url.c

Go to the documentation of this file.

 
#include "config.h"
#include <stdbool.h>
#include <string.h>
#include "mutt/lib.h"
#include "url.h"
#include "mime.h"

static const struct Mapping UrlMap[] = {
  { "file", U_FILE },     { "imap", U_IMAP },       { "imaps", U_IMAPS },
  { "pop", U_POP },       { "pops", U_POPS },       { "news", U_NNTP },
  { "nntp", U_NNTP },     { "snews", U_NNTPS },     { "nntps", U_NNTPS },
  { "mailto", U_MAILTO }, { "notmuch", U_NOTMUCH }, { "smtp", U_SMTP },
  { "smtps", U_SMTPS },   { NULL, U_UNKNOWN },
};

static bool parse_query_string(struct UrlQueryList *list, char *src)
{
  if (!src || (*src == '\0'))
    return false;
 
  bool again = true;
  while (again)
  {
    regmatch_t *match = mutt_prex_capture(PREX_URL_QUERY_KEY_VAL, src);
    if (!match)
      return false;
 
    regmatch_t *mkey = &match[PREX_URL_QUERY_KEY_VAL_MATCH_KEY];
    regmatch_t *mval = &match[PREX_URL_QUERY_KEY_VAL_MATCH_VAL];
 
    again = src[mutt_regmatch_end(mval)] != '\0';
 
    char *key = src + mutt_regmatch_start(mkey);
    char *val = src + mutt_regmatch_start(mval);
    src[mutt_regmatch_end(mkey)] = '\0';
    src[mutt_regmatch_end(mval)] = '\0';
    if ((url_pct_decode(key) < 0) || (url_pct_decode(val) < 0))
      return false;
 
    struct UrlQuery *qs = MUTT_MEM_CALLOC(1, struct UrlQuery);
    qs->name = key;
    qs->value = val;
    STAILQ_INSERT_TAIL(list, qs, entries);
 
    src += mutt_regmatch_end(mval) + again;
  }
 
  return true;
}

static enum UrlScheme get_scheme(const char *src, const regmatch_t *match)
{
  enum UrlScheme rc = U_UNKNOWN;
  if (src && match)
  {
    rc = mutt_map_get_value_n(src, mutt_regmatch_len(&match[PREX_URL_MATCH_SCHEME]), UrlMap);
    if (rc == -1)
      rc = U_UNKNOWN;
  }
  return rc;
}

static struct Url *url_new(void)
{
  struct Url *url = MUTT_MEM_CALLOC(1, struct Url);
  STAILQ_INIT(&url->query_strings);
  return url;
}

void url_free(struct Url **ptr)
{
  if (!ptr || !*ptr)
    return;
 
  struct Url *url = *ptr;
 
  struct UrlQueryList *l = &url->query_strings;
  while (!STAILQ_EMPTY(l))
  {
    struct UrlQuery *np = STAILQ_FIRST(l);
    STAILQ_REMOVE_HEAD(l, entries);
    // Don't free 'name', 'value': they are pointers into the 'src' string
    FREE(&np);
  }
 
  FREE(&url->src);
  FREE(ptr);
}

void url_pct_encode(char *buf, size_t buflen, const char *src)
{
  static const char *hex = "0123456789ABCDEF";
 
  if (!buf)
    return;
 
  *buf = '\0';
  buflen--;
  while (src && *src && (buflen != 0))
  {
    if (strchr(" /:&%=", *src))
    {
      if (buflen < 3)
        break;
 
      *buf++ = '%';
      *buf++ = hex[(*src >> 4) & 0xf];
      *buf++ = hex[*src & 0xf];
      src++;
      buflen -= 3;
      continue;
    }
    *buf++ = *src++;
    buflen--;
  }
  *buf = '\0';
}

int url_pct_decode(char *s)
{
  if (!s)
    return -1;
 
  char *d = NULL;
 
  for (d = s; *s; s++)
  {
    if (*s == '%')
    {
      if ((s[1] != '\0') && (s[2] != '\0') && mutt_isxdigit(s[1]) &&
          mutt_isxdigit(s[2]) && (hexval(s[1]) >= 0) && (hexval(s[2]) >= 0))
      {
        unsigned char c = (hexval(s[1]) << 4) | (hexval(s[2]));
        if (c == '\0')
          return -1;
        *d++ = c;
        s += 2;
      }
      else
      {
        return -1;
      }
    }
    else
    {
      *d++ = *s;
    }
  }
  *d = '\0';
  return 0;
}

enum UrlScheme url_check_scheme(const char *str)
{
  return get_scheme(str, mutt_prex_capture(PREX_URL, str));
}

struct Url *url_parse(const char *src)
{
  const regmatch_t *match = mutt_prex_capture(PREX_URL, src);
  if (!match)
    return NULL;
 
  enum UrlScheme scheme = get_scheme(src, match);
  if (scheme == U_UNKNOWN)
    return NULL;
 
  const regmatch_t *userinfo = &match[PREX_URL_MATCH_USERINFO];
  const regmatch_t *user = &match[PREX_URL_MATCH_USER];
  const regmatch_t *pass = &match[PREX_URL_MATCH_PASS];
  const regmatch_t *host = &match[PREX_URL_MATCH_HOSTNAME];
  const regmatch_t *ipvx = &match[PREX_URL_MATCH_HOSTIPVX];
  const regmatch_t *port = &match[PREX_URL_MATCH_PORT];
  const regmatch_t *path = &match[PREX_URL_MATCH_PATH];
  const regmatch_t *query = &match[PREX_URL_MATCH_QUERY];
  const regmatch_t *pathonly = &match[PREX_URL_MATCH_PATH_ONLY];
 
  struct Url *url = url_new();
  url->scheme = scheme;
  url->src = mutt_str_dup(src);
 
  /* If the scheme is not followed by two forward slashes, then it's a simple
   * path (see https://tools.ietf.org/html/rfc3986#section-3). */
  if (mutt_regmatch_start(pathonly) != -1)
  {
    url->src[mutt_regmatch_end(pathonly)] = '\0';
    url->path = url->src + mutt_regmatch_start(pathonly);
    if (url_pct_decode(url->path) < 0)
      goto err;
  }
 
  /* separate userinfo part */
  if (mutt_regmatch_end(userinfo) != -1)
  {
    url->src[mutt_regmatch_end(userinfo) - 1] = '\0';
  }
 
  /* user */
  if (mutt_regmatch_end(user) != -1)
  {
    url->src[mutt_regmatch_end(user)] = '\0';
    url->user = url->src + mutt_regmatch_start(user);
    if (url_pct_decode(url->user) < 0)
      goto err;
  }
 
  /* pass */
  if (mutt_regmatch_end(pass) != -1)
  {
    url->pass = url->src + mutt_regmatch_start(pass);
    if (url_pct_decode(url->pass) < 0)
      goto err;
  }
 
  /* host */
  if (mutt_regmatch_len(host) != 0)
  {
    url->host = url->src + mutt_regmatch_start(host);
    url->src[mutt_regmatch_end(host)] = '\0';
  }
  else if (mutt_regmatch_end(ipvx) != -1)
  {
    url->host = url->src + mutt_regmatch_start(ipvx) + 1; /* skip opening '[' */
    url->src[mutt_regmatch_end(ipvx) - 1] = '\0';         /* skip closing ']' */
  }
 
  /* port */
  if (mutt_regmatch_end(port) != -1)
  {
    url->src[mutt_regmatch_end(port)] = '\0';
    const char *ports = url->src + mutt_regmatch_start(port);
    unsigned short num;
    if (!mutt_str_atous_full(ports, &num))
    {
      goto err;
    }
    url->port = num;
  }
 
  /* path */
  if (mutt_regmatch_end(path) != -1)
  {
    url->src[mutt_regmatch_end(path)] = '\0';
    url->path = url->src + mutt_regmatch_start(path);
    if (!url->host)
    {
      /* If host is not provided, restore the '/': this is an absolute path */
      *(--url->path) = '/';
    }
    if (url_pct_decode(url->path) < 0)
      goto err;
  }
 
  /* query */
  if (mutt_regmatch_end(query) != -1)
  {
    char *squery = url->src + mutt_regmatch_start(query);
    if (!parse_query_string(&url->query_strings, squery))
      goto err;
  }
 
  return url;
 
err:
  url_free(&url);
  return NULL;
}

int url_tobuffer(const struct Url *url, struct Buffer *buf, uint8_t flags)
{
  if (!url || !buf)
    return -1;
  if (url->scheme == U_UNKNOWN)
    return -1;
 
  buf_printf(buf, "%s:", mutt_map_get_name(url->scheme, UrlMap));
 
  if (url->host)
  {
    if (!(flags & U_PATH))
      buf_addstr(buf, "//");
 
    if (url->user && (url->user[0] || !(flags & U_PATH)))
    {
      char str[256] = { 0 };
      url_pct_encode(str, sizeof(str), url->user);
      buf_add_printf(buf, "%s@", str);
    }
 
    if (strchr(url->host, ':'))
      buf_add_printf(buf, "[%s]", url->host);
    else
      buf_addstr(buf, url->host);
 
    if (url->port)
      buf_add_printf(buf, ":%hu/", url->port);
    else
      buf_addstr(buf, "/");
  }
 
  if (url->path)
    buf_addstr(buf, url->path);
 
  if (STAILQ_FIRST(&url->query_strings))
  {
    buf_addstr(buf, "?");
 
    char str[256] = { 0 };
    struct UrlQuery *np = NULL;
    STAILQ_FOREACH(np, &url->query_strings, entries)
    {
      url_pct_encode(str, sizeof(str), np->name);
      buf_addstr(buf, str);
      buf_addstr(buf, "=");
      url_pct_encode(str, sizeof(str), np->value);
      buf_addstr(buf, str);
      if (STAILQ_NEXT(np, entries))
        buf_addstr(buf, "&");
    }
  }
 
  return 0;
}

int url_tostring(const struct Url *url, char *dest, size_t len, uint8_t flags)
{
  if (!url || !dest)
    return -1;
 
  struct Buffer *dest_buf = buf_pool_get();
 
  int rc = url_tobuffer(url, dest_buf, flags);
  if (rc == 0)
    mutt_str_copy(dest, buf_string(dest_buf), len);
 
  buf_pool_release(&dest_buf);
 
  return rc;
}