Merge pull request #5981 from MartinNowak/issue18114

fix Issue 18114 - regex performance regression
merged-on-behalf-of: Martin Nowak <code@dawg.eu>

std/regex/internal/backtracking.d

@@ -825,7 +825,7 @@ struct CtContext
         int addr;
     }
 
-    this(Char)(const Regex!Char re)
+    this(Char)(ref const Regex!Char re)
     {
         match = 1;
         reserved = 1; //first match is skipped

std/regex/internal/ir.d

@@ -434,7 +434,7 @@ struct Group(DataIndex)
 interface MatcherFactory(Char)
 {
 @safe:
-    Matcher!Char create(const Regex!Char, in Char[] input) const;
+    Matcher!Char create(const ref Regex!Char, in Char[] input) const;
     Matcher!Char dup(Matcher!Char m, in Char[] input) const;
     size_t incRef(Matcher!Char m) const;
     size_t decRef(Matcher!Char m) const;
@@ -448,9 +448,9 @@ abstract class GenericFactory(alias EngineType, Char) : MatcherFactory!Char
     // round up to next multiple of size_t for alignment purposes
     enum classSize = (__traits(classInstanceSize, EngineType!Char) + size_t.sizeof - 1) & ~(size_t.sizeof - 1);
 
-    Matcher!Char construct(const Regex!Char re, in Char[] input, void[] memory) const;
+    Matcher!Char construct(const ref Regex!Char re, in Char[] input, void[] memory) const;
 
-    override Matcher!Char create(const Regex!Char re, in Char[] input) const @trusted
+    override Matcher!Char create(const ref Regex!Char re, in Char[] input) const @trusted
     {
         immutable size = EngineType!Char.initialMemory(re) + classSize;
         auto memory = enforce(malloc(size), "malloc failed")[0 .. size];
@@ -496,7 +496,7 @@ abstract class GenericFactory(alias EngineType, Char) : MatcherFactory!Char
 // A factory for run-time engines
 class RuntimeFactory(alias EngineType, Char) : GenericFactory!(EngineType, Char)
 {
-    override EngineType!Char construct(const Regex!Char re, in Char[] input, void[] memory) const
+    override EngineType!Char construct(const ref Regex!Char re, in Char[] input, void[] memory) const
     {
         import std.conv : emplace;
         return emplace!(EngineType!Char)(memory[0 .. classSize],
@@ -507,7 +507,7 @@ class RuntimeFactory(alias EngineType, Char) : GenericFactory!(EngineType, Char)
 // A factory for compile-time engine
 class CtfeFactory(alias EngineType, Char, alias func) : GenericFactory!(EngineType, Char)
 {
-    override EngineType!Char construct(const Regex!Char re, in Char[] input, void[] memory) const
+    override EngineType!Char construct(const ref Regex!Char re, in Char[] input, void[] memory) const
     {
         import std.conv : emplace;
         return emplace!(EngineType!Char)(memory[0 .. classSize],
@@ -518,7 +518,7 @@ class CtfeFactory(alias EngineType, Char, alias func) : GenericFactory!(EngineTy
 // A workaround for R-T enum re = regex(...)
 template defaultFactory(Char)
 {
-    @property MatcherFactory!Char defaultFactory(const Regex!Char re)
+    @property MatcherFactory!Char defaultFactory(const ref Regex!Char re) @safe
     {
         import std.regex.internal.backtracking : BacktrackingMatcher;
         import std.regex.internal.thompson : ThompsonMatcher;

std/regex/internal/thompson.d

@@ -847,7 +847,7 @@ final:
         }
     }
 
-    this()(const Regex!Char program, Stream stream, void[] memory)
+    this()(ref const Regex!Char program, Stream stream, void[] memory)
     {
          // We are emplace'd to malloced memory w/o blitting T.init over it\
          // make sure we initialize all fields explicitly

std/regex/package.d

@@ -441,12 +441,22 @@ template ctRegexImpl(alias pattern, string flags=[])
     {
         // allow code that expects mutable Regex to still work
         // we stay "logically const"
-        @trusted @property auto getRe() const { return cast() staticRe; }
+        @property @trusted ref getRe() const { return *cast(Regex!Char*)&staticRe; }
         alias getRe this;
     }
     enum wrapper = Wrapper();
 }
 
+@safe unittest
+{
+    // test compat for logical const workaround
+    static void test(StaticRegex!char)
+    {
+    }
+    enum re = ctRegex!``;
+    test(re);
+}
+
 /++
     Compile regular expression using CTFE
     and generate optimized native machine code for matching it.
@@ -476,18 +486,18 @@ if (isSomeString!R)
     alias String = R;
 private:
     import std.conv : text;
-    R _input;
-    int _nMatch;
     enum smallString = 3;
     enum SMALL_MASK = 0x8000_0000, REF_MASK= 0x1FFF_FFFF;
     union
     {
         Group!DataIndex[] big_matches;
         Group!DataIndex[smallString] small_matches;
     }
+    const(NamedGroup)[] _names;
+    R _input;
+    int _nMatch;
     uint _f, _b;
     uint _refcount; // ref count or SMALL MASK + num groups
-    const(NamedGroup)[] _names;
 
     this(R input, uint n, const(NamedGroup)[] named)
     {
@@ -671,6 +681,17 @@ public:
 
     ///A hook for compatibility with original std.regex.
     @property ref captures(){ return this; }
+
+    void opAssign()(auto ref Captures rhs)
+    {
+        if (rhs._refcount & SMALL_MASK)
+            small_matches[0 .. rhs._refcount & 0xFF] = rhs.small_matches[0 .. rhs._refcount & 0xFF];
+        else
+            big_matches = rhs.big_matches;
+        assert(&this.tupleof[0] is &big_matches);
+        assert(&this.tupleof[1] is &small_matches);
+        this.tupleof[2 .. $] = rhs.tupleof[2 .. $];
+    }
 }
 
 ///
@@ -802,7 +823,7 @@ public:
     @property inout(Captures!R) captures() inout { return _captures; }
 }
 
-private @trusted auto matchOnce(RegEx, R)(R input, const RegEx prog)
+private @trusted auto matchOnce(RegEx, R)(R input, const auto ref RegEx prog)
 {
     alias Char = BasicElementOf!R;
     auto factory = prog.factory is null ? defaultFactory!Char(prog) : prog.factory;
@@ -813,7 +834,7 @@ private @trusted auto matchOnce(RegEx, R)(R input, const RegEx prog)
     return captures;
 }
 
-private auto matchMany(RegEx, R)(R input, RegEx re) @safe
+private auto matchMany(RegEx, R)(R input, auto ref RegEx re) @safe
 {
     return RegexMatch!R(input, re.withFlags(re.flags | RegexOption.global));
 }