                         const auto& regex = synthesiser.compileRegex(str->getConstant());
                         if (regex) {
                             out << "!std::regex_match(symTable.decode(";
@@ -2313,6 +2313,35 @@ void Synthesiser::emitCode(std::ostream& out, const Statement& stmt) {
                 case FunctorOp::URANGE:
                 case FunctorOp::FRANGE:
                     fatal("ICE: functor `%s` must map onto `NestedIntrinsicOperator`", op.getOperator());
+
+                case FunctorOp::SSADD: {
+                    const StringConstant* lstr = as<StringConstant>(args[0]);
+                    const StringConstant* rstr = as<StringConstant>(args[1]);
+                    if (lstr && rstr) {
+                        out << "RamSigned("
+                            << synthesiser.convertSymbol2Idx(lstr->getConstant() + rstr->getConstant())
+                            << ")";
+                    } else {
+                        out << "symTable.encode(";
+                        if (lstr) {
+                            out << "R\"_(" << lstr->getConstant() << ")_\"";
+                        } else {
+                            out << "symTable.decode(";
+                            dispatch(*args[0], out);
+                            out << ")";
+                        }
+                        out << " + ";
+                        if (rstr) {
+                            out << "R\"_(" << rstr->getConstant() << ")_\"";
+                        } else {
+                            out << "symTable.decode(";
+                            dispatch(*args[1], out);
+                            out << ")";
+                        }
+                        out << ")";
+                    }
+                    break;
+                }
             }
             PRINT_END_COMMENT(out);
 

tests/evaluation/CMakeLists.txt

@@ -161,6 +161,7 @@ positive_test(subtype2)
 positive_test(subtype)
 positive_test(sum-aggregate)
 positive_test(sum-aggregate2)
+positive_test(symbol_operations)
 positive_test(term)
 positive_test(unpacking)
 positive_test(unsigned_operations)

tests/evaluation/symbol_operations/Add.csv

@@ -0,0 +1,9 @@
+pre--post
+pre-(a)
+pre-(b)
+(a)-post
+(b)-post
+(b)(a)(a)
+(a)(b)(b)
+one:1, pi:3.140000
+

tests/evaluation/symbol_operations/symbol_operations.dl

@@ -0,0 +1,23 @@
+// Souffle - A Datalog Compiler
+// Copyright (c) 2023, The Souffle Developers. All rights reserved
+// Licensed under the Universal Permissive License v 1.0 as shown at:
+// - https://opensource.org/licenses/UPL
+// - <souffle root>/licenses/SOUFFLE-UPL.txt
+
+.decl Add(s:symbol)
+Add("pre-" + "-post").
+Add("pre-" + s) :- some_symbol(s).
+Add(s + "-post") :- some_symbol(s).
+Add(s1 + s2 + s3) :-
+  some_symbol(s1),
+  some_symbol(s2),
+  some_symbol(s3),
+  s1 != s2,
+  s1 != s3.
+Add("one:" + to_string(1) + ", pi:" + to_string(3.14) + "\n").
+
+.decl some_symbol(s:symbol)
+some_symbol("(a)").
+some_symbol("(b)").
+
+.output Add()