diff --git a/src/util.c b/src/util.c
index b80cb708..d815edf5 100644
--- a/src/util.c
+++ b/src/util.c
@@ -777,6 +777,450 @@ ssize_t util_http_build_header(char * out, size_t len, ssize_t offset,
     return ret;
 }
 
+#define __SELECT_BEST_MAX_ARGS  8
+
+struct __args {
+    const char *comp;
+    char *group;
+    int best_comp;
+    int best_group;
+    int best_all;
+};
+
+static inline int __fill_arg(struct __args *arg, const char *str)
+{
+    char *delm;
+    size_t len;
+
+    arg->comp = str;
+    arg->best_comp = 0;
+    arg->best_group = 0;
+    arg->best_all = 0;
+
+    len = strlen(str);
+    arg->group = malloc(len + 2);
+    if (!arg->group)
+        return -1;
+
+    memcpy(arg->group, str, len + 1);
+
+    delm = strstr(arg->group, "/");
+    if (delm) {
+        delm[0] = '/';
+        delm[1] = '*';
+        delm[2] = 0;
+    }
+
+    return 0;
+}
+
+static inline void __free_args(struct __args *arg, size_t len)
+{
+    size_t i;
+
+    for (i = 0; i < len; i++) {
+        free(arg->group);
+    }
+}
+
+    // Accept: text/html, application/xhtml+xml, */*
+    // Accept: text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8
+
+static inline int __parse_q(const char *str)
+{
+    int ret = 0;
+    int mul = 1000;
+
+    for (; *str; str++) {
+        if (*str >= '0' && *str <= '9') {
+            ret += mul * (*str - '0');
+            mul /= 10;
+        } else if (*str == '.') {
+            mul = 100;
+        } else {
+            ICECAST_LOG_ERROR("Badly formated quality parameter found.");
+            return -1;
+        }
+    }
+
+    return ret;
+}
+
+static inline int __find_q_in_index(icecast_kva_t *kv, size_t idx)
+{
+    size_t i = kv->index[idx] + 1;
+    size_t last;
+
+    if (kv->indexlen <= (idx + 1)) {
+        last = kv->kvlen - 1;
+    } else {
+        last = kv->index[idx + 1] - 1;
+    }
+
+    for (; i <= last; i++) {
+        if (kv->kv[i].key && kv->kv[i].key && strcasecmp(kv->kv[i].key, "q") == 0) {
+            return __parse_q(kv->kv[i].value);
+        }
+    }
+
+    return 1000;
+}
+
+const char *util_http_select_best(const char *input, const char *first, ...)
+{
+    struct __args arg[__SELECT_BEST_MAX_ARGS];
+    icecast_kva_t *kv;
+    const char *p;
+    size_t arglen = 1;
+    size_t i, h;
+    va_list ap;
+    int q;
+    int best_q = 0;
+
+    if (__fill_arg(&(arg[0]), first) == -1) {
+        ICECAST_LOG_ERROR("Can not allocate memory. Selecting first option.");
+        return first;
+    }
+    va_start(ap, first);
+    while ((p = (const char*)va_arg(ap, const char*))) {
+        if (arglen == __SELECT_BEST_MAX_ARGS) {
+            ICECAST_LOG_ERROR("More arguments given than supported. Currently %zu args are supported.", (size_t)__SELECT_BEST_MAX_ARGS);
+            break;
+        }
+        if (__fill_arg(&(arg[arglen]), p) == -1) {
+            ICECAST_LOG_ERROR("Can not allocate memory. Selecting first option.");
+            __free_args(arg, arglen);
+            return first;
+        }
+
+        arglen++;
+    }
+    va_end(ap);
+
+    kv = util_parse_http_cn(input);
+    if (!kv) {
+        ICECAST_LOG_ERROR("Input string does not parse as KVA. Selecting first option.");
+        __free_args(arg, arglen);
+        return first;
+    }
+
+    ICECAST_LOG_DEBUG("--- DUMP ---");
+    for (i = 0; i < kv->kvlen; i++) {
+        ICECAST_LOG_DEBUG("kv[%zu] = {.key='%H', .value='%H'}", i, kv->kv[i].key, kv->kv[i].value);
+    }
+    for (i = 0; i < kv->indexlen; i++) {
+        ICECAST_LOG_DEBUG("index[%zu] = %zu", i, kv->index[i]);
+    }
+    ICECAST_LOG_DEBUG("--- END OF DUMP ---");
+
+    for (h = 0; h < arglen; h++) {
+        for (i = 0; i < kv->indexlen; i++) {
+            p = kv->kv[kv->index[i]].key;
+            if (!p) {
+                continue;
+            }
+
+            q = __find_q_in_index(kv, i);
+            if (best_q < q) {
+                best_q = q;
+            }
+
+            if (strcasecmp(p, arg[h].comp) == 0) {
+                if (arg[h].best_comp < q) {
+                    arg[h].best_comp = q;
+                }
+            }
+
+            if (strcasecmp(p, arg[h].group) == 0) {
+                if (arg[h].best_group < q) {
+                    arg[h].best_group = q;
+                }
+            }
+        }
+    }
+
+    util_kva_free(kv);
+
+    p = NULL;
+    for (h = 0; p == NULL && h < arglen; h++) {
+        if (arg[h].best_comp == best_q) {
+            p = arg[h].comp;
+        }
+    }
+    for (h = 0; p == NULL && h < arglen; h++) {
+        if (arg[h].best_group == best_q) {
+            p = arg[h].comp;
+        }
+    }
+
+    __free_args(arg, arglen);
+
+    if (p == NULL) {
+        p = first;
+    }
+
+    return p;
+}
+
+static inline void __skip_space(char **p)
+{
+    for (; **p == ' '; (*p)++);
+}
+
+static inline int __is_token(const char p)
+{
+    return (p >= 'a' && p <= 'z') || (p >= 'A' && p <= 'Z') || (p >= '0' && p <= '9') ||
+            p == '!' || p == '#'  ||  p == '$' || p == '%'  ||  p == '&' || p == '\'' ||
+            p == '*' || p == '+'  ||  p == '-' || p == '.'  ||  p == '^' || p == '_'  ||
+            p == '|' || p == '~'  ||  p == '/';
+
+}
+
+enum __tokenizer_result {
+    __TOKENIZER_RESULT_COMMA,
+    __TOKENIZER_RESULT_EQ,
+    __TOKENIZER_RESULT_SEMICOLON,
+    __TOKENIZER_RESULT_ILSEQ,
+    __TOKENIZER_RESULT_EOS
+};
+
+static inline enum __tokenizer_result __tokenizer_res_from_char(const char p)
+{
+    switch (p) {
+        case 0:
+            return __TOKENIZER_RESULT_EOS;
+        break;
+        case ',':
+            return __TOKENIZER_RESULT_COMMA;
+        break;
+        case '=':
+            return __TOKENIZER_RESULT_EQ;
+        break;
+        case ';':
+            return __TOKENIZER_RESULT_SEMICOLON;
+        break;
+        default:
+            return __TOKENIZER_RESULT_ILSEQ;
+        break;
+    }
+}
+
+static enum __tokenizer_result __tokenizer_str(char **out, char **in)
+{
+    char *p, *o;
+    char c;
+
+    __skip_space(in);
+
+    p = *in;
+    if (*p != '"')
+        return __TOKENIZER_RESULT_ILSEQ;
+
+    p++;
+    o = p;
+
+    for (; (c = *p); p++) {
+        if (c == '\t' || c == ' ' || c == 0x21 || (c >= 0x23 && c <= 0x5B) || (c >= 0x5D && c <= 0x7E) || (c >= 0x80 && c <= 0xFF)) {
+            *(o++) = c;
+        } else if (c == '\\') {
+            p++;
+            c = *p;
+            if (c == 0) {
+                return __TOKENIZER_RESULT_ILSEQ;
+            } else {
+                *(o++) = c;
+            }
+        } else if (c == '"') {
+            *o = 0;
+            break;
+        } else {
+            return __TOKENIZER_RESULT_ILSEQ;
+        }
+    }
+
+    if (*p == '"') {
+        p++;
+        *in = p + 1;
+        __skip_space(in);
+        return __tokenizer_res_from_char(*p);
+    } else {
+        return __TOKENIZER_RESULT_ILSEQ;
+    }
+}
+
+static enum __tokenizer_result __tokenizer(char **out, char **in)
+{
+    char *p;
+    char c = 0;
+
+    __skip_space(in);
+
+    p = *in;
+
+    switch (*p) {
+        case '=':
+        /* fall through */
+        case ',':
+        /* fall through */
+        case ';':
+            return __TOKENIZER_RESULT_ILSEQ;
+        break;
+        case 0:
+            return __TOKENIZER_RESULT_EOS;
+        break;
+        case '"':
+            return __tokenizer_str(out, in);
+        break;
+    }
+
+    *out = p;
+    for (; __is_token(*p); p++);
+
+    *in = p;
+    if (*p) {
+        __skip_space(in);
+        c = **in;
+        if (c != 0) {
+            (*in)++;
+        }
+    }
+    *p = 0;
+
+    return __tokenizer_res_from_char(c);
+}
+
+#define HTTP_CN_INDEX_INCREMENT     8
+#define HTTP_CN_KV_INCREMENT        8
+
+static inline int __resize_array(void **also_ptr, void **array, size_t size, size_t *len, size_t revlen, size_t inc)
+{
+    void *n;
+
+    if (*len > revlen)
+        return 0;
+
+    n = realloc(*array, size*(*len + inc));
+    if (!n) {
+        return -1;
+    }
+
+    memset(n + size * *len, 0, size * inc);
+
+    *also_ptr = *array = n;
+    *len += inc;
+
+    return 0;
+}
+
+icecast_kva_t * util_parse_http_cn(const char *cnstr)
+{
+    icecast_kva_t *ret;
+    char *in;
+    int eos = 0;
+    size_t indexphylen = HTTP_CN_INDEX_INCREMENT;
+    size_t kvphylen = HTTP_CN_KV_INCREMENT;
+
+    if (!cnstr || !*cnstr)
+        return NULL;
+
+    ret = calloc(1, sizeof(*ret));
+
+    if (!ret)
+        return NULL;
+
+    ret->_tofree[0] = in = strdup(cnstr);
+    ret->_tofree[1] = ret->index = calloc(HTTP_CN_INDEX_INCREMENT, sizeof(*(ret->index)));
+    ret->_tofree[2] = ret->kv = calloc(HTTP_CN_KV_INCREMENT, sizeof(*(ret->kv)));
+    if (!ret->_tofree[0] || !ret->_tofree[1] || !ret->_tofree[2]) {
+        util_kva_free(ret);
+        return NULL;
+    }
+
+
+    /* we have at minimum one token */
+    ret->indexlen = 1;
+    ret->kvlen = 1;
+
+    while (!eos) {
+        char *out = NULL;
+        enum __tokenizer_result res = __tokenizer(&out, &in);
+
+        switch (res) {
+            case __TOKENIZER_RESULT_ILSEQ:
+                ICECAST_LOG_DEBUG("Illegal byte sequence error from tokenizer.");
+                util_kva_free(ret);
+                return NULL;
+            break;
+            case __TOKENIZER_RESULT_EOS:
+            /* fall through */
+            case __TOKENIZER_RESULT_COMMA:
+            /* fall through */
+            case __TOKENIZER_RESULT_EQ:
+            /* fall through */
+            case __TOKENIZER_RESULT_SEMICOLON:
+                ICECAST_LOG_DEBUG("OK from tokenizer.");
+                /* no-op */
+            break;
+        }
+
+        if (__resize_array(&(ret->_tofree[2]), (void**)&(ret->kv), sizeof(*(ret->kv)), &kvphylen, ret->kvlen, HTTP_CN_KV_INCREMENT) == -1 ||
+            __resize_array(&(ret->_tofree[1]), (void**)&(ret->index), sizeof(*(ret->index)), &indexphylen, ret->indexlen, HTTP_CN_INDEX_INCREMENT) == -1) {
+            util_kva_free(ret);
+            return NULL;
+        }
+
+        if (ret->kv[ret->kvlen-1].key == NULL) {
+            ret->kv[ret->kvlen-1].key = out;
+        } else if (ret->kv[ret->kvlen-1].value == NULL) {
+            ret->kv[ret->kvlen-1].value = out;
+        } else {
+            util_kva_free(ret);
+            return NULL;
+        }
+
+        switch (res) {
+            case __TOKENIZER_RESULT_EOS:
+                ICECAST_LOG_DEBUG("End of string from tokenizer.");
+                eos = 1;
+                continue;
+            break;
+            case __TOKENIZER_RESULT_COMMA:
+                ICECAST_LOG_DEBUG("Comma from tokenizer.");
+                ret->index[ret->indexlen++] = ret->kvlen;
+                ret->kvlen++;
+            break;
+            case __TOKENIZER_RESULT_EQ:
+                ICECAST_LOG_DEBUG("Eq from tokenizer.");
+                /* no-op */
+            break;
+            case __TOKENIZER_RESULT_SEMICOLON:
+                ICECAST_LOG_DEBUG("Semicolon from tokenizer.");
+                ret->kvlen++;
+            break;
+            default:
+                util_kva_free(ret);
+                return NULL;
+            break;
+        }
+
+        ICECAST_LOG_DEBUG("next...");
+    }
+
+    return ret;
+}
+
+void util_kva_free(icecast_kva_t *kva)
+{
+    size_t i;
+
+    if (!kva)
+        return;
+
+    for (i = 0; i < (sizeof(kva->_tofree)/sizeof(*(kva->_tofree))); i++)
+        free(kva->_tofree[i]);
+
+    free(kva);
+}
 
 util_dict *util_dict_new(void)
 {
diff --git a/src/util.h b/src/util.h
index 8ba76fa5..469aca72 100644
--- a/src/util.h
+++ b/src/util.h
@@ -88,6 +88,24 @@ ssize_t util_http_build_header(char * out, size_t len, ssize_t offset,
         struct source_tag * source,
         struct _client_tag * client);
 
+const char *util_http_select_best(const char *input, const char *first, ...);
+
+typedef struct icecast_kv_tag {
+    char *key;
+    char *value;
+} icecast_kv_t;
+
+typedef struct icecast_kva_tag {
+    void   *_tofree[3];
+    size_t kvlen;
+    size_t indexlen;
+    size_t *index;
+    icecast_kv_t *kv;
+} icecast_kva_t;
+
+icecast_kva_t * util_parse_http_cn(const char *cnstr);
+void util_kva_free(icecast_kva_t *kva);
+
 /* String dictionary type, without support for NULL keys, or multiple
  * instances of the same key */
 typedef struct _util_dict {