[TableGen] Add !interleave operator to concatenate a list of values with delimiters

[Paul-C-Anagnostopoulos]

Summary

This patch adds the !adjoin bang operator. It takes a list of string or int/bits/bit and concatenates them with a specified delimiter in between each pair. I updated the documentation to describe it.

I added a test for it.

I modified some TableGen files to make use of it.

[Paul-C-Anagnostopoulos]

Oops, Builders.cpp should not be in this patch. I will remove it.

[Paul-C-Anagnostopoulos]

I removed Builders.cpp from this revision.

[jpienaar]

Nice reduction

llvm/include/llvm/TableGen/Record.h

812

OOC what is the ordering here?

[lattner]

The word "adjoin" is a bit obscure and I don't think it quote connotes the right thing here. How about !interleave instead?

[Paul-C-Anagnostopoulos]

I wouldn't use !interleave. That sounds like it is doing something with multiple lists. Another choice was !listify, but that seems like it would construct a list. I agree, though, that !adjoin sounds more like an operation on sets.

How about !delimconcat?

llvm/include/llvm/TableGen/Record.h

812

I've pondered the ordering of various lists in TableGen and the best I can come up with is that they are grouped logically. Do you think I should alphabetize this list?

[Paul-C-Anagnostopoulos]

If I hear no objection, I will name this bang operator !delimconcat.

[lattner]

FWIW, I suggested interleave because of the llvm::interleave algorithm in llvm/ADT/STLExtras.h

[Paul-C-Anagnostopoulos]

Ah, I didn't realize there was a precedent for the name. Okay, let's call it !interleave. It reads nicely, too.

[Paul-C-Anagnostopoulos]

Change the name of the new bang operator to !interleave.

[lattner]

nice!

llvm/docs/CommandGuide/tblgen.rst

.. option:: -gen-register-info
   Generate registers and register classes info.
 
 .. option:: -register-info-debug
 
   Make -gen-register-info dump register information for debugging.
 
 .. option:: -gen-searchable-tables
 
-  Generate generic searchable tables. See :doc:`TableGen BackEnds <./BackEnds>`
+  Generate generic searchable tables. See
+  :doc:`TableGen BackEnds <../TableGen/BackEnds>`
   for a detailed description.
 
 .. option:: -gen-subtarget
 
  Generate subtarget enumerations.
 
 .. option:: -gen-x86-EVEX2VEX-tables
 

llvm/docs/TableGen/ProgRef.rst

 
 Bang operators
 --------------
 
 TableGen provides "bang operators" that have a wide variety of uses:
 
 .. productionlist::
    BangOperator: one of
-               : !add        !and         !cast        !con         !dag
-               : !empty      !eq          !foldl       !foreach     !ge
-               : !getdagop   !gt          !head        !if          !isa
-               : !le         !listconcat  !listsplat   !lt          !mul
-               : !ne         !not         !or          !setdagop    !shl
-               : !size       !sra         !srl         !strconcat   !sub
-               : !subst      !tail        !xor
+               : !add        !adjoin      !and         !cast        !con
+               : !dag        !empty       !eq          !foldl       !foreach
+               : !ge         !getdagop    !gt          !head        !if
+               : !isa        !le          !listconcat  !listsplat   !lt
+               : !mul        !ne          !not         !or          !setdagop
+               : !shl        !size        !sra         !srl         !strconcat
+               : !sub        !subst       !tail        !xor
 
 The ``!cond`` operator has a slightly different
 syntax compared to other bang operators, so it is defined separately:
 
 .. productionlist::
    CondOperator: !cond
 
 See `Appendix A: Bang Operators`_ for a description of each bang operator.

 
 .. warning::
   The ``!getop`` and ``!setop`` bang operators are deprecated in favor of
   ``!getdagop`` and ``!setdagop``.
 
 ``!add(``\ *a*\ ``,`` *b*\ ``, ...)``
     This operator adds *a*, *b*, etc., and produces the sum.
 
+``!adjoin(``\ *list*\ ``,`` *delim*\ ``)``
+    This operator concatenates the items in the *list* with the *delim*
+    string between each pair and produces the resulting string. The list
+    can be a list of string, int, bits, or bit. An empty list
+    results in an empty string. The delimiter can be the empty string.
+
 ``!and(``\ *a*\ ``,`` *b*\ ``, ...)``
     This operator does a bitwise AND on *a*, *b*, etc., and produces the
     result. A logical AND can be performed if all the arguments are either
     0 or 1.
 
 ``!cast<``\ *type*\ ``>(``\ *a*\ ``)``
     This operator performs a cast on *a* and produces the result.
     If *a* is not a string, then a straightforward cast is performed, say

llvm/docs/TableGen/index.rst

 =================
 TableGen Overview
 =================
 
 .. contents::
    :local:
 
 .. toctree::
    :hidden:
 
    BackEnds
    BackGuide
    ProgRef
-   Deficiencies
 
 Introduction
 ============
 
 TableGen's purpose is to help a human develop and maintain records of
 domain-specific information.  Because there may be a large number of these
 records, it is specifically designed to allow writing flexible descriptions and
 for common features of these records to be factored out.  This reduces the
 amount of duplication in the description, reduces the chance of error, and makes
 it easier to structure domain specific information.
 
 The TableGen front end parses a file, instantiates the declarations, and
 hands the result off to a domain-specific `backend`_ for processing.  See
 the :doc:`TableGen Programmer's Reference <./ProgRef>` for an in-depth
-description of TableGen. See :doc:`xxx-tblgen: Target Description to C++
+description of TableGen. See :doc:`xxx-tblgen - Target Description to C++
 Code <../CommandGuide/tblgen>` for details on the various
 ``xxx-tblgen`` commands that invoke TableGen.
 
 The current major users of TableGen are :doc:`The LLVM Target-Independent
 Code Generator <../CodeGenerator>` and the `Clang diagnostics and attributes
 <https://clang.llvm.org/docs/UsersManual.html#controlling-errors-and-warnings>`_.
 
 Note that if you work on TableGen much, and use emacs or vim, that you can find

llvm/include/llvm/TableGen/Record.h

public:
   /// Return true if all values of 'this' type can be converted to the specified
   /// type.
   virtual bool typeIsConvertibleTo(const RecTy *RHS) const;
 
   /// Return true if 'this' type is equal to or a subtype of RHS. For example,
   /// a bit set is not an int, but they are convertible.
   virtual bool typeIsA(const RecTy *RHS) const;
 
-  /// Returns the type representing list<this>.
+  /// Returns the type representing list<thistype>.
   ListRecTy *getListTy();
 };
 
 inline raw_ostream &operator<<(raw_ostream &OS, const RecTy &Ty) {
   Ty.print(OS);
   return OS;
 }
 
public:
 
   std::string getAsString() const override;
 };
 
 /// !op (X, Y) - Combine two inits.
 class BinOpInit : public OpInit, public FoldingSetNode {
 public:
   enum BinaryOp : uint8_t { ADD, SUB, MUL, AND, OR, XOR, SHL, SRA, SRL, LISTCONCAT,
-                            LISTSPLAT, STRCONCAT, CONCAT, EQ, NE, LE, LT, GE,
+                            LISTSPLAT, STRCONCAT, ADJOIN, CONCAT, EQ, NE, LE, LT, GE,
                             GT, SETDAGOP };
 
 private:
   Init *LHS, *RHS;
 
   BinOpInit(BinaryOp opc, Init *lhs, Init *rhs, RecTy *Type) :
       OpInit(IK_BinOpInit, Type, opc), LHS(lhs), RHS(rhs) {}
 
 public:
   BinOpInit(const BinOpInit &) = delete;
   BinOpInit &operator=(const BinOpInit &) = delete;
 
   static bool classof(const Init *I) {
     return I->getKind() == IK_BinOpInit;
   }
 
   static BinOpInit *get(BinaryOp opc, Init *lhs, Init *rhs,
                         RecTy *Type);
   static Init *getStrConcat(Init *lhs, Init *rhs);
   static Init *getListConcat(TypedInit *lhs, Init *rhs);
-  static Init *getListSplat(TypedInit *lhs, Init *rhs);
 
   void Profile(FoldingSetNodeID &ID) const;
 
   // Clone - Clone this operator, replacing arguments with the new list
   OpInit *clone(ArrayRef<Init *> Operands) const override {
     assert(Operands.size() == 2 &&
            "Wrong number of operands for binary operation");
     return BinOpInit::get(getOpcode(), Operands[0], Operands[1], getType());

llvm/lib/TableGen/Record.cpp

 
 static StringInit *ConcatStringInits(const StringInit *I0,
                                      const StringInit *I1) {
   SmallString<80> Concat(I0->getValue());
   Concat.append(I1->getValue());
   return StringInit::get(Concat);
 }
 
+static StringInit *adjoinStringList(const ListInit *List,
+                                    const StringInit *Delim) {
+  if (List->size() == 0)
+    return StringInit::get("");
+  SmallString<80> Result(dyn_cast<StringInit>(List->getElement(0))->getValue());
+  
+  for (unsigned I = 1, E = List->size(); I < E; ++I) {
+    Result.append(Delim->getValue());
+    Result.append(dyn_cast<StringInit>(List->getElement(I))->getValue());
+  }
+  return StringInit::get(Result);
+}
+
+static StringInit *adjoinIntList(const ListInit *List,
+                                 const StringInit *Delim) {
+  if (List->size() == 0)
+    return StringInit::get("");
+  SmallString<80> Result(dyn_cast<IntInit>(List->getElement(0)->
+                             getCastTo(IntRecTy::get()))->getAsString());
+  
+  for (unsigned I = 1, E = List->size(); I < E; ++I) {
+    Result.append(Delim->getValue());
+    Result.append(dyn_cast<IntInit>(List->getElement(I)->
+                      getCastTo(IntRecTy::get()))->getAsString());
+  }
+  return StringInit::get(Result);
+}
+
 Init *BinOpInit::getStrConcat(Init *I0, Init *I1) {
   // Shortcut for the common case of concatenating two strings.
   if (const StringInit *I0s = dyn_cast<StringInit>(I0))
     if (const StringInit *I1s = dyn_cast<StringInit>(I1))
       return ConcatStringInits(I0s, I1s);
   return BinOpInit::get(BinOpInit::STRCONCAT, I0, I1, StringRecTy::get());
 }
 
Init *BinOpInit::getListConcat(TypedInit *LHS, Init *RHS) {
 
   // Shortcut for the common case of concatenating two lists.
    if (const ListInit *LHSList = dyn_cast<ListInit>(LHS))
      if (const ListInit *RHSList = dyn_cast<ListInit>(RHS))
        return ConcatListInits(LHSList, RHSList);
    return BinOpInit::get(BinOpInit::LISTCONCAT, LHS, RHS, LHS->getType());
 }
 
-Init *BinOpInit::getListSplat(TypedInit *LHS, Init *RHS) {
-  return BinOpInit::get(BinOpInit::LISTSPLAT, LHS, RHS, LHS->getType());
-}
-
 Init *BinOpInit::Fold(Record *CurRec) const {
   switch (getOpcode()) {
   case CONCAT: {
     DagInit *LHSs = dyn_cast<DagInit>(LHS);
     DagInit *RHSs = dyn_cast<DagInit>(RHS);
     if (LHSs && RHSs) {
       DefInit *LOp = dyn_cast<DefInit>(LHSs->getOperator());
       DefInit *ROp = dyn_cast<DefInit>(RHSs->getOperator());
Init *BinOpInit::Fold(Record *CurRec) const {
   }
   case STRCONCAT: {
     StringInit *LHSs = dyn_cast<StringInit>(LHS);
     StringInit *RHSs = dyn_cast<StringInit>(RHS);
     if (LHSs && RHSs)
       return ConcatStringInits(LHSs, RHSs);
     break;
   }
+  case ADJOIN: {
+    ListInit *List = dyn_cast<ListInit>(LHS);
+    StringInit *Delim = dyn_cast<StringInit>(RHS);
+    if (List && Delim) {
+      if (isa<StringRecTy>(List->getElementType()))
+        return adjoinStringList(List, Delim);
+      else
+        return adjoinIntList(List, Delim);
+    }
+    break;
+  }
   case EQ:
   case NE:
   case LE:
   case LT:
   case GE:
   case GT: {
     // try to fold eq comparison for 'bit' and 'int', otherwise fallback
     // to string objects.
std::string BinOpInit::getAsString() const {
   case NE: Result = "!ne"; break;
   case LE: Result = "!le"; break;
   case LT: Result = "!lt"; break;
   case GE: Result = "!ge"; break;
   case GT: Result = "!gt"; break;
   case LISTCONCAT: Result = "!listconcat"; break;
   case LISTSPLAT: Result = "!listsplat"; break;
   case STRCONCAT: Result = "!strconcat"; break;
+  case ADJOIN: Result = "!adjoin"; break;
   case SETDAGOP: Result = "!setdagop"; break;
   }
   return Result + "(" + LHS->getAsString() + ", " + RHS->getAsString() + ")";
 }
 
 static void
 ProfileTernOpInit(FoldingSetNodeID &ID, unsigned Opcode, Init *LHS, Init *MHS,
                   Init *RHS, RecTy *Type) {

llvm/lib/TableGen/TGLexer.h

enum TokKind {
 
     // Keywords. ('ElseKW' is named to distinguish it from the existing 'Else'
     // that means the preprocessor #else.)
     Bit, Bits, Class, Code, Dag, Def, Foreach, Defm, Field, In, Int, Let, List,
     MultiClass, String, Defset, Defvar, If, Then, ElseKW,
 
     // !keywords.
     XConcat, XADD, XSUB, XMUL, XNOT, XAND, XOR, XXOR, XSRA, XSRL, XSHL,
-    XListConcat, XListSplat, XStrConcat, XCast, XSubst, XForEach, XFoldl,
-    XHead, XTail, XSize, XEmpty, XIf, XCond, XEq, XIsA, XDag, XNe, XLe,
-    XLt, XGe, XGt, XSetDagOp, XGetDagOp,
+    XListConcat, XListSplat, XStrConcat, XAdjoin, XCast, XSubst, XForEach,
+    XFoldl, XHead, XTail, XSize, XEmpty, XIf, XCond, XEq, XIsA, XDag, XNe,
+    XLe, XLt, XGe, XGt, XSetDagOp, XGetDagOp,
 
     // Integer value.
     IntVal,
 
     // Binary constant.  Note that these are sized according to the number of
     // bits given.
     BinaryIntVal,
 

llvm/lib/TableGen/TGLexer.cpp

tgtok::TokKind Kind =
     .Case("cast", tgtok::XCast)
     .Case("empty", tgtok::XEmpty)
     .Case("subst", tgtok::XSubst)
     .Case("foldl", tgtok::XFoldl)
     .Case("foreach", tgtok::XForEach)
     .Case("listconcat", tgtok::XListConcat)
     .Case("listsplat", tgtok::XListSplat)
     .Case("strconcat", tgtok::XStrConcat)
+    .Case("adjoin", tgtok::XAdjoin)
     .Cases("setdagop", "setop", tgtok::XSetDagOp) // !setop is deprecated.
     .Cases("getdagop", "getop", tgtok::XGetDagOp) // !getop is deprecated.
     .Default(tgtok::Error);
 
   return Kind != tgtok::Error ? Kind : ReturnError(Start-1, "Unknown operator");
 }
 
 bool TGLexer::prepExitInclude(bool IncludeStackMustBeEmpty) {

llvm/lib/TableGen/TGParser.cpp

 
 /// ParseOperation - Parse an operator.  This returns null on error.
 ///
 /// Operation ::= XOperator ['<' Type '>'] '(' Args ')'
 ///
 Init *TGParser::ParseOperation(Record *CurRec, RecTy *ItemType) {
   switch (Lex.getCode()) {
   default:
-    TokError("unknown operation");
+    TokError("unknown bang operator");
     return nullptr;
   case tgtok::XNOT:
   case tgtok::XHead:
   case tgtok::XTail:
   case tgtok::XSize:
   case tgtok::XEmpty:
   case tgtok::XCast:
   case tgtok::XGetDagOp: { // Value ::= !unop '(' Value ')'
Init *TGParser::ParseOperation(Record *CurRec, RecTy *ItemType) {
   case tgtok::XNe:
   case tgtok::XLe:
   case tgtok::XLt:
   case tgtok::XGe:
   case tgtok::XGt:
   case tgtok::XListConcat:
   case tgtok::XListSplat:
   case tgtok::XStrConcat:
+  case tgtok::XAdjoin:
   case tgtok::XSetDagOp: { // Value ::= !binop '(' Value ',' Value ')'
     tgtok::TokKind OpTok = Lex.getCode();
     SMLoc OpLoc = Lex.getLoc();
     Lex.Lex();  // eat the operation
 
     BinOpInit::BinaryOp Code;
     switch (OpTok) {
     default: llvm_unreachable("Unhandled code!");
case tgtok::XSetDagOp: { // Value ::= !binop '(' Value ',' Value ')'
     case tgtok::XNe:     Code = BinOpInit::NE; break;
     case tgtok::XLe:     Code = BinOpInit::LE; break;
     case tgtok::XLt:     Code = BinOpInit::LT; break;
     case tgtok::XGe:     Code = BinOpInit::GE; break;
     case tgtok::XGt:     Code = BinOpInit::GT; break;
     case tgtok::XListConcat: Code = BinOpInit::LISTCONCAT; break;
     case tgtok::XListSplat:  Code = BinOpInit::LISTSPLAT; break;
     case tgtok::XStrConcat:  Code = BinOpInit::STRCONCAT; break;
+    case tgtok::XAdjoin:     Code = BinOpInit::ADJOIN; break;
     case tgtok::XSetDagOp:   Code = BinOpInit::SETDAGOP; break;
     }
 
     RecTy *Type = nullptr;
     RecTy *ArgType = nullptr;
     switch (OpTok) {
     default:
       llvm_unreachable("Unhandled code!");
case tgtok::XListConcat:
       break;
     case tgtok::XListSplat:
       // Can't do any typechecking until we parse the first argument.
       break;
     case tgtok::XStrConcat:
       Type = StringRecTy::get();
       ArgType = StringRecTy::get();
       break;
+    case tgtok::XAdjoin:
+      Type = StringRecTy::get();
+      // The first argument type is not yet known.
     }
 
     if (Type && ItemType && !Type->typeIsConvertibleTo(ItemType)) {
       Error(OpLoc, Twine("expected value of type '") +
                    ItemType->getAsString() + "', got '" +
                    Type->getAsString() + "'");
       return nullptr;
     }
 
     if (!consume(tgtok::l_paren)) {
       TokError("expected '(' after binary operator");
       return nullptr;
     }
 
     SmallVector<Init*, 2> InitList;
 
+    // Note that this loop consumes an arbitrary number of arguments.
+    // The actual count is checked later.
     for (;;) {
       SMLoc InitLoc = Lex.getLoc();
       InitList.push_back(ParseValue(CurRec, ArgType));
       if (!InitList.back()) return nullptr;
 
       TypedInit *InitListBack = dyn_cast<TypedInit>(InitList.back());
       if (!InitListBack) {
         Error(OpLoc, Twine("expected value to be a typed value, got '" +
                            InitList.back()->getAsString() + "'"));
         return nullptr;
       }
       RecTy *ListType = InitListBack->getType();
+
       if (!ArgType) {
+        // Argument type must be determined from the argument itself.
         ArgType = ListType;
 
         switch (Code) {
         case BinOpInit::LISTCONCAT:
           if (!isa<ListRecTy>(ArgType)) {
             Error(InitLoc, Twine("expected a list, got value of type '") +
                            ArgType->getAsString() + "'");
             return nullptr;
for (;;) {
         case BinOpInit::NE:
           if (!ArgType->typeIsConvertibleTo(IntRecTy::get()) &&
               !ArgType->typeIsConvertibleTo(StringRecTy::get())) {
             Error(InitLoc, Twine("expected int, bits, or string; got value of "
                                  "type '") + ArgType->getAsString() + "'");
             return nullptr;
           }
           break;
+        case BinOpInit::ADJOIN:
+          switch (InitList.size()) {
+          case 1: // First argument must be a list of strings or integers.
+            if (ArgType != StringRecTy::get()->getListTy() &&
+                !ArgType->typeIsConvertibleTo(IntRecTy::get()->getListTy())) {
+              Error(InitLoc, Twine("expected list of string, int, bits, or bit; "
+                                   "got value of type '") +
+                                   ArgType->getAsString() + "'");
+              return nullptr;
+            }
+            break;
+          case 2: // Second argument must be a string.
+            if (!isa<StringRecTy>(ArgType)) {
+              Error(InitLoc, Twine("expected second argument to be a string, "
+                                   "got value of type '") +
+                                 ArgType->getAsString() + "'");
+              return nullptr;
+            }
+            break;
+          default: ;
+          }
+          ArgType = nullptr; // Broken invariant: types not identical.
+          break;
         default: llvm_unreachable("other ops have fixed argument types");
         }
+
       } else {
+        // Desired argument type is a known and in ArgType.
         RecTy *Resolved = resolveTypes(ArgType, ListType);
         if (!Resolved) {
           Error(InitLoc, Twine("expected value of type '") +
                              ArgType->getAsString() + "', got '" +
                              ListType->getAsString() + "'");
           return nullptr;
         }
         if (Code != BinOpInit::ADD && Code != BinOpInit::SUB &&
Init *TGParser::ParseSimpleValue(Record *CurRec, RecTy *ItemType,
   case tgtok::XNe:
   case tgtok::XLe:
   case tgtok::XLt:
   case tgtok::XGe:
   case tgtok::XGt:
   case tgtok::XListConcat:
   case tgtok::XListSplat:
   case tgtok::XStrConcat:
+  case tgtok::XAdjoin:
   case tgtok::XSetDagOp: // Value ::= !binop '(' Value ',' Value ')'
   case tgtok::XIf:
   case tgtok::XCond:
   case tgtok::XFoldl:
   case tgtok::XForEach:
   case tgtok::XSubst: { // Value ::= !ternop '(' Value ',' Value ',' Value ')'
     return ParseOperation(CurRec, ItemType);
   }

llvm/lib/Target/AMDGPU/MIMGInstructions.td

 }
 
 // This is a separate class so that TableGen memoizes the computations.
 class MIMGNSAHelper<int num_addrs> {
   list<string> AddrAsmNames =
     !foldl([]<string>, [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11], lhs, i,
            !if(!lt(i, num_addrs), !listconcat(lhs, ["vaddr"#!size(lhs)]), lhs));
   dag AddrIns = !dag(ins, !foreach(arg, AddrAsmNames, VGPR_32), AddrAsmNames);
-  string AddrAsm = "[" # !foldl("$" # !head(AddrAsmNames), !tail(AddrAsmNames), lhs, rhs,
-                                lhs # ", $" # rhs) # "]";
+  string AddrAsm = "[$" # !adjoin(AddrAsmNames, ", $") # "]";
 
   int NSA = !if(!le(num_addrs, 1), ?,
             !if(!le(num_addrs, 5), 1,
             !if(!le(num_addrs, 9), 2,
             !if(!le(num_addrs, 13), 3, ?))));
 }
 
 // Base class of all pre-gfx10 MIMG instructions.

llvm/lib/Target/NVPTX/NVPTXIntrinsics.td

class WMMA_REGINFO<WMMA_REGS r>
 
   // Instruction input/output arguments for the fragment.
   list<NVPTXRegClass> ptx_regs = !foreach(tmp, regs, regclass);
 
   // List of register names for the fragment -- ["ra0", "ra1",...]
   list<string> reg_names = RegSeq<!size(ptx_regs), "r"#frag>.ret;
 
   // Generates "{{$r0, $r1,.... $rN-1}}" for use in asm string construction.
-  string regstring = "{{$" # !head(reg_names)
-                           # !foldl("", !tail(reg_names), a, b,
-                                    !strconcat(a, ", $", b))
-                     # "}}";
+  string regstring = "{{$" # !adjoin(reg_names, ", $") # "}}";
 
   // Predicates for particular fragment variant. Technically those are
   // per-instruction predicates, but currently all fragments that can be used in
   // a given instruction are subject to the same constraints, so an instruction
   // can use predicates from any of its fragments. If/when this is no
   // longer the case, we can concat all per-fragment predicates to enforce that
   // all fragments of the instruction are viable.
   list<Predicate> Predicates = !cond(

llvm/test/TableGen/adjoin.td

+// RUN: llvm-tblgen %s | FileCheck %s
+// RUN: not llvm-tblgen -DERROR1 %s 2>&1 | FileCheck --check-prefix=ERROR1 %s
+
+defvar EmptyList = []<string>;
+defvar OneList = ["hello"];
+defvar StringList = ["foo", "bar", "zoo", "snork", "quux"];
+defvar IntList = [0, 1, 2, 3, 4, 5, 6, 7];
+defvar BitsList = [ {0, 1, 0}, {1, 1, 1}, {0, 0, 1} ];
+defvar BitList = [0, 1, 1, 0, 1]<bit>;
+
+class Ishify<list<string> words> {
+  list<string> ret = !foreach(w, words, w # "ify");
+}
+
+// CHECK: def Rec1
+// CHECK:   Test1 = "";
+// CHECK:   Test2 = "hello";
+// CHECK:   Test3 = "foobarzoosnorkquux";
+// CHECK:   Test4 = "foo, bar, zoo, snork, quux";
+// CHECK:   Test5 = "foo & bar & zoo & snork & quux & grits";
+
+def Rec1 {
+  string Test1 = !adjoin(EmptyList, "/");
+  string Test2 = !adjoin(OneList, ":");
+  string Test3 = !adjoin(StringList, "");
+  string Test4 = !adjoin(StringList, ", ");
+  string Test5 = !adjoin(!listconcat(StringList, ["grits"]), " & ");
+}
+
+// CHECK: def Rec2
+// CHECK: Test1 = "01234567";
+// CHECK: Test2 = "0, 1, 2, 3, 4, 5, 6, 7";
+// CHECK: Test3 = "0 & 1 & 2 & 3 & 4 & 5 & 6 & 7 & 42";
+
+def Rec2 {
+  string Test1 = !adjoin(IntList, "");
+  string Test2 = !adjoin(IntList, ", ");
+  string Test3 = !adjoin(!listconcat(IntList, [42]), " & ");
+}
+
+// CHECK: def Rec3
+// CHECK: Test1 = "271";
+// CHECK: Test2 = "2, 7, 1";
+// CHECK: Test3 = "2 & 7 & 1 & 0";
+
+def Rec3 {
+  string Test1 = !adjoin(BitsList, "");
+  string Test2 = !adjoin(BitsList, ", ");
+  string Test3 = !adjoin(!listconcat(BitsList, [ {0, 0, 0} ]), " & ");
+}
+
+// CHECK: def Rec4
+// CHECK: Test1 = "01101";
+// CHECK: Test2 = "0, 1, 1, 0, 1";
+// CHECK: Test3 = "0 and 1 and 1 and 0 and 1 and 1";
+
+def Rec4 {
+  string Test1 = !adjoin(BitList, "");
+  string Test2 = !adjoin(BitList, ", ");
+  string Test3 = !adjoin(!listconcat(BitList, [1]), " and ");
+}
+
+// CHECK: def Rec5
+// CHECK: Colors = ["red", "green", "yellow"];
+// CHECK: ColorList = "redify, greenify, yellowify";
+
+def Rec5 {
+  list<string> Colors = ["red", "green", "yellow"];
+  string ColorList = !adjoin(Ishify<Colors>.ret, ", ");
+}
+
+#ifdef ERROR1
+def op;
+
+// ERROR1: expected list of string, int, bits, or bit; got value of type
+
+def Rec6 {
+  string Bad = !adjoin([(op), (op "hello")], " = ");
+}
+#endif

mlir/include/mlir/IR/OpBase.td

 //
 // StrConcat<["a", "b", "c"]>.result
 //
 // to get the string "abc"
 class StrFunc<string r> {
   string result = r;
 }
 
+// TODO: Use !adjoin() directly rather than through StrJoin/StrJoinInt.
+
 // Concatenates a list of strings with a separator (default ", ")
 class StrJoin<list<string> strings, string sep = ", "> :
-    StrFunc<!if(!empty(strings), "",
-         !foldl(!head(strings), !tail(strings), prev, cur, prev # sep # cur))>;
+    StrFunc<!adjoin(strings, sep)>;
 
 // Concatenates a list of integers into a string with a separator (default ", ")
 class StrJoinInt<list<int> integers, string sep = ", "> :
-    StrJoin<!foreach(i, integers, !cast<string>(i)), sep>;
+    StrFunc<!adjoin(integers, sep)>;
 
 //===----------------------------------------------------------------------===//
 // Predicate definitions
 //===----------------------------------------------------------------------===//
 
 // Base class for logical predicates.
 //
 // Predicates are used to compose constraints (see next section for details).
class Confined<Attr attr, list<AttrConstraint> constraints> : Attr<
   let baseAttr = attr;
 }
 
 // An AttrConstraint that holds if all attr constraints specified in
 // 'constraints' hold.
 class AllAttrConstraintsOf<list<AttrConstraint> constraints> : AttrConstraint<
     And<!listconcat([!head(constraints).predicate],
                       !foreach(pred, !tail(constraints), pred.predicate))>,
-    !foldl(/*init*/!head(constraints).description, /*list*/!tail(constraints),
-           prev, cur, prev # " and " # cur.description)> {
+    !adjoin(!foreach(con, constraints, con.description), " and ")> {
 }
 
 class IntMinValue<int n> : AttrConstraint<
     CPred<"$_self.cast<::mlir::IntegerAttr>().getInt() >= " # n>,
     "whose minimum value is " # n>;
 
 class IntMaxValue<int n> : AttrConstraint<
     CPred<"$_self.cast<::mlir::IntegerAttr>().getInt() <= " # n>,