fishlim: Implement correct handling of long and UTF-8 messages

plugins/fishlim/utils.c

@@ -23,6 +23,7 @@
 */
 
 #include "utils.h"
+#include "fish.h"
 
 /**
  * Calculate the length of Base64-encoded string
@@ -67,4 +68,82 @@ unsigned long cbc_len(size_t plaintext_len) {
  */
 unsigned long ecb_len(size_t plaintext_len) {
     return base64_fish_len(plaintext_len);
+}
+
+/**
+ * Calculate the length of encrypted string in 'mode' mode
+ *
+ * @param plaintext_len Length of plaintext
+ * @param mode          Encryption mode
+ * @return Size of encoded string
+ */
+unsigned long encoded_len(size_t plaintext_len, enum fish_mode mode) {
+    switch (mode) {
+
+        case FISH_CBC_MODE:
+            return cbc_len(plaintext_len);
+            break;
+
+        case FISH_ECB_MODE:
+            return ecb_len(plaintext_len);
+    }
+
+    return 0;
+}
+
+/**
+ * Determine the maximum length of plaintext for a 'max_len' limit taking care the overload of encryption
+ *
+ * @param max_len   Limit for plaintext
+ * @param mode      Encryption mode
+ * @return Maximum allowed plaintext length
+ */
+int max_text_command_len(size_t max_len, enum fish_mode mode) {
+    int len;
+
+    for (len = max_len; encoded_len(len, mode) > max_len; --len);
+    return len;
+}
+
+/**
+ * Iterate over 'data' in chunks of 'max_chunk_len' taking care the UTF-8 characters
+ *
+ * @param data              Data to iterate
+ * @param max_chunk_len     Size of biggest chunk
+ * @param [out] chunk_len   Current chunk length
+ * @return Pointer to current chunk position or NULL if not have more chunks
+ */
+const char *foreach_utf8_data_chunks(const char *data, int max_chunk_len, int *chunk_len) {
+    int data_len, last_chunk_len = 0;
+
+    if (!*data) {
+        return NULL;
+    }
+
+    /* Last chunk of data */
+    data_len = strlen(data);
+    if (data_len <= max_chunk_len) {
+        *chunk_len = data_len;
+        return data;
+    }
+
+    *chunk_len = 0;
+    const char *utf8_character = data;
+
+    /* Not valid UTF-8, but maybe valid text, just split into max length */
+    if (!g_utf8_validate(data, -1, NULL)) {
+        *chunk_len = max_chunk_len;
+        return utf8_character;
+    }
+
+    while (*utf8_character && *chunk_len <= max_chunk_len) {
+        last_chunk_len = *chunk_len;
+        *chunk_len = (g_utf8_next_char(utf8_character) - data) * sizeof(*utf8_character);
+        utf8_character = g_utf8_next_char(utf8_character);
+    }
+
+    /* We need the previous length before overflow the limit */
+    *chunk_len = last_chunk_len;
+
+    return utf8_character;
 }