initial commit

There is some work to be done on top of this such as creating a way to
manage memory to be freed later without needing to manually manage every
pointer.

Signed-off-by: Chloe M. <chloe@aptel.org>

1 files changed, 191 insertions, 0 deletions

diff --git a/core/lexer.c b/core/lexer.c
new file mode 100644
index 0000000..2eb50a1
--- /dev/null
+++ b/core/lexer.c
@@ -0,0 +1,191 @@
+#include <ctype.h>
+#include <stdbool.h>
+#include <stdlib.h>
+#include <string.h>
+#include "defconf/lexer.h"
+
+/* Constants */
+#define MAX_KW_LEN 32
+
+/*
+ * Returns true if the given character is a whitespace character
+ *
+ * @c:  Character to check
+ */
+static inline bool
+lexer_is_ws(char c)
+{
+    switch (c) {
+    case '\t':
+    case ' ':
+    case '\f':
+    case '\r':
+    case '\n':
+        return true;
+    }
+
+    return false;
+}
+
+/*
+ * Place a character back in the putback buffer
+ *
+ * @state:  Defconf state
+ * @c:      Character to put back
+ */
+static inline void
+lexer_putback(struct defconf_state *state, int c)
+{
+    if (state == NULL) {
+        return;
+    }
+
+    state->lex_cache = c;
+}
+
+/*
+ * Nom a single character from the input source file
+ *
+ * @state:      Defconf state machine
+ * @skip_ws:    If true, skip whitespace
+ *
+ * Returns the character on success, otherwise EOF on failure
+ */
+static int
+lexer_nom(struct defconf_state *state, bool skip_ws)
+{
+    int c;
+
+    if (state == NULL) {
+        return EOF;
+    }
+
+    if ((c = state->lex_cache) != '\0') {
+        state->lex_cache = '\0';
+        if (lexer_is_ws(c) && !skip_ws)
+            return c;
+        if (!lexer_is_ws(c))
+            return c;
+    }
+
+    while ((c = fgetc(state->in_fp)) != EOF) {
+        if (lexer_is_ws(c) && skip_ws)
+            continue;
+
+        break;
+    }
+
+    return c;
+}
+
+/*
+ * Scan a keyword
+ *
+ * @state: Defconf state
+ * @lc:    Last character provided
+ */
+static char *
+lexer_scan_ident(struct defconf_state *state, int lc)
+{
+    char c, buf[MAX_KW_LEN];
+    size_t bufind = 0;
+
+    if (state == NULL) {
+        return NULL;
+    }
+
+    if (!isalpha(lc)) {
+        return NULL;
+    }
+
+    buf[bufind++] = lc;
+    for (;;) {
+        if (bufind >= sizeof(buf) - 1) {
+            printf("fatal: token exceeds maximum length\n");
+            return NULL;
+        }
+
+        c = lexer_nom(state, false);
+        if (!isalnum(c) && c != '_') {
+            lexer_putback(state, c);
+            buf[bufind] = '\0';
+            return strdup(buf);
+        }
+
+        buf[bufind++] = c;
+    }
+
+    return NULL;
+}
+
+/*
+ * Check if a keyword matches against known keywords
+ *
+ * @kw:         Keyword to check
+ * @result:     Result is written here
+ *
+ * Returns zero on success
+ */
+static int
+lexer_check_kw(const char *kw, struct token *result)
+{
+    if (kw == NULL || result == NULL) {
+        return -1;
+    }
+
+    switch (*kw) {
+    case 'd':
+        if (strcmp(kw, "define") == 0) {
+            result->type = TT_DEFINE;
+            return 0;
+        }
+
+        break;
+    case 't':
+        if (strcmp(kw, "true") == 0) {
+            result->type = TT_TRUE;
+            return 0;
+        }
+
+        break;
+    case 'f':
+        if (strcmp(kw, "false") == 0) {
+            result->type = TT_FALSE;
+            return 0;
+        }
+
+        break;
+    }
+
+    return -1;
+}
+
+int
+lexer_scan(struct defconf_state *state, struct token *result)
+{
+    int c;
+    char *ident;
+
+    if (state == NULL || result == NULL) {
+        return -1;
+    }
+
+    if ((c = lexer_nom(state, true)) == EOF) {
+        return -1;
+    }
+
+    /* Scan for an identifier */
+    if ((ident = lexer_scan_ident(state, c)) != NULL) {
+        if (lexer_check_kw(ident, result) == 0) {
+            free(ident);
+            return 0;
+        }
+
+        result->type = TT_IDENT;
+        result->s = ident;
+        return 0;
+    }
+
+    printf("fatal: unexpected token '%c'\n", c);
+    return -1;
+}