Add floating point number parsing

5 files changed, 135 insertions, 22 deletions

diff --git a/Makefile b/Makefile
index 35447d3..7fed16b 100644
--- a/Makefile
+++ b/Makefile
@@ -17,7 +17,7 @@ DOT := dot -Gmargin=0.7 -Gcolor=white -Gfontcolor=white \
                 -T png | kitty +kitten icat
 
 # Output executable.
-TARGET := bdl
+TARGET := badc
 BIN    := $(BUILD_DIR)/$(TARGET)
 
 # Compiler and linker configuration.
diff --git a/src/badlib.h b/src/badlib.h
index 9802278..d334fa7 100644
--- a/src/badlib.h
+++ b/src/badlib.h
@@ -586,6 +586,9 @@ str_to_int(Str s) {
         s = str_remove_prefix(s, cstr("0b"));
         while (s.size) {
             char c = str_next(&s);
+            if (c == '_') {
+                continue;
+            }
             assert(c == '0' || c == '1');
             num = num * 2 + (c - '0');
         }
@@ -594,6 +597,9 @@ str_to_int(Str s) {
         s = str_remove_prefix(s, cstr("0x"));
         while (s.size) {
             char c = str_next(&s);
+            if (c == '_') {
+                continue;
+            }
             assert((c >= '0' && c <= '9') || (c >= 'a' && c <= 'f') ||
                    (c >= 'A' && c <= 'F'));
             if (c >= '0' && c <= '9') {
@@ -606,11 +612,95 @@ str_to_int(Str s) {
         }
     } else {
         // Decimal number.
+        char c = str_peek(s);
+        sz neg = 1;
+        if (c == '-') {
+            neg = -1;
+            str_next(&s);
+        } else if (c == '+') {
+            str_next(&s);
+        }
+        // TODO: check if it fits within the s64 range.
         while (s.size) {
             char c = str_next(&s);
+            if (c == '_') {
+                continue;
+            }
             assert(c >= '0' && c <= '9');
             num = num * 10 + (c - '0');
         }
+        num *= neg;
+    }
+    return num;
+}
+
+f64
+str_to_float(Str s) {
+    char c = str_peek(s);
+    f64 neg = 1.0;
+    if (c == '-') {
+        neg = -1.0;
+        str_next(&s);
+    } else if (c == '+') {
+        str_next(&s);
+    }
+    f64 num = 0.0;
+    // Integral part.
+    while (s.size) {
+        char c = str_next(&s);
+        if (c == '_') {
+            continue;
+        }
+        if (c == '.') {
+            break;
+        }
+        assert(c >= '0' && c <= '9');
+        num = num * 10 + (c - '0');
+    }
+    // Fractional part.
+    f64 frac = 0;
+    sz frac_digits = 1;
+    bool has_exponent = false;
+    while (s.size) {
+        char c = str_next(&s);
+        if (c == '_') {
+            continue;
+        }
+        if (c == 'e' || c == 'E') {
+            has_exponent = true;
+            break;
+        }
+        assert(c >= '0' && c <= '9');
+        frac = frac * 10 + (c - '0');
+        frac_digits *= 10;
+    }
+    num *= neg;
+    num += frac / frac_digits;
+    if (has_exponent) {
+        bool exp_neg = false;
+        char c = str_peek(s);
+        if (c == '-') {
+            exp_neg = true;
+            str_next(&s);
+        } else if (c == '+') {
+            str_next(&s);
+        }
+        sz exponent = 0;
+        while (s.size) {
+            c = str_next(&s);
+            if (c == '_') {
+                continue;
+            }
+            assert(c >= '0' && c <= '9');
+            exponent = exponent * 10 + (c - '0');
+        }
+        for (sz i = 0; i < exponent; i++) {
+            if (!exp_neg) {
+                num *= 10;
+            } else {
+                num /= 10;
+            }
+        }
     }
     return num;
 }
diff --git a/src/lexer.c b/src/lexer.c
index 404082b..997a9f3 100644
--- a/src/lexer.c
+++ b/src/lexer.c
@@ -14,7 +14,8 @@ typedef enum TokenKind {
     TOK_RCURLY,   // }
 
     // Basic literals.
-    TOK_NUMBER,
+    TOK_NUM_INT,
+    TOK_NUM_FLOAT,
     TOK_SYMBOL,
     TOK_STRING,
 
@@ -81,7 +82,8 @@ Str token_str[] = {
     [TOK_RCURLY] = cstr("RCURLY"),
 
     // Basic literals.
-    [TOK_NUMBER] = cstr("NUMBER"),
+    [TOK_NUM_INT] = cstr("INUMBER"),
+    [TOK_NUM_FLOAT] = cstr("FNUMBER"),
     [TOK_SYMBOL] = cstr("SYMBOL"),
     [TOK_STRING] = cstr("STRING"),
 
@@ -304,7 +306,7 @@ emit_token_number(Scanner *scanner) {
                 continue;
             }
             if (scan_is_valid_split(c)) {
-                return emit_token(current, scanner, TOK_NUMBER);
+                return emit_token(current, scanner, TOK_NUM_INT);
             }
             scan_skip_until_valid(scanner);
             return emit_token_err(
@@ -321,7 +323,7 @@ emit_token_number(Scanner *scanner) {
                 continue;
             }
             if (scan_is_valid_split(c)) {
-                return emit_token(current, scanner, TOK_NUMBER);
+                return emit_token(current, scanner, TOK_NUM_INT);
             }
             scan_skip_until_valid(scanner);
             return emit_token_err(&current,
@@ -340,7 +342,7 @@ emit_token_number(Scanner *scanner) {
                 continue;
             }
             if (scan_is_valid_split(c)) {
-                return emit_token(current, scanner, TOK_NUMBER);
+                return emit_token(current, scanner, TOK_NUM_INT);
             }
             scan_skip_until_valid(scanner);
             return emit_token_err(&current, cstr("malformed number"));
@@ -362,7 +364,7 @@ emit_token_number(Scanner *scanner) {
                 continue;
             }
             if (scan_is_valid_split(c)) {
-                return emit_token(current, scanner, TOK_NUMBER);
+                return emit_token(current, scanner, TOK_NUM_FLOAT);
             }
             scan_skip_until_valid(scanner);
             return emit_token_err(&current, cstr("malformed number"));
@@ -385,7 +387,7 @@ emit_token_number(Scanner *scanner) {
                     cstr("malformed number: decimals not allowed on exponent"));
             }
             if (scan_is_valid_split(c)) {
-                return emit_token(current, scanner, TOK_NUMBER);
+                return emit_token(current, scanner, TOK_NUM_FLOAT);
             }
             scan_skip_until_valid(scanner);
             return emit_token_err(&current, cstr("malformed number"));
diff --git a/src/main.c b/src/main.c
index c321d7c..f15afed 100644
--- a/src/main.c
+++ b/src/main.c
@@ -39,19 +39,27 @@ print_tokens(Str path, Token *tokens) {
 //
 
 typedef enum NodeKind {
-    NODE_NUMBER,
+    NODE_NUM_INT,
+    NODE_NUM_FLOAT,
     // TODO: probably want to handle ints/unsigneds/floats separately.
     NODE_ADD,
     NODE_SUB,
     NODE_DIV,
     NODE_MUL,
-    // TODO: MOD
+    NODE_MOD,
 } NodeKind;
 
 Str node_str[] = {
-    [NODE_NUMBER] = cstr("NUM"), [NODE_ADD] = cstr("ADD"),
-    [NODE_SUB] = cstr("SUB"),    [NODE_DIV] = cstr("DIV"),
+    // Arithmetic.
+    [NODE_ADD] = cstr("ADD"),
+    [NODE_SUB] = cstr("SUB"),
+    [NODE_DIV] = cstr("DIV"),
     [NODE_MUL] = cstr("MUL"),
+    [NODE_MOD] = cstr("MOD"),
+
+    // Literals.
+    [NODE_NUM_INT] = cstr("INT"),
+    [NODE_NUM_FLOAT] = cstr("FLOAT"),
 };
 
 typedef struct Node {
@@ -62,9 +70,7 @@ typedef struct Node {
     NodeKind kind;
     union {
         f64 d;
-        f32 f;
         sz i;
-        u64 u;
     } value;
     struct Node *left;
     struct Node *right;
@@ -108,7 +114,7 @@ typedef enum {
     PREC_EQUALITY,    // == !=
     PREC_COMPARISON,  // < > <= >=
     PREC_TERM,        // + -
-    PREC_FACTOR,      // * /
+    PREC_FACTOR,      // * / %
     PREC_UNARY,       // ! -
     PREC_CALL,        // . ()
     PREC_PRIMARY      // highest precedence
@@ -136,7 +142,9 @@ ParseRule parse_rules[] = {
     [TOK_ADD] = {NULL, parse_binary, PREC_TERM},
     [TOK_DIV] = {NULL, parse_binary, PREC_FACTOR},
     [TOK_MUL] = {NULL, parse_binary, PREC_FACTOR},
-    [TOK_NUMBER] = {parse_number, NULL, PREC_NONE},
+    [TOK_MOD] = {NULL, parse_binary, PREC_FACTOR},
+    [TOK_NUM_INT] = {parse_number, NULL, PREC_NONE},
+    [TOK_NUM_FLOAT] = {parse_number, NULL, PREC_NONE},
     [TOK_EOF] = {NULL, NULL, PREC_NONE},
 };
 
@@ -201,7 +209,7 @@ parse_unary(Parser *parser) {
     print("parsing unary ");
     print_token(prev);
 #endif
-    TokenKind kind = parser->previous.kind;
+    TokenKind kind = prev.kind;
     parse_expr(parser, PREC_LOW);
     // TODO: ...
     switch (kind) {
@@ -227,6 +235,7 @@ parse_binary(Parser *parser) {
         case TOK_SUB: node = node_alloc(NODE_SUB, prev, parser->storage); break;
         case TOK_MUL: node = node_alloc(NODE_MUL, prev, parser->storage); break;
         case TOK_DIV: node = node_alloc(NODE_DIV, prev, parser->storage); break;
+        case TOK_MOD: node = node_alloc(NODE_MOD, prev, parser->storage); break;
         default: {
             parse_emit_err(parser, prev, cstr("unreachable"));
             return;
@@ -244,9 +253,18 @@ parse_number(Parser *parser) {
     print("parsing number ");
     print_token(prev);
 #endif
-    Node *node = node_alloc(NODE_NUMBER, prev, parser->storage);
-    node->value.i = str_to_int(prev.val);
-    // TODO: handle sign and/or floating point values.
+    Node *node = NULL;
+    switch (prev.kind) {
+        case TOK_NUM_INT: {
+            node = node_alloc(NODE_NUM_INT, prev, parser->storage);
+            node->value.i = str_to_int(prev.val);
+        } break;
+        case TOK_NUM_FLOAT: {
+            node = node_alloc(NODE_NUM_FLOAT, prev, parser->storage);
+            node->value.d = str_to_float(prev.val);
+        } break;
+        default: break;
+    }
     array_push(parser->nodes, node, parser->storage);
 }
 
@@ -268,7 +286,8 @@ graph_node(Node *node) {
     print("%d [width=2.5,shape=Mrecord,label=\"", node->id);
     print("<top> %s ", node_str[node->kind]);
     switch (node->kind) {
-        case NODE_NUMBER: print("| Value: %d", node->value.i); break;
+        case NODE_NUM_INT: print("| Value: %d", node->value.i); break;
+        case NODE_NUM_FLOAT: print("| Value: %f{2}", node->value.d); break;
         default: break;
     }
     print("| Line: %d | Col: %d ", node->line, node->col);
diff --git a/tests/expressions.bad b/tests/expressions.bad
index af60b9b..5b114d1 100644
--- a/tests/expressions.bad
+++ b/tests/expressions.bad
@@ -1,7 +1,9 @@
-1 + 2
+1 + -2
 
 1 + 2 * 3
 
 1 + 2 * 3 - 4
 
 1 + 2 * (3 - 4)
+
+1.0 - 1234.56e-3