From 741d78b6e840c88b7b183010c812ad982b342da1 Mon Sep 17 00:00:00 2001 From: Cliff Click Date: Mon, 10 Aug 2026 11:11:18 -0700 Subject: Chapter 28a: Modules and Lazy Class Initialization Co-authored-by: Hsiang-Ying Fu <66062782+fuhsnn@users.noreply.github.com> Co-authored-by: fuhsnn <66062782+fuhsnn@users.noreply.github.com> Co-authored-by: Symious <14933944+symious@users.noreply.github.com> Co-authored-by: Jun Kawasaki Co-authored-by: Claude Opus 5 --- src/main/c/simple_crt.c | 61 + src/main/java/com/seaofnodes/simple/IterPeeps.java | 199 +- src/main/java/com/seaofnodes/simple/Parser.java | 2612 ++++++++++++-------- src/main/java/com/seaofnodes/simple/Var.java | 18 +- .../com/seaofnodes/simple/codegen/BuildLRG.java | 20 +- .../simple/codegen/ClassInitDependencies.java | 214 ++ .../com/seaofnodes/simple/codegen/CodeGen.java | 597 +++-- .../com/seaofnodes/simple/codegen/CompUnit.java | 98 + .../com/seaofnodes/simple/codegen/ElfExpr.java | 87 - .../com/seaofnodes/simple/codegen/ElfFile.java | 382 --- .../com/seaofnodes/simple/codegen/ElfWriter.java | 460 ++++ .../com/seaofnodes/simple/codegen/Encoding.java | 371 +-- .../com/seaofnodes/simple/codegen/FinalizeIR.java | 90 +- .../com/seaofnodes/simple/codegen/FinishCalls.java | 57 + .../simple/codegen/GlobalCodeMotion.java | 71 +- .../java/com/seaofnodes/simple/codegen/IFG.java | 91 +- .../com/seaofnodes/simple/codegen/InstSelect.java | 80 +- .../java/com/seaofnodes/simple/codegen/Iter.java | 38 + .../java/com/seaofnodes/simple/codegen/LRG.java | 7 +- .../com/seaofnodes/simple/codegen/LinkMem.java | 22 +- .../seaofnodes/simple/codegen/ListScheduler.java | 2 +- .../com/seaofnodes/simple/codegen/LocalIds.java | 33 + .../com/seaofnodes/simple/codegen/Machine.java | 26 +- .../java/com/seaofnodes/simple/codegen/Opto.java | 218 +- .../com/seaofnodes/simple/codegen/ParseAll.java | 285 +++ .../com/seaofnodes/simple/codegen/RIPRelSize.java | 2 +- .../com/seaofnodes/simple/codegen/RegAlloc.java | 150 +- .../com/seaofnodes/simple/codegen/RegMask.java | 1 - .../com/seaofnodes/simple/codegen/SourceError.java | 49 + .../com/seaofnodes/simple/codegen/StaticData.java | 177 ++ .../com/seaofnodes/simple/codegen/TypeCheck.java | 61 +- .../java/com/seaofnodes/simple/node/AddFNode.java | 32 - .../java/com/seaofnodes/simple/node/AddNode.java | 3 +- .../java/com/seaofnodes/simple/node/AndNode.java | 1 + .../com/seaofnodes/simple/node/ArithFNode.java | 31 - .../java/com/seaofnodes/simple/node/ArithNode.java | 2 + .../java/com/seaofnodes/simple/node/BoolNode.java | 36 +- .../com/seaofnodes/simple/node/BulkMemPhiNode.java | 76 +- .../java/com/seaofnodes/simple/node/CFGNode.java | 143 +- .../com/seaofnodes/simple/node/CLInitNode.java | 165 ++ .../java/com/seaofnodes/simple/node/CProjNode.java | 18 +- .../com/seaofnodes/simple/node/CallEndMach.java | 6 +- .../com/seaofnodes/simple/node/CallEndNode.java | 421 ++-- .../java/com/seaofnodes/simple/node/CallNode.java | 100 +- .../java/com/seaofnodes/simple/node/CastMach.java | 19 - .../java/com/seaofnodes/simple/node/CastNode.java | 79 - .../com/seaofnodes/simple/node/ConFldOffNode.java | 51 +- .../com/seaofnodes/simple/node/ConstantNode.java | 57 +- .../com/seaofnodes/simple/node/ConvertNode.java | 16 +- .../java/com/seaofnodes/simple/node/CtrlNode.java | 1 + .../java/com/seaofnodes/simple/node/DivFNode.java | 27 - .../java/com/seaofnodes/simple/node/DivNode.java | 1 + .../com/seaofnodes/simple/node/EscapeMach.java | 12 - .../com/seaofnodes/simple/node/EscapeNode.java | 131 +- .../com/seaofnodes/simple/node/ExternNode.java | 20 +- .../seaofnodes/simple/node/ExternOffsetNode.java | 16 + .../java/com/seaofnodes/simple/node/FRefNode.java | 27 +- .../com/seaofnodes/simple/node/FldOffNode.java | 56 +- .../java/com/seaofnodes/simple/node/FunNode.java | 367 ++- .../com/seaofnodes/simple/node/FunPtrNode.java | 62 +- .../java/com/seaofnodes/simple/node/GuardNode.java | 4 - .../java/com/seaofnodes/simple/node/IfNode.java | 15 +- .../java/com/seaofnodes/simple/node/LoadNode.java | 243 +- .../java/com/seaofnodes/simple/node/LoopNode.java | 36 +- .../seaofnodes/simple/node/MachConcreteNode.java | 5 +- .../java/com/seaofnodes/simple/node/MachNode.java | 2 +- .../com/seaofnodes/simple/node/MemMergeNode.java | 199 +- .../java/com/seaofnodes/simple/node/MemOpNode.java | 124 +- .../com/seaofnodes/simple/node/MemPhiNode.java | 53 +- .../com/seaofnodes/simple/node/MinusFNode.java | 32 - .../java/com/seaofnodes/simple/node/MulFNode.java | 37 - .../java/com/seaofnodes/simple/node/MulNode.java | 3 +- .../java/com/seaofnodes/simple/node/MultiNode.java | 2 +- .../java/com/seaofnodes/simple/node/NeverNode.java | 3 +- .../java/com/seaofnodes/simple/node/NewNode.java | 108 +- src/main/java/com/seaofnodes/simple/node/Node.java | 129 +- .../java/com/seaofnodes/simple/node/NotNode.java | 1 - .../java/com/seaofnodes/simple/node/OrNode.java | 1 + .../java/com/seaofnodes/simple/node/ParmNode.java | 80 +- .../java/com/seaofnodes/simple/node/PhiNode.java | 149 +- src/main/java/com/seaofnodes/simple/node/Proj.java | 2 + .../java/com/seaofnodes/simple/node/ProjNode.java | 11 +- .../com/seaofnodes/simple/node/PtrToIntNode.java | 1 + .../com/seaofnodes/simple/node/ReadOnlyNode.java | 2 +- .../com/seaofnodes/simple/node/RegionNode.java | 16 +- .../com/seaofnodes/simple/node/ReturnNode.java | 139 +- .../com/seaofnodes/simple/node/RoundF32Node.java | 1 - .../java/com/seaofnodes/simple/node/SarNode.java | 1 + .../java/com/seaofnodes/simple/node/ScopeNode.java | 481 ++-- .../java/com/seaofnodes/simple/node/ShlNode.java | 3 +- .../java/com/seaofnodes/simple/node/ShrNode.java | 1 + .../java/com/seaofnodes/simple/node/SplitNode.java | 8 +- .../com/seaofnodes/simple/node/StartCUNode.java | 22 + .../java/com/seaofnodes/simple/node/StartNode.java | 82 +- .../com/seaofnodes/simple/node/StopCUNode.java | 126 + .../java/com/seaofnodes/simple/node/StopNode.java | 90 +- .../java/com/seaofnodes/simple/node/StoreNode.java | 219 +- .../com/seaofnodes/simple/node/StructNode.java | 56 - .../java/com/seaofnodes/simple/node/SubFNode.java | 28 - .../java/com/seaofnodes/simple/node/SubNode.java | 1 + .../com/seaofnodes/simple/node/ToFloatNode.java | 1 - .../com/seaofnodes/simple/node/ToIntegerNode.java | 1 + .../java/com/seaofnodes/simple/node/TypeNode.java | 32 + .../java/com/seaofnodes/simple/node/XCtrlNode.java | 2 +- .../java/com/seaofnodes/simple/node/XorNode.java | 1 + .../seaofnodes/simple/node/cpus/arm/BranchARM.java | 8 +- .../seaofnodes/simple/node/cpus/arm/CallARM.java | 3 +- .../seaofnodes/simple/node/cpus/arm/FloatARM.java | 44 - .../seaofnodes/simple/node/cpus/arm/FltARM.java | 45 + .../seaofnodes/simple/node/cpus/arm/IntARM.java | 7 +- .../seaofnodes/simple/node/cpus/arm/LoadARM.java | 2 +- .../seaofnodes/simple/node/cpus/arm/MemOpARM.java | 12 +- .../seaofnodes/simple/node/cpus/arm/NewARM.java | 3 +- .../seaofnodes/simple/node/cpus/arm/SetARM.java | 6 +- .../seaofnodes/simple/node/cpus/arm/SplitARM.java | 12 +- .../simple/node/cpus/arm/SplitFlagsARM.java | 2 +- .../simple/node/cpus/arm/SplitGPRARM.java | 2 +- .../seaofnodes/simple/node/cpus/arm/StoreARM.java | 2 +- .../seaofnodes/simple/node/cpus/arm/TFPARM.java | 37 +- .../seaofnodes/simple/node/cpus/arm/TMPARM.java | 18 +- .../com/seaofnodes/simple/node/cpus/arm/arm.java | 116 +- .../seaofnodes/simple/node/cpus/riscv/AUIPC.java | 11 +- .../simple/node/cpus/riscv/CallRISC.java | 3 +- .../simple/node/cpus/riscv/CallRRISC.java | 2 +- .../seaofnodes/simple/node/cpus/riscv/FltRISC.java | 5 +- .../simple/node/cpus/riscv/Int8RISC.java | 8 +- .../seaofnodes/simple/node/cpus/riscv/IntRISC.java | 2 +- .../simple/node/cpus/riscv/LoadRISC.java | 35 + .../simple/node/cpus/riscv/MemOpRISC.java | 41 +- .../seaofnodes/simple/node/cpus/riscv/NewRISC.java | 5 +- .../simple/node/cpus/riscv/SplitRISC.java | 12 +- .../simple/node/cpus/riscv/StoreRISC.java | 8 +- .../seaofnodes/simple/node/cpus/riscv/TFPRISC.java | 31 +- .../seaofnodes/simple/node/cpus/riscv/TMPRISC.java | 12 +- .../seaofnodes/simple/node/cpus/riscv/riscv.java | 119 +- .../simple/node/cpus/x86_64_v2/AddFMemX86.java | 2 +- .../simple/node/cpus/x86_64_v2/CallX86.java | 12 +- .../simple/node/cpus/x86_64_v2/CmpMemX86.java | 4 +- .../simple/node/cpus/x86_64_v2/FltX86.java | 7 +- .../simple/node/cpus/x86_64_v2/FunX86.java | 1 + .../simple/node/cpus/x86_64_v2/IntX86.java | 7 +- .../simple/node/cpus/x86_64_v2/LoadX86.java | 4 +- .../simple/node/cpus/x86_64_v2/MemOpX86.java | 8 +- .../simple/node/cpus/x86_64_v2/NewX86.java | 3 +- .../simple/node/cpus/x86_64_v2/SplitFlagsX86.java | 4 +- .../simple/node/cpus/x86_64_v2/SplitX86.java | 17 +- .../simple/node/cpus/x86_64_v2/StoreX86.java | 4 +- .../simple/node/cpus/x86_64_v2/TFPX86.java | 48 +- .../simple/node/cpus/x86_64_v2/TMPX86.java | 21 +- .../simple/node/cpus/x86_64_v2/x86_64_v2.java | 155 +- .../com/seaofnodes/simple/print/ASMPrinter.java | 40 +- .../com/seaofnodes/simple/print/IRPrinter.java | 17 +- .../simple/print/SimpleGraphAdapter.java | 43 +- .../simple/print/SimpleGraphObserver.java | 2 +- .../java/com/seaofnodes/simple/type/Field.java | 82 +- src/main/java/com/seaofnodes/simple/type/Type.java | 141 +- .../com/seaofnodes/simple/type/TypeConAry.java | 9 +- .../com/seaofnodes/simple/type/TypeConAryB.java | 5 +- .../java/com/seaofnodes/simple/type/TypeFloat.java | 17 +- .../com/seaofnodes/simple/type/TypeFunPtr.java | 195 +- .../com/seaofnodes/simple/type/TypeInteger.java | 34 +- .../java/com/seaofnodes/simple/type/TypeMem.java | 268 +- .../com/seaofnodes/simple/type/TypeMemPtr.java | 90 +- .../java/com/seaofnodes/simple/type/TypeNil.java | 1 - .../java/com/seaofnodes/simple/type/TypePtr.java | 8 +- .../java/com/seaofnodes/simple/type/TypeRPC.java | 95 +- .../com/seaofnodes/simple/type/TypeScalar.java | 32 +- .../com/seaofnodes/simple/type/TypeStruct.java | 398 ++- .../java/com/seaofnodes/simple/type/TypeTuple.java | 76 +- src/main/java/com/seaofnodes/simple/type/XInt.java | 15 +- src/main/java/com/seaofnodes/simple/util/BAOS.java | 162 ++ .../com/seaofnodes/simple/util/IntHashMap.java | 87 + src/main/java/com/seaofnodes/simple/util/SB.java | 2 +- src/main/smp/sys.smp | 206 +- src/main/smp/sys/adt.smp | 3 + src/main/smp/sys/adt/bitset.smp | 57 + src/main/smp/sys/ary.smp | 24 + src/main/smp/sys/aryi64.smp | 82 + src/main/smp/sys/aryu8.smp | 87 + src/main/smp/sys/char.smp | 8 + src/main/smp/sys/io.smp | 64 + src/main/smp/sys/libc.smp | 44 + src/main/smp/sys/scan.smp | 39 + 183 files changed, 9107 insertions(+), 5344 deletions(-) create mode 100644 src/main/c/simple_crt.c create mode 100644 src/main/java/com/seaofnodes/simple/codegen/ClassInitDependencies.java create mode 100644 src/main/java/com/seaofnodes/simple/codegen/CompUnit.java delete mode 100644 src/main/java/com/seaofnodes/simple/codegen/ElfExpr.java delete mode 100644 src/main/java/com/seaofnodes/simple/codegen/ElfFile.java create mode 100644 src/main/java/com/seaofnodes/simple/codegen/ElfWriter.java create mode 100644 src/main/java/com/seaofnodes/simple/codegen/FinishCalls.java create mode 100644 src/main/java/com/seaofnodes/simple/codegen/Iter.java create mode 100644 src/main/java/com/seaofnodes/simple/codegen/LocalIds.java create mode 100644 src/main/java/com/seaofnodes/simple/codegen/ParseAll.java create mode 100644 src/main/java/com/seaofnodes/simple/codegen/SourceError.java create mode 100644 src/main/java/com/seaofnodes/simple/codegen/StaticData.java delete mode 100644 src/main/java/com/seaofnodes/simple/node/AddFNode.java delete mode 100644 src/main/java/com/seaofnodes/simple/node/ArithFNode.java create mode 100644 src/main/java/com/seaofnodes/simple/node/CLInitNode.java delete mode 100644 src/main/java/com/seaofnodes/simple/node/CastMach.java delete mode 100644 src/main/java/com/seaofnodes/simple/node/CastNode.java delete mode 100644 src/main/java/com/seaofnodes/simple/node/DivFNode.java delete mode 100644 src/main/java/com/seaofnodes/simple/node/EscapeMach.java create mode 100644 src/main/java/com/seaofnodes/simple/node/ExternOffsetNode.java delete mode 100644 src/main/java/com/seaofnodes/simple/node/MinusFNode.java delete mode 100644 src/main/java/com/seaofnodes/simple/node/MulFNode.java create mode 100644 src/main/java/com/seaofnodes/simple/node/Proj.java create mode 100644 src/main/java/com/seaofnodes/simple/node/StartCUNode.java create mode 100644 src/main/java/com/seaofnodes/simple/node/StopCUNode.java delete mode 100644 src/main/java/com/seaofnodes/simple/node/StructNode.java delete mode 100644 src/main/java/com/seaofnodes/simple/node/SubFNode.java create mode 100644 src/main/java/com/seaofnodes/simple/node/TypeNode.java delete mode 100644 src/main/java/com/seaofnodes/simple/node/cpus/arm/FloatARM.java create mode 100644 src/main/java/com/seaofnodes/simple/node/cpus/arm/FltARM.java create mode 100644 src/main/java/com/seaofnodes/simple/util/BAOS.java create mode 100644 src/main/java/com/seaofnodes/simple/util/IntHashMap.java create mode 100644 src/main/smp/sys/adt.smp create mode 100644 src/main/smp/sys/adt/bitset.smp create mode 100644 src/main/smp/sys/ary.smp create mode 100644 src/main/smp/sys/aryi64.smp create mode 100644 src/main/smp/sys/aryu8.smp create mode 100644 src/main/smp/sys/char.smp create mode 100644 src/main/smp/sys/io.smp create mode 100644 src/main/smp/sys/libc.smp create mode 100644 src/main/smp/sys/scan.smp (limited to 'src/main') diff --git a/src/main/c/simple_crt.c b/src/main/c/simple_crt.c new file mode 100644 index 00000000..8656a3cc --- /dev/null +++ b/src/main/c/simple_crt.c @@ -0,0 +1,61 @@ +#include +#include +#include +#include + +// errno may be thread-local or a macro calling a platform accessor. +int32_t simple_errno(void) { return errno; } + +// Native process arguments converted to ordinary Simple array layouts. +// The outer array is mutable; every contained byte array is published as +// immutable and non-null to Simple code. +typedef struct SimpleBytes { + uint32_t len; + unsigned char bytes[]; +} SimpleBytes; + +typedef struct SimpleArgs { + uint32_t len; + uint32_t pad; + SimpleBytes *args[]; +} SimpleArgs; + +__attribute__((CALL_CONV)) +extern int64_t simple_main(SimpleArgs *args); + +static SimpleArgs *simple_args(int argc, char **argv) { + if (argc < 0 || (uint64_t)argc > UINT32_MAX) + return NULL; + + SimpleArgs *args = calloc(1, sizeof(*args) + (size_t)argc * sizeof(args->args[0])); + if (args == NULL) + return NULL; + args->len = (uint32_t)argc; + + for (int i = 0; i < argc; i++) { + size_t len = strlen(argv[i]); + if (len > UINT32_MAX) { + free(args); + return NULL; + } + SimpleBytes *str = malloc(sizeof(*str) + len); + if (str == NULL) { + for (int j = 0; j < i; j++) free(args->args[j]); + free(args); + return NULL; + } + str->len = (uint32_t)len; + memcpy(str->bytes, argv[i], len); + args->args[i] = str; + } + return args; +} + +int main(int argc, char **argv) { + SimpleArgs *args = simple_args(argc, argv); + if (args == NULL) + return 1; + // The language-level hidden receiver is not part of the emitted ABI for + // free functions; source argument zero occupies native argument zero. + return (int)simple_main(args); +} diff --git a/src/main/java/com/seaofnodes/simple/IterPeeps.java b/src/main/java/com/seaofnodes/simple/IterPeeps.java index 2fe4e172..e65ef6d9 100644 --- a/src/main/java/com/seaofnodes/simple/IterPeeps.java +++ b/src/main/java/com/seaofnodes/simple/IterPeeps.java @@ -1,12 +1,14 @@ package com.seaofnodes.simple; import com.seaofnodes.simple.codegen.CodeGen; +import com.seaofnodes.simple.codegen.Opto; import com.seaofnodes.simple.type.Type; import com.seaofnodes.simple.type.TypeInteger; import com.seaofnodes.simple.util.Ary; import com.seaofnodes.simple.node.*; import java.util.Arrays; import java.util.BitSet; +import java.util.IdentityHashMap; import java.util.Random; /** @@ -38,7 +40,7 @@ import java.util.Random; *
  • Our strong invariant is that for all Nodes, either they are on the worklist * OR no peephole applies. This invariant is easy to check, although expensive. * Basically the normal "iterate peepholes to a fixed point" is linear, and this - * check is linear at each peephole step... so quadratic overall. Its a useful + * check is linear at each peephole step... so quadratic overall. It's a useful * assert, but one we can disable once the overall algorithm is stable - and * then turn it back on again when some new set of peepholes is misbehaving. * The code for this is turned on in `IterPeeps.iterate` as `assert @@ -48,11 +50,11 @@ import java.util.Random; public class IterPeeps { public final WorkList _work; - - private final WorkList _workInline; + final WorkList _workInline; + private int _cnt; public IterPeeps( long seed ) { - _work = new WorkList<>(seed); + _work = new WorkList<>(seed); _workInline = new WorkList<>(seed); } @@ -62,36 +64,53 @@ public class IterPeeps { public void addAll( Ary ary ) { _work.addAll(ary); } /** - * Iterate peepholes to a fixed point + * Iterate peepholes and inlining to a fixed point */ public void iterate( CodeGen code ) { + boolean didInline = false; + Ary defer = new Ary<>(Node.class); while( true ) { + // Clean up everything that does not grow the code iteratePeeps(code); + + // Pick an inline candidate, no real heuristic, first come, first served + boolean inlined = false; CallEndNode cend; - while( (cend=_workInline.pop())!=null ) - if( cend.inlineSmall() ) break; - if( cend==null ) return; + while( (cend=_workInline.pop()) != null) { + if( cend.isDead() ) continue; + byte inline = cend.maybeInline(); + if( inline == -1 ) defer.add(cend); + if( inline > 0 ) { + inlined = true; + break; + } + } + // Inlined, run peeps until clean again + if( inlined ) { didInline = true; continue; } + // No more candidates, check the defer list + if( !didInline ) break; // No more progress, so all the "maybe inline after cleanup" do not progress + // Some inlining happened, retry all the "try again after cleanup" calls + didInline = false; + code.addAll(defer); + defer.clear(); } } - // Only try growing inlines once the shrinking peepholes reach a fixed point. + // Run all the code-reduction and type-lifting peeps as possible private void iteratePeeps( CodeGen code ) { - assert progressOnList(code, _work, true); - int cnt=0; + assert !CodeGen.expensiveAssert(0) || (progressOnList(code, _work) && schedulableUses(code)); Node n; while( (n=_work.pop()) != null ) { if( n.isDead() ) continue; - if( n instanceof CallEndNode cend ) _workInline.push(cend); - cnt++; // Useful for debugging, searching which peephole broke things - var obs = CodeGen.CODE._obs; + _cnt++; // Useful for debugging, searching which peephole broke things + var obs = code._obs; // Collect peephole neighbors for graph display if( obs != null ) obs.before(n); Node x = n.peepholeOpt(); + if( n instanceof CallEndNode cend ) + _workInline.push(cend); if( x != null ) { - if( x.isDead() ) { - if( obs != null ) obs.after(n, x, true); - continue; - } + assert !x.isDead(); // Peepholes return alive answers // peepholeOpt can return brand-new nodes, needing an initial type set if( x._type==null ) x.setType(x.compute()); // Changes require neighbors onto the worklist @@ -101,22 +120,31 @@ public class IterPeeps { // Everybody gets a free "go again" in case they didn't get // made in their final form. _work.push(x); + // A self-returning peephole can have rewritten its input + // edges. The new defs gained a user and may have + // backwards, user-sensitive peepholes of their own. + for( Node z : x._inputs ) _work.push(z); // If the result is not self, revisit all inputs (because // there's a new user), and replace in the graph. if( x != n ) { - for( Node z : n. _inputs ) _work.push(z); + //for( Node z : n. _inputs ) _work.push(z); for( Node z : x._outputs ) _work.push(z); n.subsume(x); } } // If there are distant neighbors, move to worklist n.moveDepsToWorklist(); - assert progressOnList(code, _work, true); // Very expensive assert + // Very expensive assert. + assert !CodeGen.expensiveAssert(_cnt) || (progressOnList(code, _work) && schedulableUses(code)); } - if( n.isUnused() && !(n instanceof StopNode) ) + boolean unused = n.isUnused(); + if( unused ) { + assert !(n instanceof StopNode); // StopNodes can die if all code in the compunit dies n.kill(); // Just plain dead - if( obs != null ) obs.after(n, x, true); + } + if( obs != null ) obs.after(n, x!=null ? x : unused ? n : null, true); } + assert !CodeGen.expensiveAssert(0) || schedulableUses(code); } @@ -133,27 +161,112 @@ public class IterPeeps { // {@link #Node.addDep} which is a side effect in an assert. The {@link // #midAssert} is used to stop this side effect. // Pessimistic solver assert - public static boolean progressOnList(CodeGen code, WorkList list, boolean dir ) { + public static boolean progressOnList(CodeGen code, WorkList list ) { code._midAssert = true; Node changed = code._stop.walk( n -> { - Node m = n; + // Ignore most in-progress things + if( n.iskeep() ) return null; + // Types must be forwards, even if on the worklist. Type nval = n.compute(); - if( (!n.iskeep() || n._nid<=8) && // Types must be forwards, even if on worklist - ( dir - ? nval.isa(n._type) // Pesi: new value lifts over old - : n._type.isa(nval) // Opto: new value falls over old - ) ) { - if( list.on(n) ) return null; - m = n.peepholeOpt(); - if( m==null ) return null; + assert nval.isa(n._type) : "Non-monotonic peep: "+n+"#"+n._nid+" old="+n._type+" new="+nval+" inputs="+inputTypes(n)+" peep="+n; + if( list.on(n) ) + return null; // On worklist is ok! + if( n instanceof CallEndNode cend ) { + if( code._iter._workInline.on(cend) ) + return null; // On inline worklist is ok! + assert cend.maybeInline() <= 0 : "Inline fired and not on worklist, CallEndNode#"+cend._nid; } - System.err.println("BREAK HERE FOR BUG"); + // Off list: must be live, in GVN table, and no peeps. + assert n.nOuts() > 0 : "Unused live node: "+n; + assert code._gvn.get(n)==n : "Not in GVN table: "+n; + Node m = n.peepholeOpt(); + assert m==null : "Peep fired and not on worklist, "+n.getClass().getSimpleName()+"#"+n._nid+" -> "+m; return m; }); code._midAssert = false; return changed==null; } + private static String inputTypes(Node n) { + StringBuilder sb = new StringBuilder("["); + for( Node in : n._inputs ) + sb.append(in==null ? "null" : in.getClass().getSimpleName()+"#"+in._nid+"="+in+":"+in._type+":keep="+in.iskeep()).append(','); + return sb.append(']').toString(); + } + + // GCM assumes every movable node can be placed no earlier than its inputs + // and no later than the LCA of its uses. Catch graphs where a use escapes + // above an input-defined branch before the later scheduling pass walks off + // the top of the idom tree. + private static final IdentityHashMap earlyCache = new IdentityHashMap<>(); + private static final BitSet VISIT = new BitSet(); + public static boolean schedulableUses(CodeGen code) { + earlyCache.clear(); + Node bad = code._stop.walk( n -> { + if( n.iskeep() || n.isDead() || n.isConst() || + n instanceof CFGNode || n instanceof PhiNode || n instanceof ProjNode ) + return null; + assert VISIT.isEmpty(); + CFGNode early = CFGNode.earlyCFG(n,earlyCache,VISIT); + VISIT.clear(); + if( early == null ) return null; + CFGNode lca = null; + for( Node use : n._outputs ) { + CFGNode ublk = useBlock(n,use); + if( ublk != null ) + lca = ublk.domLCA(lca,null); + } + if( lca == null ) return null; + if( !early.sameFun(lca) ) return null; + assert early.dominates(lca,null) : badSchedule(n,early,lca); + return null; + }); + return bad == null; + } + + private static CFGNode useBlock(Node n, Node use) { + if( use == null ) + return null; + if( use instanceof PhiNode phi ) { + CFGNode found = null; + for( int i=1; i= phi.region().nIns() ) + return null; + found = phi.region().cfg(i).domLCA(found,null); + } + return found; + } + return CFGNode.safeCFG(use); + } + + private static String badSchedule(Node n, CFGNode early, CFGNode lca) { + StringBuilder sb = new StringBuilder("Unschedulable data node ") + .append(n.getClass().getSimpleName()).append('#').append(n._nid) + .append(" early=").append(cfg(early)) + .append(" use-lca=").append(cfg(lca)) + .append(" node=").append(n).append("\ninputs:"); + for( int i=0; i _xScopes = new Stack<>(); - ScopeNode _continueScope; - ScopeNode _breakScope; // Merge all the while-breaks here - FunNode _fun; // Current function being parsed + ScopeNode _breakScope; // Merge all the while-breaks here + ScopeNode _continueScope; // Merge all the while-continues here + ScopeNode _returnScope; // Merge all the function exits here + Ary _frefs; // Forward refs created during this parse + TypeStruct _ctorOpenStruct; // Open struct being initialized by an allocation constructor block. + // Constructor exit types survive scope teardown until all fields are known. + // A null entry denotes a constructor with no returning path. + private ArrayList> _ctorExits; // Mapping from a type name to a Type. The string name matches // `type.str()` call. - public static HashMap TYPES = defaultTypes(); - - // Mapping from a type name to the constructor for a Type. - public final HashMap INITS; - - - public Parser(CodeGen code, TypeInteger arg) { - _code = code; - _scope = new ScopeNode(); - _continueScope = _breakScope = null; - ZERO = con(TypeInteger.ZERO).keep(); - NIL = con(Type.NIL).keep(); - XCTRL= new XCtrlNode().peephole().keep(); - TYPES = defaultTypes(); - INITS = new HashMap<>(); - } - - @Override - public String toString() { return _lexer.toString(); } - - public static HashMap defaultTypes() { - return new HashMap<>() {{ + public static final HashMap TYPES = new HashMap<>(); + public static final HashMap INIT_TYPES = + new HashMap<>() {{ put("bool",TypeInteger.U1 ); put("byte",TypeInteger.U8 ); put("f32" ,TypeFloat .F32); @@ -123,242 +127,442 @@ public class Parser { put("val" ,Type.TOP); // Marker type, indicates type inference put("var" ,Type.BOTTOM); // Marker type, indicates type inference }}; + + + public Parser(CodeGen code ) { + _code = code; } + @Override + public String toString() { return _lexer==null ? "" : _lexer.toString(); } + // Debugging utility to find a Node by index public Node f(int nid) { return _code.f(nid); } + // Read and set CTRL private Node ctrl() { return _scope.ctrl(); } private N ctrl(N n) { return _scope.ctrl(n); } - - public void parse() { - // Graph callbacks during scope setup need a source position. - _lexer = new Lexer(sys.SYS); - + // Read and set Memory + private Node mem() { return _scope.mem(1); } + private void mem(Node mem) { _scope.mem(mem); } + // Structural label for precise memory Phis. The '$' prefix cannot be + // referenced in Simple source. + public static String memName(int alias) { return ("$"+alias).intern(); } + + // Ordinary parsing carries one whole-memory value. Alias partitioning is + // graph optimization, not parser state. + private void storeMem( Node st, int alias, Node prior ) { + mem(alias==1 ? st : mergeAlias(prior,alias,st)); + } + + // A semantic graph partition representing the complete memory after a + // precise alias update. ScopeNode treats this as one opaque value. + private Node mergeAlias(Node bulk, int alias, Node precise) { + assert bulk != null : "Missing bulk memory for alias #"+alias; + MemMergeNode merge = new MemMergeNode(null,bulk); + merge.alias(alias,precise); + return peep(merge); + } + + + public Ary parse( CompUnit ref ) { + assert _scope == null && _breakScope == null && _continueScope == null && _returnScope == null; + assert _frefs == null; + _lexer = new Lexer(ref); + _frefs = new Ary<>(FRefNode.class); + // Starting Scope has control, memory, initial arguments + _scope = new ScopeNode(); _scope.define(ScopeNode.CTRL, Type.CONTROL , false, null, _lexer); _scope.define(ScopeNode.MEM0, TypeMem.BOT , false, null, _lexer); - _scope.define(ScopeNode.ARG0, TypeInteger.BOT, false, null, _lexer); - - ctrl(XCTRL); - _scope.mem(con(TypeMem.BOT)); + // Track active scopes for Graph display + _xScopes.push(_scope); + ctrl(_code.XCTRL); + mem(con(TypeMem.BOT)); - // Parse the sys import - while( !_lexer.isEOF() ) { - parseStatement(); - _lexer.skipWhiteSpace(); - } + // File-level struct declaration + _ref = ref; + _nestedType = ref._cname; + String typeName = addClzPrefix(ref._cname); - // Reset lexer for program text - _lexer = new Lexer(_code._src); - _xScopes.push(_scope); + ReturnNode clzret = parseStruct( true, typeName ); - // Parse whole program, as-if function header "{ int arg -> body }" - parseFunctionBody( _code._main,loc(),"arg"); - - // Kill an empty default main. Keep only if it was explicitly defined - // (programmer asked for a "main") or it has stuff (i.e. beginner - // default main). - FunNode main = _code.link(_code._main); - StopNode stop = _code._stop; - // Gather an explicit main - FunNode xmain = null; - for( Node n : stop._inputs ) - if( n instanceof ReturnNode ret && "main".equals(ret.fun()._name) ) - if( xmain != null ) throw Utils.TODO(); - else xmain = ret.fun(); - - if( main.ret().expr()._type==Type.TOP && main.uctrl()==null ) { - // Kill an empty default main; so it does not attempt to put a - // "main" in any final ELF file - main.setDef(1,XCTRL); // Delete default start input - stop.delDef(stop._inputs.find(main.ret())); - if( xmain != null ) - _code.setMain(xmain); + if( !_lexer.isEOF() ) throw _errorSyntax("unexpected"); - } else { - // We have a non-empty default main. - if( xmain != null ) - // Found an explicit "main" AND we have a default "main" - throw error("Cannot define both an explicit main and a default main"); - main._name = "main"; + // All still-unresolved forward refs + // should immediately resolve to a file-based name which is a Simple + // class type. + Ary frefs = new Ary<>(FRefNode.class); + for( FRefNode fref : _frefs ) { + if( fref.nIns() != 1 && fref._con == FRefNode.FREF_TYPE ) continue; + CompUnit cu = ParseAll.findCompUnitOrThrow(_code,_ref,fref._name); + fref._con = TypeMemPtr.make((byte)2,TypeStruct.make(addClzPrefix(cu._cname),true),true); + frefs.push(fref); } - if( !_lexer.isEOF() ) throw _errorSyntax("unexpected"); + // The class being published + ref._clz = (TypeStruct)TYPES.get(typeName); // Clean up and reset _xScopes.pop(); _scope.kill(); - for( StructNode init : INITS.values() ) - init.unkeep().kill(); - INITS.clear(); - for( Ary captures : CAPTURES.values() ) - for( Node value : captures ) value.unkill(); - CAPTURES.clear(); - stop.peephole(); - // Helpers are built with init(), so their complete call graph exists - // before any inlining. Seed those nodes now along with the parsed body. - _code._start.walk(n -> { _code.add(n); return null; }); + _scope = null; + _frefs = null; + + return frefs; } + /** * Parses a function body, assuming the header is parsed. */ - private ReturnNode parseFunctionBody( TypeFunPtr sig, Lexer loc, String... ids) { + private ReturnNode parseFunctionBody( TypeFunPtr sig, Lexer loc, String[] ids) { return parseFunctionBody(sig, loc, null, ids); } - private FunNode _privateFun; - private HashMap _fieldVars; - private record Constructor(TypeFunPtr sig, Lexer loc, int body, String[] ids, java.util.BitSet finals) {} - private Constructor _constructor; - private final HashMap ALLOCS = new HashMap<>(); - private final HashMap INITIALIZERS = new HashMap<>(); - private final HashMap> CAPTURES = new HashMap<>(); - private Ary _captures; + private ReturnNode parseFunctionBody( TypeFunPtr sig, Lexer loc, java.util.BitSet finals, String[] ids) { + // Record & restore the existing scope vars. The function parse might + // update fields, so when the parse is done reset back to the pre-parse + // state. + Node[] preParse = _scope.save(); + + // Parse function body normally + ReturnNode ret = _parseFunctionBody(null, sig,loc,finals,ids); + + // Reset all fields to pre-parse days + _scope.restore(preParse); + return ret; + } + + + // Parse a function body + + // Stack the {break,continue,return} scopes and a new lexical scope. + // Build a FunNode and parameters, extending scope by parm names. + // Parse the function body + // - while( not end-of-function ) + // - - parseStatement + // - - If isInit, also incrementally extend the self-type. + // If isInit + // - upgrade and close the self-type + // - Gather fields and store values + // - returns self, not last + // Unstack the {break,continue,return} scopes. - private ReturnNode parseFunctionBody(TypeFunPtr sig, Lexer loc, java.util.BitSet finals, String... ids) { - return parseFunctionBody(sig,loc,finals,false,ids); + private ReturnNode _parseFunctionBody( String funName, TypeFunPtr sig, Lexer loc, String[] ids) { + return _parseFunctionBody(funName, sig, loc, null, ids); } - private ReturnNode parseFunctionBody(TypeFunPtr sig, Lexer loc, java.util.BitSet finals, boolean initializer, String... ids) { - // If this is a method, record & restore the existing struct fields, - // which are also the struct initializers. The method might update - // mutable fields, so when the function parse is done reset back to - // pre-function-parse values. - Node[] preParse = new Node[_scope.nIns()]; - for( int i=0; i<_scope.nIns(); i++ ) - preParse[i] = _scope.in(i)==null ? null : _scope.in(i).keep(); + private ReturnNode _parseFunctionBody( String funName, TypeFunPtr sig, Lexer loc, java.util.BitSet finals, String[] ids) { + Constructor outerCtor=_pendingCtor; + _pendingCtor=null; + boolean isInit = FunNode.isInit(funName); + boolean isCtor = isInit || isExplicitConstructorName(funName); + String typeName = isCtor ? ((TypeMemPtr)sig._sig[0])._obj._name : null; // Stack parser state on the local Java stack, and unstack it later - FunNode oldfun = _fun; - FunNode oldPrivate = _privateFun; - HashMap oldFields = _fieldVars; - Constructor oldConstructor = _constructor; - _fieldVars = new HashMap<>(); - _constructor = null; - Ary oldCaptures = _captures; - _captures = initializer ? new Ary<>(Node.class) : null; - ScopeNode breakScope = _breakScope; _breakScope = null; + _scope.push(new Kind.Func(funName,isCtor && !funName.endsWith("."))); + ScopeNode breakScope = _breakScope; _breakScope = null; ScopeNode continueScope = _continueScope; _continueScope = null; + ScopeNode returnScope = _returnScope; _returnScope = null; - FunNode fun = _fun = (FunNode)peep(new FunNode(loc(),sig,_nestedType,null,_code._start)); + int oldUID = _code.UID(); // Used to approximate function size + FunNode fun = (FunNode)peep(new FunNode(loc(),sig, funName, _ref, null, _ref._start)); // Once the function header is available, install in linker table - // allowing recursive functions. Linker matches on declared args and // exact fidx, and ignores the return (because the fidx will only match // the exact single function). _code.link(fun); - _privateFun = sig.ret() instanceof TypeMem ? fun : null; - fun._constructor = _privateFun!=null; - - Node rpc = new ParmNode("$rpc",0,TypeRPC.BOT,fun,con(TypeRPC.BOT)).peephole(); - - // Build a multi-exit return point for all function returns - RegionNode r = new RegionNode( null,null,null).init(); - assert r.inProgress(); - PhiNode rmem = new BulkMemPhiNode(ScopeNode.MEM0,r,null).init(); - PhiNode rrez = (sig.ret() instanceof TypeMem - ? new BulkMemPhiNode("#selfMem",r,null) : new PhiNode(ScopeNode.ARG0,r,null)).init(); - ReturnNode ret = new ReturnNode(r, rmem, rrez, rpc, fun).init(); - fun.setRet(ret); - assert ret.inProgress(); - _code._stop.addDef(ret); + Node rpc = new ParmNode("$rpc",0,TypeRPC.BOT,fun,ConstantNode.seed(TypeRPC.BOT).peephole()).peephole(); // Pre-call the function from Start, with worse-case arguments. This // represents all the future, yet-to-be-parsed functions calls and // external calls. - _scope.push(new Kind.Func(_privateFun!=null)); - ctrl(fun); // Scope control from function - // Unknown callers supply whole memory; field aliases are discovered lazily. - _scope.mem(new ParmNode(ScopeNode.MEM0,1,TypeMem.BOT,fun,new ProjNode(_code._start,1,ScopeNode.MEM0).peephole()).peephole()); + _scope.ctrl(fun); // Scope control from function + // Private mem alias tracking per function + Node defaultMem = new ProjNode(_ref._start,1,ScopeNode.MEM0).peephole(); + Node privMem = new ParmNode(ScopeNode.MEM0,1,TypeMem.BOT,fun,defaultMem).peephole(); + mem(privMem); // All args, "as-if" called externally for( int i=0; i=0; j-- ) - if( _scope.var(j)._name.equals(f._fname) ) { v=_scope.var(j); break; } - assert v!=null; - _fieldVars.put(f._fname,v); - if( !f._one ) { - Node value=missingInit(defaults.in(i),new java.util.BitSet()) ? con(Type.BOTTOM) - : new LoadNode(loc(),f._fname,f._alias,privateMem(),self(),off(shape,f._fname),true).peephole(); - _scope.update(v,value); - } + // Args take a default input if called from Start/external-world + _scope.define(ids[i], t, (i==0 && ids[0]=="self") || (finals!=null && finals.get(i)), new ParmNode(ids[i],i+2,t,fun,ConstantNode.seed(t).peephole()).peephole(), loc); + } + + if( isCtor && !isInit ) { + int lex=_scope.klast()._lexSize; + TypeStruct shape=(TypeStruct)TYPES.get(typeName); + for( Field fld : shape._fields ) { + int idx=fieldVarIdxBefore(lex,fld._fname); + if( idx<0 ) continue; + Var v=_scope.var(idx); + if( v._fref || v._extern!=0 || v._uninit || _scope.in(v)._type.isConstant() ) continue; + Node value=convertExpr( + new LoadNode(loc,fld._fname,fld._alias,ctrl(),fun.parm(3),fun.parm(2),off(shape,fld._fname)).peephole(),v.type()); + _scope.update(v,value); } } - Node last=ZERO; + + // The initializer owns its run-once flag store. Parent initialization + // is a structural requirement, executed before this body by CLInit lowering. + if( fun.isClz() ) { + Node self = fun.parm(2); + Field flag = ((TypeMemPtr)self._type)._obj.field("$init"); + Node prior = mem().keep(); + Node st = peep(new StoreNode(loc,"$init",flag._alias,flag._t,ctrl(),prior,self, + fldoff(self,"$init"),con(1),false)); + storeMem(st,flag._alias,prior); + prior.unkeep(); + } + + // Parse the function body. + Node last=_code.ZERO; // Last statement as the default result + + int nvar = 0; while (!peek('}') && !_lexer.isEOF()) { - last = parseStatement(); - if( initializer ) fields = initFields(fields); - } - if( initializer ) { - finishFields(); - CAPTURES.put(_nestedType,_captures); - if( _constructor!=null ) { - Constructor ctor=_constructor; - int end=pos(ctor.body()); - ReturnNode body=parseFunctionBody(ctor.sig(),ctor.loc(),ctor.finals(),ctor.ids()); - body.fun().setName(shortName(_nestedType)); - require("}"); - pos(end); + // Parse a statement; record last statement as default return + last = _code.add(parseStatement()); + // For and , local vars are really struct fields and + // get exposed as discovered. This allows late local functions to + // use early declared fields. + if( isInit ) + nvar = updateSelfAsFieldsDiscovered( typeName, nvar ); + } + + if( isInit ) finishConstructor(); + _pendingCtor=outerCtor; + + ParmNode self = fun.parm(2); + // Last expression is the return + if( ctrl()._type==Type.CONTROL ) { + // Constructors get special return types + if( fun.isClz() ) { + assert last!=null; + } else if( isCtor ) { + last = fun.parm(3); + } + addReturn(last); + // exit path is now dead (nothing live after a "return") + } + if( isCtor ) + upgradeSelfTypeAndStoreFields( typeName, fun.isClz(), self, fun.parm(3), isInit ); + + // Build a return from the _returnScope. + // Can be no returns for never-exit functions + Node rctl = _returnScope==null ? _code.XCTRL : _returnScope.ctrl().peephole(); + Node rmem = rctl==_code.XCTRL ? con(TypeMem.TOP) : _returnScope.mem(); + Node expr = rctl==_code.XCTRL ? con(Type.TOP) : _returnScope._inputs.last().peephole(); + ReturnNode ret = (ReturnNode)peep(new ReturnNode(rctl, rmem, expr, rpc, fun)); + fun.setRet(ret); + // Approximate function size, for inlining heuristics + fun._approxUIDs = _code.UID() - oldUID; + for( Node cend : ret.outs() ) + if( cend instanceof CallEndNode ) + throw TODO(); //_code.add(cend); // have call ends ALREADY? just barely made the Return + + // Export (and keep-alive) functions until prove dead + _ref.addFun(_code,fun); + + // Unstack parser state + if( _returnScope != null ) _returnScope.kill(); + _returnScope = returnScope; + _continueScope = continueScope; + _breakScope = breakScope; + _scope.pop(); + + return ret; + } + + private int updateSelfAsFieldsDiscovered( String typeName, int nvar ) { + int lex = _scope.klast()._lexSize + 2/*Skip self and selfMem*/; + // Load self type in and out of the TYPES hashtable. + TypeStruct tself = (TypeStruct)TYPES.get(typeName); + // Start looking for newly declared vars at the lexical scope start, + // skipping two ("self" and "selfmem") and fields seen already. + while( lex + nvar < _scope.nIns() ) { + Var v = _scope.var(lex + nvar++); + // Field with sharper type + Field old = tself.field(v._name); + Field fld = v._extern!=0 + ? Field.make(v._name,v.type(),v._extern,false,true) + : old == null + ? Field.make(v._name,v.type(),_code.alias(_ref._cname),v._final) + : old.makeFrom(v.type()); + tself = tself.addOrUpdate(fld); + } + TYPES.put(typeName,tself); + return nvar; + } + + + void upgradeSelfTypeAndStoreFields( String typeName, boolean isClz, ParmNode self, ParmNode smem, boolean upgrade ) { + int lex = _scope.klast()._lexSize; + assert _scope.var(lex)._name=="self"; + TypeStruct tself = (TypeStruct)TYPES.get(typeName); + int base = lex+2; // Skip ctrl, mem + if( upgrade ) { + // Default / defines the struct and can sharpen fields. + // A named constructor only initializes the already-closed struct. + for( int i=0; i receives the + // broad allocation pointer, while private memory tracks initialized fields. + self._declaredType = self._type = isClz ? TypeMemPtr.make((byte)2,tself,true) : TypeMemPtr.make(constructorRecv(tself)); + FunNode fun = self.fun(); + fun.setSig(fun.sig().makeFrom(self._type,0)); + if( !isClz ) // returns upgraded private memory + smem._declaredType = smem._type = TypeMem.makePrivate(tself); + } + + if( !upgrade ) { + IdentityHashMap exit = null; + if( _returnScope!=null && !_returnScope.ctrl()._type.isHigh() ) { + exit = new IdentityHashMap<>(); + int first = _scope._kinds.at(_scope.depth()-2)._lexSize+2; + for( int i=first; i updates the private memory that got + // passed in - is treated like a normal argument and not like memory. + Node mmm = isClz ? _returnScope.mem(1) : smem; + + + // Store constructor results into fields + for( int i=0; i return memory. + _returnScope.mem(mmm); + //_code.add(mmm); + } + + // Validate each user constructor, or an implicit empty one, before the + // declaration scope goes away. Later fields use their declaration defaults. + private static boolean initialized(Type t) { + return t!=Type.BOTTOM && !t.isHigh() && (!(t instanceof TypeNil tn) || tn.notNull()); + } + + private void checkInitFields(TypeStruct ts, int base) { + if( _returnScope==null || _returnScope.ctrl()._type.isHigh() ) return; + boolean implicit = _ctorExits.isEmpty(); + if( implicit ) _ctorExits.add(new IdentityHashMap<>()); + for( IdentityHashMap exit : _ctorExits ) { + if( exit==null ) continue; + for( Field fld : ts._fields ) { + int idx = fieldVarIdx(base,fld._fname); + if( idx == -1 ) continue; + Var v = _scope.var(idx); + if( !v._uninit || v._fref ) continue; + Type t = exit.getOrDefault(v,_returnScope.in(idx)._type); + if( !initialized(t) ) + throw error("'"+ts._name+"' is not fully initialized, field '"+fld._fname+ + (implicit ? "' needs to be set in a constructor" : + "' is only partially set in the constructor"),v._loc); } + } + } - // Pop off the inProgress node on the multi-exit Region merge - assert r.inProgress(); - r ._inputs.pop(); - rmem._inputs.pop(); - rrez._inputs.pop(); - r._loc = loc(); // Final position - assert !r.inProgress(); - - // Force peeps, which have been avoided due to inProgress - ret.setDef(1,rmem.peephole()); - ret.setDef(2,rrez.peephole()); - ret.setDef(0,r.peephole()); - ret = (ReturnNode)ret.peephole(); - - // Function scope ends - if( initializer ) _scope.pop(); - _scope.pop(); - _fun = oldfun; - _privateFun = oldPrivate; - _fieldVars = oldFields; - _constructor = oldConstructor; - _captures = oldCaptures; - _breakScope = breakScope; - _continueScope = continueScope; + private int fieldVarIdx( int base, String name ) { + for( int i=base; i<_scope.nIns(); i++ ) + if( _scope.var(i)._name == name ) + return i; + return -1; + } - // Reset all fields to pre-parse days - for( int i=0; i=0; i-- ) + if( _scope.var(i)._name == name ) + return i; + return -1; + } - return ret; + private boolean isConstructorDecl(String name) { + TypeStruct self = _scope.constructorSelf(); + return self != null && name == initFieldName(self._name); + } + + private boolean inExplicitConstructor() { + Kind.Func func = (Kind.Func)_scope._kinds.at(_scope.enclosingFuncOrDecl()); + return isExplicitConstructorName(func._name); } + private boolean isExplicitConstructorName(String funName) { + TypeStruct self = _scope.constructorSelf(); + return self != null && funName == (self._name+"."+initFieldName(self._name)).intern(); + } + + private static TypeStruct constructorRecv(TypeStruct self) { + Field[] fs = new Field[self._fields.length]; + for( int i=0; i body }; + private record Constructor(String name, Lexer body) {} + private Constructor _pendingCtor; + private Node parseConstructorDeclaration() { - Lexer loc=loc(); - String name=requireId(); - String expect=shortName(_nestedType); - if( !name.equals(expect) ) - throw error("Constructor name '"+name+"' must match type name '"+expect+"'"); + TypeStruct self = _scope.constructorSelf(); + assert self != null; + String name = requireId(); + String expect = initFieldName(self._name); + if( name != expect ) + throw error("Constructor name '" + name + "' must match type name '" + expect + "'"); require("="); - Node ctor=constructorFunc(); - TypeFunPtr sig=(TypeFunPtr)ctor._type; - Node lift=liftExpr(ctor,sig,true); - if( !_scope.define(name,sig,true,lift,loc) ) - throw error("Redefining constructor '"+name+"'"); - _scope.lookup(name)._init=true; - return require(lift,";"); + if( _pendingCtor!=null ) throw error("Redefining constructor '"+name+"'"); + _pendingCtor=new Constructor(name,loc()); + // Parse the body after all field declarations are known, so self.x + // and x share a local slot even when x is declared below this body. + require("{"); + int depth=1; + while( depth>0 ) { + _lexer.skipWhiteSpace(); + if( _lexer.isEOF() ) throw error("Unclosed constructor body"); + char ch=_lexer.nextChar(); + if( ch=='{' ) depth++; + else if( ch=='}' ) depth--; + else if( ch=='"' || ch=='\'' ) { + char quote=ch; + while( !_lexer.isEOF() ) { + ch=_lexer.nextChar(); + if( ch=='\\' ) _lexer.nextChar(); + else if( ch==quote ) break; + } + } + } + require(";"); + return _code.ZERO; + } + + private void finishConstructor() { + if( _pendingCtor==null ) return; + Constructor ctor=_pendingCtor; + _pendingCtor=null; + Lexer after=_lexer; + _lexer=new Lexer(ctor.body()); + defineConstructor(ctor.name()); + _lexer=after; + } + + private void defineConstructor(String name) { + TypeStruct self=_scope.constructorSelf(); + Node ctor = constructorFunc(name); + TypeFunPtr sig = (TypeFunPtr)ctor._type; + String clzName = addClzPrefix(self._name); + TypeStruct clz = (TypeStruct)TYPES.get(clzName); + Field old = clz==null ? null : clz.field(""); + if( old != null ) + throw error("Redefining constructor '" + name + "'"); + Field fld = Field.make("",sig,_code.alias(_ref._cname),true); + TYPES.put(clzName,clz==null ? TypeStruct.make(clzName,true,fld) : clz.add(fld)); + require(ctor,";"); } /** @@ -475,7 +728,7 @@ public class Parser { // used as an indicator to switch off peepholes of the region and // associated phis; see {@code inProgress()}. - ctrl(new LoopNode(loc(),ctrl()).peephole()); // Note we set back edge to null here + ctrl(new LoopNode(loc(),null,ctrl()).peephole()); // Note we set back edge to null here // At loop head, we clone the current Scope (this includes all // names in every nesting level within the Scope). @@ -494,8 +747,8 @@ public class Parser { // IfNode takes current control and predicate Node ifNode = new IfNode(ctrl(), pred.keep()).peephole(); // Setup projection nodes - Node ifT = new CProjNode(ifNode. keep(), 0, "True" ).peephole().keep(); - Node ifF = new CProjNode(ifNode.unkeep(), 1, "False").peephole(); + CFGNode ifT = new CProjNode(ifNode. keep(), 0, "True" ).peephole().keep(); + CFGNode ifF = (CFGNode)new CProjNode(ifNode.unkeep(), 1, "False").peephole(); // for( ;;next ) body int nextPos = -1, nextEnd = -1; @@ -548,6 +801,8 @@ public class Parser { // it sets the second input to the corresponding input from the back // edge. If the phi is redundant, it is replaced by its sole input. var exit = _breakScope; + head.balanceLoopFRefs(_scope); + exit.balanceLoopFRefs(_scope); head.endLoop(_scope, exit); head.unkeep().kill(); @@ -565,12 +820,12 @@ public class Parser { _xScopes.push( exit ); } _scope = exit; - return ZERO; + return _code.ZERO; } private ScopeNode jumpTo(ScopeNode toScope) { ScopeNode cur = _scope.dup(); - ctrl(XCTRL); // Kill current scope + ctrl(_code.XCTRL); // Kill current scope // Prune nested lexical scopes that have depth > than the loop head // We use _breakScope as a proxy for the loop head scope to obtain the depth while( cur.depth() > _breakScope.depth() ) @@ -588,7 +843,7 @@ public class Parser { private void checkLoopActive() { if (_breakScope == null) throw error("No active loop for a break or continue"); } - private Node parseContinue() { checkLoopActive(); _continueScope = require(jumpTo( _continueScope ),";"); return ZERO; } + private Node parseContinue() { checkLoopActive(); _continueScope = require(jumpTo( _continueScope ),";"); return _code.ZERO; } private Node parseBreak () { checkLoopActive(); // At the time of the break, and loop-exit conditions are only valid if @@ -596,8 +851,8 @@ public class Parser { // conditions here, not the union. _breakScope.removeGuards(_breakScope.ctrl()); _breakScope = require(jumpTo(_breakScope ),";"); - _breakScope.addGuards(_breakScope.ctrl(), null, false); - return ZERO; + _breakScope.addGuards((CFGNode)_breakScope.ctrl(), null, false); + return _code.ZERO; } // Look for an unbalanced `)`, skipping balanced @@ -607,7 +862,7 @@ public class Parser { // Next X char handles skipping complex comments switch( _lexer.nextXChar() ) { case Character.MAX_VALUE: - throw Utils.TODO(); + throw TODO(); case ')': if( --paren<0 ) { posT( pos() - 1 ); @@ -641,8 +896,8 @@ public class Parser { pred.keep(); Node ifNode = new IfNode(ctrl(), pred).peephole(); // Setup projection nodes - Node ifT = new CProjNode(ifNode. keep(), 0, "True" ).peephole().keep(); - Node ifF = new CProjNode(ifNode.unkeep(), 1, "False").peephole().keep(); + CFGNode ifT = new CProjNode(ifNode. keep(), 0, "True" ).peephole().keep(); + CFGNode ifF = new CProjNode(ifNode.unkeep(), 1, "False").peephole().keep(); // In if true branch, the ifT proj node becomes the ctrl // But first clone the scope and set it as current ScopeNode fScope = _scope.dup(); // Duplicate current scope @@ -658,7 +913,7 @@ public class Parser { case ":" -> parseAsgn(); // pred ? asgn; case "&&" -> parseLogical(); // pred && expr case "||" -> pred; // pred || expr_is_ignored - default -> throw Utils.TODO(); + default -> throw TODO(); }; lhs.keep(); _scope.removeGuards(ifT); @@ -681,12 +936,14 @@ public class Parser { Node rhs = switch( fside ) { case "else" -> (doRHS=match(fside)) ? parseStatement() : con(lhs._type.makeZero()); case ":" -> (doRHS=match(fside)) ? parseAsgn() : con(lhs._type.makeZero()); - case "&&" -> rhs = pred; - case "||" -> rhs = parseLogical(); - default -> throw Utils.TODO(); + case "&&" -> con(lhs._type.makeZero()); // pred && expr_is_ignored + case "||" -> parseLogical(); // pred || expr + default -> throw TODO(); }; rhs.keep(); _scope.removeGuards(ifF); + // The RHS of `||` is parsed on the false side, so retain its updated + // control, memory and locals just as an explicit else arm does. if( doRHS || fside.equals("||") ) fScope = _scope; pred.unkeep(); @@ -709,9 +966,10 @@ public class Parser { Node ret = peep(new PhiNode("",r,lhs.unkeep(),rhs.unkeep())); // Immediately fail e.g. `arg ? 7 : ptr` ParseException err; - if( !fside.equals("else") && (err=ret.err()) !=null ) throw err; + if( !fside.equals("else") && (err=ret.err()) !=null ) throw err; + ret.keep(); r.peephole(); - return ret; + return ret.unkeep(); } /** @@ -725,22 +983,69 @@ public class Parser { */ private Node parseReturn() { var expr = require(parseAsgn(), ";"); - // Force the lazy whole-memory Phi before closing this exit. - if( _privateFun==_fun ) - throw error("A constructor cannot return a value"); - _fun.addReturn(ctrl(), _scope.mem(), expr); - ctrl(XCTRL); // Kill control - return expr; + return addReturn(expr); } + private Node addReturn(Node expr) { + expr.keep(); + // Need default memory, since it can be lazy, need to force + // a non-lazy Phi + mem(); // Force a pending lazy bulk-memory Phi + // For and - all fields in scope will be live-on-exit + // and stored into `self` and need to have lazy-phi's inserted. + int lexN = _scope.enclosingFuncOrDecl(); + Kind.Func k = (Kind.Func)_scope._kinds.at(lexN); + if( FunNode.isInit(k._name) ) { + int nestedLexSize = lexN+1 < _scope.depth() ? _scope._kinds.at(lexN+1)._lexSize : _scope.nIns(); + for( int i=k._lexSize; i and - ALL FIELDS IN LAST SCOPE are live + // and need to have lazy-phi's inserted. + + // And fields might not match.... will need matching Var/define with default values + int rlen = _returnScope.nIns()-1; + RegionNode r = ctrl(new RegionNode(null, null,_returnScope.ctrl(), _scope.ctrl()).init().keep()); + _returnScope._merge(_scope, r, rlen); + Node oldExpr = _returnScope.in(rlen); + _returnScope.setDef(rlen,new PhiNode("$expr", r, oldExpr, expr).peephole()); + _code.add(r); + _returnScope.ctrl(r.unkeep()); + } + ctrl(_code.XCTRL); // Kill control + return expr.unkeep(); + } /** Parse: [name '='] expr */ private Node parseAsgn() { int old = pos(); String name = _lexer.matchId(); - int old_pre_equals = pos(); // Just a plain expression, no assignment. // Distinguish `var==expr` from `var=expr` if( name==null || KEYWORDS.contains(name) || !matchOpx('=','=') ) @@ -748,17 +1053,13 @@ public class Parser { // Find variable to update Var def = _scope.lookup(name); + if( def==null ) + def = defineOpenConstructorField(name); if( def==null ) throw error("Undefined name '" + name + "'"); - // If a method var, allow nested blocks and methods, but not fcns or other structs. - // Else safety check for accessing fields out of function scope. - if( instance(def) ) { - pos(old_pre_equals); // setup for `self.name = expr` - Var selfVar = _scope.lookup("self"); - Node self = liftExpr(_scope.in(selfVar),selfVar.type(),true); - return parsePostfixName(self,name); - } + if( def._extern!=0 ) // Parse the C binding as a memory lvalue + { pos(old); return parseExpression(); } // TOP fields are for late-initialized fields; these have never // been written to, and this must be the final write. Other writes @@ -768,34 +1069,100 @@ public class Parser { // The alloc() call added the allocation scope !_scope.canInit(def) ) throw error("Cannot reassign final '"+name+"'"); - checkSharedField(def); + checkSharedFunction(def); // Parse assignment expression - checkCaptureWrite(def); Node expr = parseAsgn(); // Lift expression, based on type - Node lift = liftExpr(expr.keep(), def.type(), true); + Type decl = def.type(); // Declared type + Node lift = convertExpr(expr.keep(), decl).keep(); // Update _scope.update(name,lift); + + // Check for storing into an instance field instead of a local var + int kx = _scope.kindx(def); // Declaration scope + int fx = _scope.enclosingFunction(); // Enclosing function scope + Kind kk = _scope._kinds.at(kx); + Kind fk = _scope._kinds.at(fx); + // Access instance field from 'self' + if( kx+1 == fx && kk instanceof Kind.Func inst && FunNode.isInstance(inst._name) && + fk instanceof Kind.Func method && !inExplicitConstructor() ) { + // Named field offset on an unknown struct type + TypeMemPtr self = (TypeMemPtr)_scope.var(fk._lexSize).type(); + TypeStruct sobj = (TypeStruct)TYPES.get(self._obj._name); + Field fld = sobj.field(name); + int alias = fld._alias; + Node ptr = _scope.in(method._lexSize); + Node off = fldoff(ptr,name); + Node prior = mem().keep(); + Node st = new StoreNode(loc(), name, alias, decl, ctrl(), prior, ptr, off, lift, inExplicitConstructor()).peephole().keep(); + storeMem(st,alias,prior); + st.unkeep(); + prior.unkeep(); + } + lift.unkill(); + + // Return un-lifted expr return expr.unkeep(); } - // Apply declared access permissions; widen floats and narrow integers. + // Widen ints to floats; narrow wide int types. // Early error if types do not match variable. - private Node liftExpr(Node expr, Type t, boolean isLoad) { - if( t==Type.BOTTOM || t==Type.TOP ) return expr; - // Constructor sentinels record missing initialization, not a value - // needing conversion. Keep them visible to the completion check. - if( expr instanceof ConstantNode && !(expr instanceof FRefNode) && - (expr._type==Type.TOP || expr._type==Type.BOTTOM) ) return expr; + private Node convertExpr( Node expr, Type t ) { + if( !expr._type.accessISA(t) ) + throw error("Type "+expr._type.str()+" is not of declared type "+t.str()); if( t instanceof TypeMemPtr tmp && tmp.isFinal() ) expr = peep(new ReadOnlyNode(expr)); - if( expr instanceof FunPtrNode && t instanceof TypeFunPtr ) return expr; + // A FunPtr carries the semantic edge to its function's Return. Do not + // hide that edge behind a Convert which can fold to a plain constant. + if( expr instanceof FunPtrNode && t instanceof TypeFunPtr ) + return expr; + // Auto-widen array to i64 (cast ptr to raw int bits) + if( t == TypeInteger.BOT && expr._type instanceof TypeMemPtr tmp && tmp._obj.isAry() ) + return peep(new AddNode(peep(new PtrToIntNode(expr)),off(tmp._obj,"[]"))); + // Widen ints to floats, narrow wide types return peep(new ConvertNode(t,expr)); } + // Inferred `var` declarations describe writable slots, not the exact + // value used to initialize them. Explicitly typed locals retain the + // user's nullability and numeric-width guarantees; `val` retains the + // initializer's precise type. + private static Type mutableInferredType(Type t) { + return switch( t ) { + case TypeInteger ti -> TypeInteger.BOT; + case TypeFloat tf -> TypeFloat.F64; + case TypeMemPtr tmp -> tmp.makeVar(); + case TypeFunPtr tfp -> tfp; + case TypeConAry a -> a; + case TypePtr ptr -> TypePtr.PTR; + case TypeScalar ts -> TypeScalar.BOT; + default -> t; + }; + } + + private Type constructorFieldType(Type t) { + if( t instanceof TypeMemPtr ptr && ptr._obj.isConAry() ) { + TypeStruct declared=(TypeStruct)TYPES.get(ptr._obj._name); + TypeStruct shape=typeAry(declared._fields[1]._t,ptr._obj._fields[1]._final); + return TypeMemPtr.make((byte)2,shape,false).withAccess(ptr._ro); + } + return mutableInferredType(t); + } + + // Seed an inferred return merge with the declared numeric family instead + // of a transient constant/range. Pointer nullability remains part of the + // inferred function result and is combined by the later Phi. + private static Type inferredReturnType(Type t) { + return switch( t ) { + case TypeInteger ti -> TypeInteger.BOT; + case TypeFloat tf -> TypeFloat.F64; + default -> t; + }; + } + private Node widenInt( Node expr, Type t ) { return (expr._type instanceof TypeInteger || expr._type==Type.NIL) && t instanceof TypeFloat ? peep(new ToFloatNode(expr)) : expr; @@ -822,23 +1189,6 @@ public class Parser { return require(n,";"); } - // Widen a reassignable binding without discarding the initializer's - // permissions. Lattice glb includes readonly memory and is too broad here. - private Type inferredType(Type t) { - if( t instanceof TypeMemPtr tmp ) { - if( tmp._obj.isAry() ) { - Type elem = tmp._obj._fields[1]._t; - if( elem instanceof TypeConAry ) - elem = ((TypeMemPtr)TYPES.get(tmp._obj._name))._obj._fields[1]._t; - return typeAry(elem,tmp._obj._fields[1]._final).withAccess(tmp._ro).makeNullable(); - } - Type base = TYPES.get(tmp._obj._name); - TypeMemPtr ptr = (TypeMemPtr)base; - return ptr.withAccess(tmp._ro).makeNullable(); - } - return t.glb(false); - } - // The binding modifier sits immediately before the name. It never // changes permissions on the referenced object. private boolean bindingFinal(Type t) { @@ -857,35 +1207,50 @@ public class Parser { Lexer loc = loc(); String name = requireId(); // Optional initializing expression follows - boolean fld_final = false; // Fixed field independent of pointee access - boolean initialized = match("="); + boolean fld_final = false; // Field is final, but not deeply final + boolean uninit = false; // Required non-null constructor field Node expr; - if( initialized ) { - expr = isExternDecl() ? externDecl(name,t) - : _scope.inConstructor() && name.equals(shortName(_nestedType)) - ? constructorFunc() : parseAsgn(); + if( match("=") ) { + if( isExternDecl() ) { + if( !(t instanceof TypeFunPtr) ) { + Node ptr = externData(name,t); + // The lexical binding is a fixed address; its contents are mutable. + if( !_scope.define(name,t,true,ptr,loc) ) + throw error("Redefining name '"+name+"'",loc); + _scope.lookup(name)._extern = _code.externAlias(name); + return _code.ZERO; + } + expr = externDecl(name,t); + t = expr._type; // Upgrade declared type to the exact extern decl type + } else if( isConstructorDecl(name) ) { + expr = constructorFunc(name); + } else { + expr = parseAsgn(); + } // var/val, then type comes from expression if( inferType ) { if( expr._type==Type.NIL ) throw error("a not-null/non-zero expression"); t = expr._type; - if( !xfinal ) t = inferredType(t); // Widen if not final - // A final field is still a local variable during construction. - // Its constant default is a value, not a one-value declaration - // type: later assignments and loop Phis may change it. - else if( _scope.inConstructor() && !(t instanceof TypeFunPtr) ) { - Type wide=inferredType(t); - t = t instanceof TypeMemPtr ptr && ptr.notNull() - ? TypeMemPtr.make((byte)2,((TypeMemPtr)wide)._obj,false).withAccess(ptr._ro) : wide; - } + if( !xfinal ) t = mutableInferredType(t); + // Instance fields may change until the constructor finishes. + // Keep the declared family/access, not the default's constant + // value or string contents. Class constants remain static. + else if( _scope.klast() instanceof Kind.Func fun && FunNode.isInstance(fun._name) && !(t instanceof TypeFunPtr) ) + t = constructorFieldType(t); } + // expr is a constant function - if( t instanceof TypeFunPtr && expr._type instanceof TypeFunPtr tfp && tfp.isConstant() ) { - if( expr instanceof ExternNode ) t = expr._type; // Upgrade declared type to exact function - else if( _code._link(tfp)!=null ) _code.link(tfp).setName(name); // Assign debug name to Simple function + if( t instanceof TypeFunPtr decl && expr._type instanceof TypeFunPtr tfp && tfp.isConstant() ) { + t = decl.makeFrom(tfp.fidx()); + FunNode fun = _code.link(tfp); + if( fun != null ) + fun.setName(name); // Assign debug name to Simple function } } else { + // Since no initializer, not-final but might need an initializer + // Need an expression to infer the type. // Also, if not-null then need an initializing expression. // Allowed in a class def, because type/init will happen in the constructor. @@ -893,281 +1258,393 @@ public class Parser { throw errorSyntax("=expression"); // Initial value for uninitialized struct fields. expr = switch( t ) { - // Nullable pointers get a NIL; not-null get a TOP which - // signals that they *must* be initialized in the constructor. - case TypeNil tn -> tn.nullable() ? NIL : con(Type.TOP); - case TypeInteger ti -> xfinal ? con(Type.TOP) : ZERO; - case TypeFloat tf -> con(xfinal ? Type.TOP : TypeFloat.FZERO); + // Required pointers retain their declared shape, but include + // null until every constructor path assigns them. + case TypeNil tn -> tn.nullable() ? con(Type.NIL) : con(t.meet(Type.NIL)); + case TypeInteger ti -> xfinal ? con(Type.BOTTOM) : _code.ZERO; + case TypeFloat tf -> con(xfinal ? Type.BOTTOM : TypeFloat.FZERO); // Bottom signals type inference: they must be initialized in // the constructor and that's when we'll discover the type. case Type tt -> { assert tt==Type.BOTTOM; yield con(tt); } }; - // Nullable fields are set in the constructor, but remain shallow final. - // e.g. final pointer to a r/w array - if( t instanceof TypeNil tn && !tn.nullable() && xfinal ) - fld_final = true; + // Every non-null pointer without a default is constructor-required. + // Without `!` it is also shallow-final; mutability and definite + // initialization are independent properties. + if( (xfinal && !(t instanceof TypeNil)) || (t instanceof TypeNil tn && !tn.nullable()) ) { + uninit = true; + if( xfinal ) fld_final = true; + } } // Lift expression, based on type - Node lift = liftExpr(expr, t, true); + // Do not let an authoritative Convert hide the nullable + // not-yet-initialized state from constructor-local type flow. + Node lift = uninit ? expr : convertExpr(expr, t); - // Lift type to the declaration. This will report as an error later if - // we cannot lift the type. // Define a new name if( !_scope.define(name,t,xfinal || fld_final,lift, loc) ) throw error("Redefining name '" + name + "'", loc); - _scope.lookup(name)._init = initialized; + _scope.lookup(name)._uninit = uninit; return lift; } - /** * Parse a struct declaration, and return the following statement. * Structs cannot be redefined, but can be nested. * * @return zero */ - private String _nestedType; - private static String shortName(String name) { return name.substring(name.lastIndexOf('.')+1); } + private Node parseStruct( ) { + // keyword "struct" already parsed, so expect the struct name next. + String typeName = requireId(); + require("{"); + + // create type class:NESTED.typeName; no fields. - private Node parseStruct() { String old = _nestedType; - String name = requireId(); - if( old!=null ) name=old+"."+name; - _nestedType=name; - TypeMemPtr ptr = TYPES.get(name) instanceof TypeMemPtr p ? p : TypeMemPtr.make(TypeStruct.open(name)); - TYPES.put(name,ptr); - require("{"); - TypeFunPtr sig = _code.makeFun(TypeFunPtr.make(false,false,new Type[]{ptr,TypeMem.BOT},TypeMem.BOT)); - ReturnNode init = parseFunctionBody(sig,loc(),null,true,"self","#selfMem"); - init.fun()._name=name+"."; - INITIALIZERS.put(name,init); - TypeStruct shape = ((TypeMemPtr)TYPES.get(name))._obj; - ALLOCS.put(name,makeAllocator(shape,init)); - _nestedType=old; + String fullName = _scope.depth()==1 && _nestedType.endsWith(typeName) ? old : (old+"."+typeName).intern(); + _nestedType = fullName; // Recursive structs start with this as their basename + + ReturnNode ret = parseStruct( false, fullName ); + + _nestedType = old; + + // Insert a field into class:NESTED.typeName of "val typeName = fcn-ptr-to-init" + // Function does malloc internally. + // Future self: means subtyping has to deal with "who does the malloc" + String clzName = addClzPrefix(fullName); + TypeFunPtr sig = ret.fun().sig(); + TypeStruct ts = (TypeStruct)TYPES.get(clzName); + String initName = initFieldName(fullName); + Field oldFld = ts==null ? null : ts.field(initName); + Field fld = oldFld==null + ? Field.make(initName,sig,_code.alias(_ref._cname),true) + : oldFld.makeFrom(sig); + Field ctorFld = ts==null ? null : ts.field(""); + // A same-named struct shares its file's class object and flag. + Field flag = ts==null ? null : ts.field("$init"); + ts = ctorFld==null + ? TypeStruct.make(clzName,false,fld) + : TypeStruct.make(clzName,false,fld,ctorFld); + if( flag!=null ) ts=ts.add(flag); + TypeFunPtr alloc = makeAllocator((TypeStruct)TYPES.get(fullName),ret,ctorFld); + ts = ts.add(Field.make("",alloc,_code.alias(_ref._cname),true)); + TYPES.put(clzName,ts); + + // Insert a field in the containing class with the nested class type. + // This is basically sugar for "val CLZNAME = " + // and is a nicer version of updateSelfAsFieldsDiscovered. + _scope.define(typeName, sig, true, new FunPtrNode(sig,_code._start,ret).peephole(), loc()); + require("}"); - require(";"); - return ZERO; + return require(_code.ZERO,";"); + } + + // Build the declaration-side allocation helper. Allocation sites only + // call this function; all layout, initialization and publication choices + // are made here, where the complete struct and constructor are known. + private TypeFunPtr makeAllocator(TypeStruct ts, ReturnNode initRet, Field ctorFld) { + TypeFunPtr ctor = ctorFld==null ? null : (TypeFunPtr)ctorFld._t; + int nargs = ctor==null ? 0 : ctor.nargs()-2; // Hide self and selfMem + Type[] atypes = new Type[nargs]; + for( int i=0; i").intern(); + + int oldUID = _code.UID(); + FunNode fun = (FunNode)new FunNode(loc(),sig,name,_ref,null,_ref._start).peephole(); + _code.link(fun); + Node rpc = new ParmNode("$rpc",0,TypeRPC.BOT,fun,ConstantNode.seed(TypeRPC.BOT).peephole()).peephole(); + Node pub = new ParmNode(ScopeNode.MEM0,1,TypeMem.BOT,fun, + new ProjNode(_ref._start,1,ScopeNode.MEM0).peephole()).peephole(); + Node[] parms = new Node[nargs]; + for( int i=0; i(); + // Record & restore global state set during parsing the code + Node oldCtrl= _scope.ctrl().keep(); + Node oldMem = _scope.mem ().keep(); + + // Make the future clazz/instance struct. + TypeStruct tself = (TypeStruct)TYPES.get(typeName); + if( tself==null ) + TYPES.put(typeName, tself = TypeStruct.make(typeName,true)); + else { + assert tself._open; + TYPES.put(typeName,tself = tself.defined()); + } + + // Class object keeps a runtime flag to signal the has run. + if( isClz ) + TYPES.put(typeName,tself=tself.add(Field.make("$init",TypeInteger.BOOL,_code.alias(_ref._cname),false))); + + // Signature of the default constructor + TypeFunPtr sig = structInitSig(isClz,tself); + // Function and argument names of default constructor + String fname = (typeName + (isClz ? "." : ".")).intern(); + String[] ids = new String[]{"self", isClz ? "arg" : "#selfMem"}; + + // Struct decls look like function bodies. Parse function body normally. + ReturnNode ret = _parseFunctionBody(fname,sig,loc(),ids); + _ctorExits = oldExits; + // Unwind global state. + ctrl(oldCtrl.unkeep()); + mem (oldMem .unkeep()); + return ret; } - private void finishFields() { - TypeStruct shape = ((TypeMemPtr)TYPES.get(_nestedType))._obj.close(); - TYPES.put(_nestedType,TypeMemPtr.make(shape)); - StructNode constants = new StructNode(shape); - int first = _scope._kinds.last()._lexSize; - for( int i=first; i<_scope.nIns(); i++ ) - if( !_scope.var(i).isFRef() ) constants.addDef(_scope.in(i)); - constants.setType(constants.compute()); - INITS.put(_nestedType,constants.keep()); + // signature: { self arg/selfMem -> BOT/selfMem } + // Self-memory is the rare *private* (never aliased) memory for the new object. + private TypeFunPtr structInitSig( boolean isClz, TypeStruct tself ) { + // Private uninitialized memory from new object, no escaped values. + TypeMem privMem = isClz ? null : TypeMem.make(1,tself,true,isClz,false,null,null); + Type targ = isClz ? TypeInteger.BOT : privMem; + // Return for private memory has to name all escaped fields, which are not known yet. + // Escape ALL fields, and sharpen later. + Type tret = isClz ? Type.BOTTOM : TypeMem.makePrivate(tself); + Type[] targs = new Type[]{ TypeMemPtr.make((byte)2,tself,isClz), targ }; + // FIDX for FREFs is in the global namespace, needs the correct clz name. + String clzBare = tself._name; + int order = 3; // 3 is always the order, 2nd in the file just behind the + if( isClz ) { + assert startsClzPrefix( clzBare ); + clzBare = clzBare.substring(clzPrefix.length()).intern(); + order = 2; // 2 is always the order, first in file and skipping reserved #1 + } + return TypeFunPtr.make1((byte)2, true, targs, tret, _code.fidx(clzBare,order)); } - private Node self() { return _scope.in(_scope.lookup("self")); } - private Node privateMem() { return _scope.in(_scope.lookup("#selfMem")); } - - private boolean privateSelf(Node ptr) { - if( _privateFun!=_fun || _privateFun==null ) return false; - while( ptr instanceof ConvertNode || ptr instanceof ReadOnlyNode || ptr instanceof GuardNode ) ptr=ptr.in(1); - Node self=self(); - while( self instanceof ConvertNode || self instanceof ReadOnlyNode || self instanceof GuardNode ) self=self.in(1); - return ptr==self; + // A class type carries the constructor function as a required final field. + private TypeStruct classStruct( String typeName, TypeFunPtr init, boolean open ) { + return TypeStruct.make(addClzPrefix(typeName),open,Field.make(initFieldName(typeName),init,_code.alias(_ref._cname),true)); } - private boolean mayBeSelf(Node ptr, java.util.BitSet seen) { - if( _privateFun==null || _privateFun!=_fun || seen.get(ptr._nid) ) return false; - seen.set(ptr._nid); - if( privateSelf(ptr) ) return true; - if( ptr instanceof ConvertNode || ptr instanceof ReadOnlyNode || ptr instanceof GuardNode ) - return mayBeSelf(ptr.in(1),seen); - if( ptr instanceof PhiNode && !(ptr instanceof ParmNode) ) - for( int i=1; i types = new Ary<>(Type.class); - Ary names = new Ary<>(String.class); - types.add(TYPES.get(_nestedType)); names.add("self"); - types.add(TypeMem.BOT); names.add("#selfMem"); - java.util.BitSet finals = new java.util.BitSet(); - finals.set(0); // Rebinding self would detach the pointer from its private memory. - while( true ) { - Type t=type(); - if( t==null ) break; - finals.set(names.size(),bindingFinal(t)); - types.add(t); names.add(requireId()); - match(","); - } - require("->"); - TypeFunPtr sig = _code.makeFun(TypeFunPtr.make(false,false,types.asAry(),TypeMem.BOT)); - // Bind the signature now, but parse the body after the field list is - // closed. Then even a later field has a local SSA slot before branches - // and loops are parsed. Strings/comments must not affect brace depth. - _constructor=new Constructor(sig,loc(),pos(),names.asAry(),finals); - int depth=1; - while( depth>0 ) { - if( peek('"') ) { parseString(); continue; } - if( match("'") ) { parseChar(); continue; } - char ch=_lexer.nextXChar(); - if( ch==Character.MAX_VALUE ) throw errorSyntax("}"); - if( ch=='{' ) depth++; - if( ch=='}' ) depth--; - } - return con(sig); + + private TypeStruct latestStruct( TypeStruct ts ) { + Type t = TYPES.get(ts._name); + return t instanceof TypeStruct ts2 ? ts2 : ts; } - private CallEndNode rawCall(Node ctrl, Node pub, ReturnNode target, Node... args) { - Node[] inputs = new Node[args.length+3]; - inputs[0]=ctrl; inputs[1]=pub; - System.arraycopy(args,0,inputs,2,args.length); - inputs[inputs.length-1]=new FunPtrNode(target.fun().sig(),target).peephole(); - // Keep the complete declaration-side caller set before any inlining. - CallNode call = new CallNode(loc(),inputs).init(); - CallEndNode end = new CallEndNode(call).init(); - call.link(target.fun()); - return end; + private Field fieldOrOpen( TypeStruct ts, String name, boolean xfinal ) { + ts = latestStruct(ts); + Field fld = ts.field(name); + if( fld != null ) return fld; + if( !ts._open || !ts._fref ) return null; + fld = Field.make(name,Type.BOTTOM,_code.alias(_ref._cname),xfinal); + TYPES.put(ts._name,ts = ts.add(fld)); + if( _ctorOpenStruct != null && _ctorOpenStruct._name == ts._name ) + _ctorOpenStruct = ts; + return fld; } - private Node publish(TypeStruct shape, Node ctrl, Node pub, Node priv, Node ptr) { - priv.keep(); ptr.keep(); - for( Field f : shape._fields ) if( !f._one ) { - pub.keep(); - Node esc = new EscapeNode(loc(),f,ctrl,pub,priv,ptr).peephole(); - Node next = new MemMergeNode(pub,f._alias,esc).peephole(); - pub.unkeep(); pub=next; - } - ptr.unkeep(); priv.unkill(); - return pub; - } - - private TypeFunPtr makeAllocator(TypeStruct shape, ReturnNode init) { - Field cf=shape.field(shortName(shape._name)); - TypeFunPtr ctor=cf!=null && cf._t instanceof TypeFunPtr f ? f : null; - int nargs=ctor==null ? 0 : ctor.nargs()-2; - Ary captures=CAPTURES.get(shape._name); - Type[] types=new Type[nargs+captures.size()]; - for( int i=0; i"; fun._constructor=true; - _code.link(fun); - Node rpc=new ParmNode("$rpc",0,TypeRPC.BOT,fun,con(TypeRPC.BOT)).init(); - Node pub=new ParmNode(ScopeNode.MEM0,1,TypeMem.BOT,fun,new ProjNode(_code._start,1,ScopeNode.MEM0).peephole()).init(); - Node[] args=new Node[nargs+2]; - for( int i=0; i")) || id==null ) + if( !(peek(',') || peek(';') || peek('=') || match("->")) || id==null ) return posT(old1); // Reset lexer to reparse pos(old2); // Reset lexer to reparse // Yes a forward ref, so declare it - TYPES.put(tname,((TypeMemPtr)t1).withAccess(false)); + forwardRefType(tname,(TypeStruct)((TypeMemPtr)t1)._obj); // Return the (array, final) type return t2; } - // Make an array type of t - private TypeMemPtr typeAry( Type t, boolean efinal ) { - if( t instanceof TypeMemPtr tmp && tmp.notNull() ) - throw error("Arrays of reference types must always be nullable"); - String tname = "[]"+t.str(); - TypeMemPtr ta = (TypeMemPtr)TYPES.get(tname); - if( ta==null ) - // Remember final version - TYPES.put(tname, ta = TypeMemPtr.make( TypeStruct.makeAry(tname,TypeInteger.U32,_code.getALIAS(),t,_code.getALIAS(),true) )); - // Reuse aliases, but preserve this occurrence's element permissions. - TypeStruct ary = ta._obj; - return TypeMemPtr.make(TypeStruct.makeAry(tname,TypeInteger.U32,ary._fields[0]._alias,t,ary._fields[1]._alias,efinal)); + private TypeMemPtr forwardRefType( String tname, TypeStruct tself ) { + TYPES.put(tname,tself); + String clzName = addClzPrefix( tname ); + TYPES.put(clzName, classStruct(tname,structInitSig(false,tself),true)); + return TypeMemPtr.make(tself); + } + + private TypeMemPtr forwardRefType( String tname ) { + tname = fullTypeName(tname); + return forwardRefType(tname,TypeStruct.forward(tname)); + } + + private String fullTypeName( String tname ) { + String nest = _nestedType, fullq; + while( true ) { + fullq = nest+"."+tname; + if( TYPES.get(fullq) != null ) + return fullq.intern(); + int idx = nest.lastIndexOf('.'); + if( idx== -1 ) return fullq.intern(); + nest = nest.substring(0,idx); + } + } + + // Make an array type of t. Always record a mutable version, + // but return the requested version. + private TypeStruct typeAry( Type t, boolean efinal ) { + // Non-null reference arrays are valid types, even though an ordinary + // zero-filled `new T[len]` cannot construct one. Allocation retains + // that check in alloc(); trusted runtimes (notably argv construction) + // may fill the private array completely before publishing it. + String tname = ("[]"+t.str()).intern(); + TypeStruct ta = (TypeStruct)TYPES.get(tname); + if( ta==null ) { + // Primitive arrays use the null compunit + String rcname = TYPES.get(t) instanceof TypeStruct ? _ref._cname : ("_."+t).intern(); + ta = TypeStruct.makeAry(tname,TypeInteger.U32,_code.alias(rcname),t,_code.alias(rcname), false ); + TYPES.put(tname,ta); + } + // Already have the aliases, just efinal is wrong + return TypeStruct.makeAry(tname,TypeInteger.U32,ta._fields[0]._alias,t,ta._fields[1]._alias,efinal ); + } + + private Type implicitSelfType() { + TypeStruct self = _scope.constructorSelf(); + // Ordinary instance methods accept the declared object shape. The + // constructor-only receiver starts at the low/default field values, + // and is built explicitly by constructorFunc(). + return self == null ? TypePtr.PTR : TypeMemPtr.make(self); } // A function type is `{ type... -> type }` or `{ type }`. @@ -1227,16 +1742,19 @@ public class Parser { match("{"); // Skip already-peeked '{' Type t0 = type(); // Either return or first arg if( t0==null ) return posT(old); // Not a function - if( match("}") ) // No-arg function { -> type } - return TypeFunPtr.make(match("?"),false,TypeFunPtr.TEMPTY,t0); Ary ts = new Ary<>(Type.class); + // Always reserve a receiver slot. Outside constructor scopes it is + // broad and unused, so call sites can uniformly pass their self. + ts.push(implicitSelfType()); + if( match("}") ) // No-arg function { -> type } + return TypeFunPtr.make(match("?"),ts.asAry(),t0); ts.push(t0); // First argument while( true ) { if( match("->") ) { // End of arguments, parse return Type ret = type(); if( ret==null || !match("}") ) return posT(old); // Not a function - return TypeFunPtr.make(match("?"),false,ts.asAry(),ret); + return TypeFunPtr.make(match("?"),ts.asAry(),ret); } Type t1 = type(); if( t1==null ) return posT(old); // Not a function @@ -1300,7 +1818,6 @@ public class Parser { } - /** * Parse an eq/ne expression */ @@ -1475,15 +1992,17 @@ public class Parser { // Pre-dec/pre-inc int old = pos(); if( match("--") || match("++") ) { - int delta = _lexer.peek(-1)=='+' ? 1 : -1; // Pre vs post + int delta = _lexer.peek(-1)=='+' ? 1 : -1; // Inc vs Dec String name = _lexer.matchId(); if( name!=null ) { Var n = _scope.lookup(name); if( n != null && !(n.type() instanceof TypeMemPtr) ) { + if( n._extern!=0 ) + return parsePostfixName(_scope.in(n),name,delta); if( n._final && !_scope.canInit(n) ) throw error("Cannot reassign final '"+n._name+"'"); - checkSharedField(n); - Node expr = liftExpr(peep(new AddNode(_scope.in(n),con(delta))),n.type(),true); + checkSharedFunction(n); + Node expr = convertExpr(peep(new AddNode(_scope.in(n),con(delta))),n.type()); _scope.update(n,expr); return expr; } @@ -1503,151 +2022,125 @@ public class Parser { * 
          *     primaryExpr : integerLiteral | "string" | 'char' | true | false | null |
      *                   new Type | '(' expression ')' | Id['++','--'] |
-     *                   [static.]* Id
+     *                   type['.' Id]*
      *
    * @return a primary {@link Node}, never {@code null} */ private Node parsePrimary() { if( _lexer.isNumber(_lexer.peek()) ) return parseLiteral(); - if( _lexer.peek('"') ) return newString(parseString()); + if( _lexer.peek('"') ) return parsePostfix(newString(parseString())); if( matchx("true" ) ) return con(1); - if( matchx("false") ) return ZERO; - if( matchx("null" ) ) return NIL; + if( matchx("false") ) return _code.ZERO; + if( matchx("null" ) ) return con(Type.NIL); if( match ("'" ) ) return parseChar(); if( match ("(" ) ) return parsePostfix(require(parseAsgn(), ")")); if( matchx("new" ) ) return parsePostfix(alloc()); if( match ("{" ) ) return parsePostfix(require(func(),"}")); - // Parse static variable lookup: `type.fld` in the type namespace - // CNC: not a full type parse, just a typename parse int pos = pos(); - Type t = type(); - if( t!=null ) { - if( peek('.') ) - // A namespace lookup is not a reference declaration. - return parsePostfix(con(TYPES.get(((TypeMemPtr)t)._obj._name))); - //pos(pos); // TODO: not a `type.fld` - //return null; - throw Utils.TODO(); - } - // Expect an identifier now - - // - if null, make fref, normal var handling - // - if final constant, normal var handling - // - if method var, scan skipping blocks & methods, must hit correct struct no fcns - // - if normal var, scan skipping blocks, allowing last fcn - // - NO fcns nested in methods, NO methods nested in fcns - String id = _lexer.matchId(); if( id == null || KEYWORDS.contains(id) ) throw errorSyntax("an identifier or expression"); + int pos2 = pos(); + // Attempt a local var lookup first Var var = _scope.lookup(id); - if( var==null ) { - // If missing, assume a forward reference - _scope.define(id, FRefNode.FREF_TYPE, true, XCTRL, loc()); - var = _scope.lookup(id); - } + // Not a local var, try again as a type name + pos(pos); + Type t = type(); + if( t!=null ) { + // Static typename lookup + if( peek('.') ) // type.fld, which is really a field lookup in the class object + return parsePostfix(con(t)); + // Direct reference to a type. Convert the instance + // type to a class type name + if( !(t instanceof TypeMemPtr tmp) ) { + //return new FRefNode(addClzPrefix(tmp._obj._name), loc()).peephole(); + // Not a local var, but yes a bare type... disallow a local + // FREF and return null, not-a-primary parse. + pos(pos); + return null; + } + } - // If a method var, allow nested blocks and methods, but not fcns or other structs. - // Else safety check for accessing fields out of function scope. - if( instance(var) ) { - Var selfVar = _scope.lookup("self"); - Node self = liftExpr(_scope.in(selfVar),selfVar.type(),true); - return parsePostfixName(self,id); + // Insert FREF outside enclosing Kind.Func scope. + pos(pos2); + t = FRefNode.FREF_TYPE; + // Check if this is a class. + CompUnit cu = ParseAll.findCompUnit(_code,_ref,id); + if( cu!=null ) { + // Must be a final class-type + String clzName = addClzPrefix(cu._cname); + TypeStruct clz = (TypeStruct)TYPES.get(clzName); + if( clz == null ) + TYPES.put(clzName,clz = TypeStruct.make(clzName,true)); + t = TypeMemPtr.make((byte)2,clz,true); + } + // Define the FRef + var = _scope.defineFRef(id,t,true,loc()); + _frefs.push((FRefNode)_scope.in(var)); } // Load local value - Node oldValue = _scope.in(var); - if( oldValue instanceof ConstantNode && !(oldValue instanceof FRefNode) && oldValue._type==Type.BOTTOM ) - throw error("Cannot read uninitialized field '"+id+"'"); - Node rvalue = liftExpr(oldValue,var.type(),true); - if( rvalue!=oldValue ) _scope.update(var,rvalue); - if( rvalue instanceof ConstantNode && rvalue._type == Type.BOTTOM ) - if( rvalue instanceof FRefNode ) - return parsePostfix(rvalue); - else throw error("Cannot read uninitialized field '"+id+"'"); + Node rvalue = _scope.in(var); + if( var._extern!=0 ) + return parsePostfixName(rvalue,id,0); + if( var._fref ) + return parsePostfix(rvalue); + // A constructor may read the unfinished object, but each field it + // reads must already be initialized on every path. Read the current + // local value; other fields may still be waiting for their first write. + if( inExplicitConstructor() && var._uninit && missingInit(rvalue,new BitSet()) ) + throw error("Field '"+id+"' might not be initialized",var._loc); + + // Check for an instance field load + int kx = _scope.kindx(var); // Declaration scope + int fx = _scope.enclosingFunction(); // Enclosing function scope + Kind kk = _scope._kinds.at(kx); + Kind fk = _scope._kinds.at(fx); + // Access instance field from 'self' + if( kx+1 == fx && kk instanceof Kind.Func inst && FunNode.isInstance(inst._name) && fk instanceof Kind.Func method && !inExplicitConstructor() ) + return parsePostfixName(_scope.in(method._lexSize),id,0); + + // Check for a function-escaping variable; these require true + // closures. Final constants are OK; final vars require a hidden var + // argument; not-final fields can just use an explicit struct arg. + if( kx < fx && !inExplicitConstructor() && // Out of scope + !(var._final && rvalue._type.isConstant()) ) { // Not a final constant + if( kx > 0 ) // Global scope is OK, only ever one of these (not one per function invoke) + throw error("Variable '"+var._name+"' is out of function scope and must be a final constant"); + + // var.idx==2 is self argument, 3 is args, 4 or more are now static globals + if( var._idx >= 4 ) { + // Read from global scope. + // Surely a nicer way to get the top-level scope compile name? + Var clzptr = _scope.var( 2 ); + // Get the most recent sharpen global + TypeStruct clz = (TypeStruct) TYPES.get( ((TypeMemPtr) clzptr.type())._obj._name ); + TypeMemPtr clztmp = TypeMemPtr.make( (byte) 2, clz, true ); + return parsePostfixName( con( clztmp ), id,0 ); + //return parsePostfixName(_scope.in(2),id); + } + } // Check for assign-update, x += e0; char ch = _lexer.matchOperAssign(); - if( ch==0 ) // Normal primary, check for postfix updates - return parsePostfix(rvalue); + if( ch==0 ) { // Normal primary, check for postfix updates + int selfx = _scope.kindFcn(var)._lexSize; + // Might be a method call from inside a method, so no explicit "self". + // Pass the in-scope "self" + return parsePostfixMethod(rvalue,_scope.in(_scope.update(_scope.var(selfx),null))); + } // Assign-update direct into Scope if( var._final && !_scope.canInit(var) ) throw error("Cannot reassign final '"+var._name+"'"); - checkSharedField(var); - checkCaptureWrite(var); + checkSharedFunction(var); Node op = opAssign(ch,rvalue, var.type() ); _scope.update(var,op); return postfix(ch) ? rvalue.unkeep() : op; } - // true if instance variable lookup (var is in any number of nested blocks, - // then an instance function then inside a struct declaration). - private boolean instance(Var var ) { - // If a method var, allow nested blocks and methods, but not fcns or other structs - Kind kind = _scope.kind(var); - if( kind instanceof Kind.Define ) { - if( var.isFRef() ) return false; // Promote an unresolved lexical name, not self.name. - if( _privateFun==_fun && _privateFun!=null ) return true; - for( int i = _scope.depth()-1; i>=0 && var._idx < _scope._kinds.at(i)._lexSize; i-- ) { - Kind xkind = _scope._kinds.at(i); - if( xkind instanceof Kind.Block ) continue; - if( xkind instanceof Kind.Func ) { - if( _scope._kinds.at(i-1)==kind ) - return true; // Instance variable: lookup in a method in a decl - } - throw Utils.TODO(); - } - } else { - // If a normal var, allow any nested blocks, and a last fcn - - // but not functions further out (which involves capture). - Node def = _scope.in(var._idx); - if( _captures!=null && _scope.kind(var) instanceof Kind.Func && - !(var._final && def._type.isConstant()) && !(def instanceof FRefNode) ) { - if( !(def instanceof ParmNode parm && parm.fun()==_fun) ) { - _captures.add(def.keep()); - Type type=def._type; - int index=_fun.sig().nargs(); - _fun.addParseArg(type); - _scope.update(var,new ParmNode(var._name,index+2,type,_fun,con(type)).init()); - } - return false; - } - for( int i = _scope.depth()-1; i>=0 && var._idx < _scope._kinds.at(i)._lexSize; i-- ) - if( _scope._kinds.at(i) instanceof Kind.Func && - !(def instanceof FRefNode) && !(var._final && def._type.isConstant()) ) - throw error("Variable '"+var._name+"' is out of function scope and must be a final constant"); - } - return false; - } - - private void checkCaptureWrite(Var var) { - if( _captures==null ) return; - for( int i=_scope.depth()-1; i>=0 && var._idx<_scope._kinds.at(i)._lexSize; i-- ) - if( _scope._kinds.at(i) instanceof Kind.Func ) - throw error("Cannot assign captured variable '"+var._name+"' in an initializer"); - } - - // Inline constructors have local field slots too, but shared function - // declarations have no instance slot to publish. Do not silently discard - // an assignment to one of those declarations. - private void checkSharedField(Var var) { - if( !_scope.canInit(var) ) return; - if( _scope.kind(var) instanceof Kind.Define ) { - checkPrivateWrite(var._name); - return; - } - for( int i=_scope.depth()-1; i>0; i-- ) - if( _scope._kinds.at(i) instanceof Kind.Alloc alloc ) { - Field field=alloc._tmp._obj.field(var._name); - if( field!=null && field._one ) - throw error("Cannot reassign shared function '"+var._name+"'"); - return; - } - } - // Check for assign-update, "x += e0". // Returns "x + e0"; the caller assigns value. // if ch is postfix, then lhs is kept and caller will unkeep and return it. @@ -1670,17 +2163,17 @@ public class Parser { case '-' -> new SubNode(lhs,rhs); case '*' -> new MulNode(lhs,rhs); case '/' -> new DivNode(lhs,rhs); - case '&' -> new AndNode(loc(),lhs,rhs); - case '|' -> new OrNode(loc(),lhs,rhs); + case '|' -> new OrNode(null,lhs,rhs); + case '&' -> new AndNode(null,lhs,rhs); case '^' -> new XorNode(loc(),lhs,rhs); case Lexer.SHL -> new ShlNode(loc(),lhs,rhs); case Lexer.SAR -> new SarNode(loc(),lhs,rhs); case Lexer.SHR -> new ShrNode(loc(),lhs,rhs); - default -> throw Utils.TODO(); + default -> throw TODO(); }; // Convert to float ops, or narrow int types; error if not declared type. // Also, if postfix LHS is still keep() - return liftExpr(peep(op),t,true); + return convertExpr(peep(op),t); } @@ -1689,161 +2182,119 @@ public class Parser { */ private Node alloc() { Type t = type(); - if( t==null ) throw error("Expected a type"); + if( t==null ) { + String tname = _lexer.matchId(); + if( tname==null || KEYWORDS.contains(tname) ) + throw error("Expected a type"); + t = forwardRefType(tname); + } // Parse ary[ length_expr ] if( match("[") ) { if( !t.makeZero().isa(t) ) - throw error("Cannot allocate a non-nullable, since arrays always zero/null fill"); - Node len = parseAsgn(); - if( !(len._type instanceof TypeInteger) ) - throw error("Cannot allocate an array with length "+len._type); + throw error("Cannot allocate a non-nullable, since arrays are always zero/null filled"); + Node len = parseAsgn().keep(); require("]"); - TypeMemPtr tmp = typeAry(t,false); - return newArray(tmp._obj,len); - } - - if( !(t instanceof TypeMemPtr tmp) ) - throw error("Cannot allocate a "+t.str()); - - TypeStruct shape=tmp._obj; - TypeFunPtr sig=ALLOCS.get(shape._name); - if( sig==null ) throw error("Unknown struct type '"+shape._name+"'"); - if( !peek('{') ) { - // Obvious missing defaults can still be diagnosed here. A user - // constructor or a merge needs the later publication check. - if( shape.field(shortName(shape._name))==null ) { - StructNode defaults=INITS.get(shape._name); - Type initial=INITIALIZERS.get(shape._name).expr()._type; - for( int i=0; i args=new Ary<>(Node.class); - if( match("(") ) { - while( !peek(')') ) { - Node arg=parseAsgn(); - checkPublication(arg); - args.add(arg.keep()); - if( !match(",") ) break; - } - require(")"); - } - for( Node capture : CAPTURES.get(shape._name) ) { - checkPublication(capture); - args.add(capture.keep()); - } - if( args.size()!=sig.nargs() ) throw error("Wrong number of constructor arguments for '"+shape._name+"'"); - for( Node arg : args ) arg.unkeep(); - return callWithArgs(fptr,args.asAry()); - } - require("{"); - Node alloc=new NewNode(TypeMemPtr.make(shape),ctrl(),null,off(shape," len")).init(); - Node ptr=new ProjNode(alloc,0,shape._name).init().keep(); - Node priv=new ProjNode(alloc,1,"#selfMem").init(); - Ary captures=CAPTURES.get(shape._name); - Node[] args=new Node[captures.size()+2]; - args[0]=ptr; args[1]=priv; - for( int i=0; i ctorArgs = constructorArgs(); + // A field declaration can leave a short forward type name in TYPES. + // Allocation happens in lexical context, so resolve that short name + // to the completed nested type before selecting its class helper. + String typeName = ts._name.indexOf('.') == -1 ? fullTypeName(ts._name) : ts._name; + Type resolved = TYPES.get(typeName); + if( resolved instanceof TypeStruct rts ) ts = rts; + TypeStruct clz = (TypeStruct)TYPES.get(addClzPrefix(typeName)); + if( clz==null ) + throw error("Unknown struct type '" + ts._name + "'"); + Field afld = clz.field(""); + Node alloc; + if( afld != null ) { + TypeFunPtr tfp = (TypeFunPtr)afld._t; + FunNode fun = _code.link(tfp); + alloc = fun==null + ? new FRefNode((addClzPrefix(typeName)+".").intern(),loc()) + : new FunPtrNode(tfp,_code._start,fun.ret()).peephole(); + } else { + if( !ts._fref ) + throw error("Constructor arguments for '" + ts._name + "' but no constructor is defined"); + Type[] sig = new Type[ctorArgs._len]; + for( int i=0; i").intern(),loc(), + TypeFunPtr.make(false,sig,TypeMemPtr.make(ts))); + } + Ary args = new Ary<>(Node.class); + args.add(ctrl()); args.add(mem()); + for( Node arg : ctorArgs ) args.add(arg.unkeep()); + args.add(alloc); + return functionCall(args); + } + + private Ary constructorArgs() { + Ary args = new Ary<>(Node.class); + if( !match("(") ) + return args; + while( !peek(')') ) { + Node arg = parseAsgn(); + if( arg==null ) break; + args.push(arg.keep()); + if( !match(",") ) break; } - _scope.pop(); - mem(publish(shape,ctrl(),mem(),priv.unkeep(),ptr)); - return ptr.unkeep(); - } - - // The old constructor model requires each field on every path. An - // unresolved expression also has BOTTOM now, so inspect the actual - // sentinel and merges instead of mistaking an unknown type for no write. - private boolean missingInit(Node n, java.util.BitSet visit) { - if( n==null || visit.get(n._nid) || n instanceof FRefNode ) return false; - visit.set(n._nid); - if( n instanceof ConstantNode ) return n._type==Type.BOTTOM || n._type==Type.TOP; - if( n instanceof ConvertNode || n instanceof ReadOnlyNode ) return missingInit(n.in(1),visit); - if( n instanceof PhiNode phi ) - for( int i=1; i