1 files changed, 104 insertions, 0 deletions
diff --git a/src/XpathTokenizer.c b/src/XpathTokenizer.c
new file mode 100644
index 0000000..32bede9
--- /dev/null
+++ b/src/XpathTokenizer.c
@@ -0,0 +1,104 @@
+#include "../include/quickmedia/XpathTokenizer.h"
+#include <stdlib.h>
+#include <string.h>
+
+void quickmedia_xpath_tokenizer_init(QuickMediaXpathTokenizer *self, const char *xpath) {
+    self->code = xpath;
+    self->identifier.data = NULL;
+    self->identifier.size = 0;
+}
+
+static int is_alpha(char c) {
+    return (c >= 'a' && c <= 'z') || (c >= 'A' && c <= 'Z');
+}
+
+static int is_num(char c) {
+    return c >= '0' && c <= '9';
+}
+
+static int is_alphanum(char c) {
+    return is_alpha(c) || is_num(c);
+}
+
+static const char* find_end_of_string(const char *str, char escape_symbol) {
+    int escape = 0;
+    while(*str != '\0') {
+        char c = *str;
+        if(c == '\\') {
+            escape = !escape;
+        } else if(c == escape_symbol) {
+            if(!escape)
+                return str;
+        } else {
+            escape = 0;
+        }
+        ++str;
+    }
+    return str;
+}
+
+QuickMediaXpathToken quickmedia_xpath_tokenizer_next(QuickMediaXpathTokenizer *self) {
+    char c = *self->code;
+    if(c == '/') {
+        ++self->code;
+        c = *self->code;
+        if(c == '/') {
+            ++self->code;
+            return QUICKMEDIA_XPATH_TOKEN_CHILD_RECURSIVE;
+        }
+        return QUICKMEDIA_XPATH_TOKEN_CHILD;
+    } else if(is_alpha(c)) {
+        self->identifier.data = self->code;
+        ++self->code;
+        while(is_alphanum(*self->code) || *self->code == '_' || *self->code == '-') {
+            ++self->code;
+        }
+        self->identifier.size = self->code - self->identifier.data;
+        return QUICKMEDIA_XPATH_TOKEN_IDENTIFIER;
+    } else if(c == '[') {
+        ++self->code;
+        return QUICKMEDIA_XPATH_TOKEN_OPEN_BRACKET;
+    } else if(c == ']') {
+        ++self->code;
+        return QUICKMEDIA_XPATH_TOKEN_CLOSING_BRACKET;
+    } else if(c == '=') {
+        ++self->code;
+        return QUICKMEDIA_XPATH_TOKEN_EQUAL;
+    } else if(c == '"' || c == '\'') {
+        char escape_symbol = c;
+        ++self->code;
+        self->string.data = self->code;
+        self->code = find_end_of_string(self->string.data, escape_symbol);
+        if(*self->code == '\0') {
+            /* Reached end of xpath before end of string */
+            return QUICKMEDIA_XPATH_TOKEN_INVALID;
+        }
+        self->string.size = self->code - self->string.data;
+        ++self->code;
+        return QUICKMEDIA_XPATH_TOKEN_STRING;
+    } else if(c == '\0') {
+        return QUICKMEDIA_XPATH_TOKEN_END_OF_FILE;
+    } else {
+        /* Invalid symbol @c */
+        return QUICKMEDIA_XPATH_TOKEN_INVALID;
+    }
+}
+
+int quickmedia_xpath_tokenizer_next_if(QuickMediaXpathTokenizer *self, QuickMediaXpathToken token) {
+    const char *restore_point = self->code;
+    if(quickmedia_xpath_tokenizer_next(self) == token)
+        return 0;
+    self->code = restore_point;
+    return -1;
+}
+
+char* quickmedia_xpath_tokenizer_copy_identifier(QuickMediaXpathTokenizer *self) {
+    char *result = malloc(self->identifier.size + 1);
+    result[self->identifier.size] = '\0';
+    memcpy(result, self->identifier.data, self->identifier.size);
+    return result;
+}
+
+char* quickmedia_xpath_tokenizer_copy_string(QuickMediaXpathTokenizer *self) {
+    return quickmedia_xpath_tokenizer_copy_identifier(self);
+}