structural regular expressions prototype

1 files changed, 93 insertions, 0 deletions

diff --git a/sys/libsre/sre.h b/sys/libsre/sre.h
new file mode 100644
index 0000000..a7ace1a
--- /dev/null
+++ b/sys/libsre/sre.h
@@ -0,0 +1,93 @@
+#pragma once
+
+#include <u.h>
+#include <libn.h>
+
+enum
+{
+    Toperator = RuneMask + 1,
+    Tstart    = Toperator,
+    Trparen,
+    Tlparen,
+    Tor,
+    Tcat,
+    Tstar,
+    Tplus,
+    Tqmark,
+
+    Tany = Toperator << 1,
+    Tnop,
+    Tbol,
+    Teol,
+    Tcclass,
+    Tnclass,
+    Tend,
+
+    isoptor = Toperator,
+    isopand = Toperator << 1,
+};
+
+typedef struct Class Class;
+typedef struct State State;
+typedef struct Patch Patch;
+typedef struct Node Node;
+
+typedef struct Parser  Parser;
+typedef struct Machine Machine;
+
+struct Class
+{
+    rune *end;
+    rune span[64];
+};
+
+struct State 
+{
+    int type;
+    union {
+        State *l;
+    };
+    union {
+        State *r;
+        State *out;
+    };
+};
+
+struct Patch
+{
+    State *s;
+    Patch *link;
+};
+
+struct Node
+{
+    State *beg;
+    State *end;
+};
+
+struct Parser
+{
+    Machine *mach;
+    byte    *re;
+    int     wasopand : 1;
+    int     *optor, optorstk[1000];
+    Node    *node,  nodestk[1000];
+};
+
+struct Machine
+{
+    /* memory buffers */
+    struct {
+        void *heap;
+        mem·Reallocator;
+    };
+    State *state, statestk[1000];
+
+    struct {
+        int   cap;
+        int   len;
+        Class *c;
+    } class;
+
+    State *entry;
+};