frontend: Add lexer+parser groundwork

Signed-off-by: Ian M. <ian@aptel.org>

1 files changed, 80 insertions, 0 deletions

diff --git a/core/lexer.c b/core/lexer.c
new file mode 100644
index 0000000..019724c
--- /dev/null
+++ b/core/lexer.c
@@ -0,0 +1,80 @@
+/*
+ * Copyright (c) 2026, Apollo Telephone Laboratories.
+ * Provided under the BSD-3 clause.
+ */
+
+#include <atas/lexer.h>
+#include <stdio.h>
+#include <stdbool.h>
+
+/*
+ * Returns true if the given char is a whitespace
+ *
+ * @c:  Character to compare against a set of whitespace
+ */
+static inline bool
+lexer_is_ws(char c)
+{
+    switch (c) {
+    case '\n':
+    case '\f':
+    case '\t':
+    case '\r':
+    case ' ':
+        return true;
+    }
+
+    return false;
+}
+
+/*
+ * Scan a single byte from the input source file
+ *
+ * @state:   Assembler state
+ * @skip_ws: If true, skip whitespaces
+ *
+ * Returns the scanned byte on success, returns '\0'
+ * on failure.
+ */
+static char
+lexer_scan_byte(struct atas_state *state, bool skip_ws)
+{
+    char c;
+
+    if (state == NULL) {
+        return '\0';
+    }
+
+    c = fgetc(state->in_file);
+    while (lexer_is_ws(c) && skip_ws) {
+        if ((c = fgetc(state->in_file)) == EOF) {
+            return '\0';
+        }
+    }
+
+    return c;
+}
+
+int
+atas_lexer_scan(struct atas_state *state, struct atas_token *token)
+{
+    char c;
+
+    if (state == NULL || token == NULL) {
+        return -1;
+    }
+
+    c = lexer_scan_byte(state, true);
+    if (c == '\0') {
+        return -1;
+    }
+
+    switch (c) {
+    case ',':
+        token->type = TT_COMMA;
+        token->c = c;
+        return 0;
+    }
+
+    return -1;
+}