Merge 41df65c455 into f48ee93931

src/link.c

@@ -77,10 +77,15 @@ static CURL *Link_to_curl(Link *link)
     if (ret) {
         lprintf(error, "%s", curl_easy_strerror(ret));
     }
-    ret = curl_easy_setopt(curl, CURLOPT_URL, link->f_url);
+
+    char *escaped_spaces = escape_char(link->f_url, ESCAPED_CHAR_SPACE);
+    ret = curl_easy_setopt(curl, CURLOPT_URL,
+                            escaped_spaces ? escaped_spaces : link->f_url);
     if (ret) {
         lprintf(error, "%s", curl_easy_strerror(ret));
     }
+    free(escaped_spaces);
+
     ret = curl_easy_setopt(curl, CURLOPT_TCP_KEEPALIVE, 1);
     if (ret) {
         lprintf(error, "%s", curl_easy_strerror(ret));

src/util.c

@@ -49,6 +49,61 @@ char *path_append(const char *path, const char *filename)
     return str;
 }
 
+char *escape_char(const char *url, const ESCAPE_CHAR c)
+{
+    char escape_me;
+
+    /* A space, for example, becomes thrice bigger as '%20' after escaping */
+    int how_bigger = 3;
+
+    switch (c) {
+        case ESCAPED_CHAR_SPACE:
+            escape_me = ' ';
+            break;
+        default:
+            lprintf(error, "unknown character to escape: %d\n", c);
+            return NULL;
+    }
+
+    if (!strchr(url, escape_me)) {
+        return NULL;
+    }
+
+    int len = strnlen(url, MAX_PATH_LEN);
+
+    /* Best case scenario of only one character to escape */
+    if (len + 3 > MAX_PATH_LEN) {
+        lprintf(fatal, "URL too long: %s\n", url);
+    }
+
+    char *escaped = CALLOC(MAX_PATH_LEN, sizeof(char));
+
+    int j = 0, k = 0;
+    for (int i = 0; i < len; i++) {
+        /* Precaution against writing beyond the buffer, since we do not count
+         * all spaces beforehand to know the exact amount of memory to calloc
+         * for the escaped URL.
+         */
+        if (k > MAX_PATH_LEN) {
+            lprintf(fatal, "URL too long: %s\n", url);
+        }
+        if (url[i] == escape_me) {
+            switch (c) {
+                case ESCAPED_CHAR_SPACE:
+                    strncpy(escaped + j, "%20", how_bigger);
+                    break;
+            }
+            k = j + how_bigger;
+            j = k;
+
+        } else {
+            escaped[j++] = url[i];
+        }
+    }
+    return escaped;
+}
+
+
 int64_t round_div(int64_t a, int64_t b)
 {
     return (a + (b / 2)) / b;

src/util.h

@@ -9,6 +9,13 @@
 #include <stdint.h>
 #include <stdio.h>
 
+/**
+ * \brief URL characters that can be escaped
+ */
+typedef enum  {
+    ESCAPED_CHAR_SPACE,
+} ESCAPE_CHAR;
+
 /**
  * \brief append a path
  * \details This function appends a path with the next level, while taking the
@@ -17,6 +24,16 @@
  */
 char *path_append(const char *path, const char *filename);
 
+/**
+ * \brief escape a single character in a URL
+ * \details This function escapes a character a URL and returns a pointer to
+ * the string with the escaped character. We don't use curl_easy_escape() on
+ * the entire URL because it would break the URL. For example, 'http://a c'
+ * becomes 'http:%2F%2Fa%20c', escaping more characters than we want.
+ * \note You need to free the char * after use.
+ */
+char *escape_char(const char *s, ESCAPE_CHAR c);
+
 /**
  * \brief division, but rounded to the nearest integer rather than truncating
  */