[svn] Attempt to quote '?' as "%3F" when linking to local files.

[wget] / src / url.c

/* URL handling.
   Copyright (C) 1995, 1996, 1997, 2000, 2001 Free Software Foundation, Inc.

This file is part of GNU Wget.

GNU Wget is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 2 of the License, or (at
your option) any later version.

GNU Wget is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
GNU General Public License for more details.

You should have received a copy of the GNU General Public License
along with Wget; if not, write to the Free Software
Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.  */

#include <config.h>

#include <stdio.h>
#include <stdlib.h>
#ifdef HAVE_STRING_H
# include <string.h>
#else
# include <strings.h>
#endif
#include <sys/types.h>
#ifdef HAVE_UNISTD_H
# include <unistd.h>
#endif
#include <errno.h>
#include <assert.h>

#include "wget.h"
#include "utils.h"
#include "url.h"
#include "host.h"
#include "hash.h"

#ifndef errno
extern int errno;
#endif

/* Is X "."?  */
#define DOTP(x) ((*(x) == '.') && (!*(x + 1)))
/* Is X ".."?  */
#define DDOTP(x) ((*(x) == '.') && (*(x + 1) == '.') && (!*(x + 2)))

static int urlpath_length PARAMS ((const char *));

struct scheme_data
{
  char *leading_string;
  int default_port;
};

/* Supported schemes: */
static struct scheme_data supported_schemes[] =
{
  { "http://",  DEFAULT_HTTP_PORT },
#ifdef HAVE_SSL
  { "https://", DEFAULT_HTTPS_PORT },
#endif
  { "ftp://",   DEFAULT_FTP_PORT },

  /* SCHEME_INVALID */
  { NULL,       -1 }
};

static char *construct_relative PARAMS ((const char *, const char *));

\f
/* Support for encoding and decoding of URL strings.  We determine
   whether a character is unsafe through static table lookup.  This
   code assumes ASCII character set and 8-bit chars.  */

enum {
  urlchr_reserved = 1,
  urlchr_unsafe   = 2
};

#define R  urlchr_reserved
#define U  urlchr_unsafe
#define RU R|U

#define urlchr_test(c, mask) (urlchr_table[(unsigned char)(c)] & (mask))

/* rfc1738 reserved chars, preserved from encoding.  */

#define RESERVED_CHAR(c) urlchr_test(c, urlchr_reserved)

/* rfc1738 unsafe chars, plus some more.  */

#define UNSAFE_CHAR(c) urlchr_test(c, urlchr_unsafe)

const static unsigned char urlchr_table[256] =
{
  U,  U,  U,  U,   U,  U,  U,  U,   /* NUL SOH STX ETX  EOT ENQ ACK BEL */
  U,  U,  U,  U,   U,  U,  U,  U,   /* BS  HT  LF  VT   FF  CR  SO  SI  */
  U,  U,  U,  U,   U,  U,  U,  U,   /* DLE DC1 DC2 DC3  DC4 NAK SYN ETB */
  U,  U,  U,  U,   U,  U,  U,  U,   /* CAN EM  SUB ESC  FS  GS  RS  US  */
  U,  0,  U, RU,   0,  U,  R,  0,   /* SP  !   "   #    $   %   &   '   */
  0,  0,  0,  R,   0,  0,  0,  R,   /* (   )   *   +    ,   -   .   /   */
  0,  0,  0,  0,   0,  0,  0,  0,   /* 0   1   2   3    4   5   6   7   */
  0,  0, RU,  R,   U,  R,  U,  R,   /* 8   9   :   ;    <   =   >   ?   */
 RU,  0,  0,  0,   0,  0,  0,  0,   /* @   A   B   C    D   E   F   G   */
  0,  0,  0,  0,   0,  0,  0,  0,   /* H   I   J   K    L   M   N   O   */
  0,  0,  0,  0,   0,  0,  0,  0,   /* P   Q   R   S    T   U   V   W   */
  0,  0,  0,  U,   U,  U,  U,  0,   /* X   Y   Z   [    \   ]   ^   _   */
  U,  0,  0,  0,   0,  0,  0,  0,   /* `   a   b   c    d   e   f   g   */
  0,  0,  0,  0,   0,  0,  0,  0,   /* h   i   j   k    l   m   n   o   */
  0,  0,  0,  0,   0,  0,  0,  0,   /* p   q   r   s    t   u   v   w   */
  0,  0,  0,  U,   U,  U,  U,  U,   /* x   y   z   {    |   }   ~   DEL */

  U, U, U, U,  U, U, U, U,  U, U, U, U,  U, U, U, U,
  U, U, U, U,  U, U, U, U,  U, U, U, U,  U, U, U, U,
  U, U, U, U,  U, U, U, U,  U, U, U, U,  U, U, U, U,
  U, U, U, U,  U, U, U, U,  U, U, U, U,  U, U, U, U,

  U, U, U, U,  U, U, U, U,  U, U, U, U,  U, U, U, U,
  U, U, U, U,  U, U, U, U,  U, U, U, U,  U, U, U, U,
  U, U, U, U,  U, U, U, U,  U, U, U, U,  U, U, U, U,
  U, U, U, U,  U, U, U, U,  U, U, U, U,  U, U, U, U,
};

/* Decodes the forms %xy in a URL to the character the hexadecimal
   code of which is xy.  xy are hexadecimal digits from
   [0123456789ABCDEF] (case-insensitive).  If x or y are not
   hex-digits or `%' precedes `\0', the sequence is inserted
   literally.  */

static void
decode_string (char *s)
{
  char *t = s;                  /* t - tortoise */
  char *h = s;                  /* h - hare     */

  for (; *h; h++, t++)
    {
      if (*h != '%')
        {
        copychar:
          *t = *h;
        }
      else
        {
          /* Do nothing if '%' is not followed by two hex digits. */
          if (!*(h + 1) || !*(h + 2)
              || !(ISXDIGIT (*(h + 1)) && ISXDIGIT (*(h + 2))))
            goto copychar;
          *t = (XCHAR_TO_XDIGIT (*(h + 1)) << 4) + XCHAR_TO_XDIGIT (*(h + 2));
          h += 2;
        }
    }
  *t = '\0';
}

/* Like encode_string, but return S if there are no unsafe chars.  */

static char *
encode_string_maybe (const char *s)
{
  const char *p1;
  char *p2, *newstr;
  int newlen;
  int addition = 0;

  for (p1 = s; *p1; p1++)
    if (UNSAFE_CHAR (*p1))
      addition += 2;            /* Two more characters (hex digits) */

  if (!addition)
    return (char *)s;

  newlen = (p1 - s) + addition;
  newstr = (char *)xmalloc (newlen + 1);

  p1 = s;
  p2 = newstr;
  while (*p1)
    {
      if (UNSAFE_CHAR (*p1))
        {
          unsigned char c = *p1++;
          *p2++ = '%';
          *p2++ = XDIGIT_TO_XCHAR (c >> 4);
          *p2++ = XDIGIT_TO_XCHAR (c & 0xf);
        }
      else
        *p2++ = *p1++;
    }
  *p2 = '\0';
  assert (p2 - newstr == newlen);

  return newstr;
}

/* Encode the unsafe characters (as determined by UNSAFE_CHAR) in a
   given string, returning a malloc-ed %XX encoded string.  */
  
char *
encode_string (const char *s)
{
  char *encoded = encode_string_maybe (s);
  if (encoded != s)
    return encoded;
  else
    return xstrdup (s);
}

/* Encode unsafe characters in PTR to %xx.  If such encoding is done,
   the old value of PTR is freed and PTR is made to point to the newly
   allocated storage.  */

#define ENCODE(ptr) do {                        \
  char *e_new = encode_string_maybe (ptr);      \
  if (e_new != ptr)                             \
    {                                           \
      xfree (ptr);                              \
      ptr = e_new;                              \
    }                                           \
} while (0)
\f
enum copy_method { CM_DECODE, CM_ENCODE, CM_PASSTHROUGH };

/* Decide whether to encode, decode, or pass through the char at P.
   This used to be a macro, but it got a little too convoluted.  */
static inline enum copy_method
decide_copy_method (const char *p)
{
  if (*p == '%')
    {
      if (ISXDIGIT (*(p + 1)) && ISXDIGIT (*(p + 2)))
        {
          /* %xx sequence: decode it, unless it would decode to an
             unsafe or a reserved char; in that case, leave it as
             is. */
          char preempt = (XCHAR_TO_XDIGIT (*(p + 1)) << 4) +
            XCHAR_TO_XDIGIT (*(p + 2));

          if (UNSAFE_CHAR (preempt) || RESERVED_CHAR (preempt))
            return CM_PASSTHROUGH;
          else
            return CM_DECODE;
        }
      else
        /* Garbled %.. sequence: encode `%'. */
        return CM_ENCODE;
    }
  else if (UNSAFE_CHAR (*p) && !RESERVED_CHAR (*p))
    return CM_ENCODE;
  else
    return CM_PASSTHROUGH;
}

/* Translate a %-quoting (but possibly non-conformant) input string S
   into a %-quoting (and conformant) output string.  If no characters
   are encoded or decoded, return the same string S; otherwise, return
   a freshly allocated string with the new contents.

   After a URL has been run through this function, the protocols that
   use `%' as the quote character can use the resulting string as-is,
   while those that don't call decode_string() to get to the intended
   data.  This function is also stable: after an input string is
   transformed the first time, all further transformations of the
   result yield the same result string.

   Let's discuss why this function is needed.

   Imagine Wget is to retrieve `http://abc.xyz/abc def'.  Since a raw
   space character would mess up the HTTP request, it needs to be
   quoted, like this:

       GET /abc%20def HTTP/1.0

   So it appears that the unsafe chars need to be quoted, as with
   encode_string.  But what if we're requested to download
   `abc%20def'?  Remember that %-encoding is valid URL syntax, so what
   the user meant was a literal space, and he was kind enough to quote
   it.  In that case, Wget should obviously leave the `%20' as is, and
   send the same request as above.  So in this case we may not call
   encode_string.

   But what if the requested URI is `abc%20 def'?  If we call
   encode_string, we end up with `/abc%2520%20def', which is almost
   certainly not intended.  If we don't call encode_string, we are
   left with the embedded space and cannot send the request.  What the
   user meant was for Wget to request `/abc%20%20def', and this is
   where reencode_string kicks in.

   Wget used to solve this by first decoding %-quotes, and then
   encoding all the "unsafe" characters found in the resulting string.
   This was wrong because it didn't preserve certain URL special
   (reserved) characters.  For instance, URI containing "a%2B+b" (0x2b
   == '+') would get translated to "a%2B%2Bb" or "a++b" depending on
   whether we considered `+' reserved (it is).  One of these results
   is inevitable because by the second step we would lose information
   on whether the `+' was originally encoded or not.  Both results
   were wrong because in CGI parameters + means space, while %2B means
   literal plus.  reencode_string correctly translates the above to
   "a%2B+b", i.e. returns the original string.

   This function uses an algorithm proposed by Anon Sricharoenchai:

   1. Encode all URL_UNSAFE and the "%" that are not followed by 2
      hexdigits.

   2. Decode all "%XX" except URL_UNSAFE, URL_RESERVED (";/?:@=&") and
      "+".

   ...except that this code conflates the two steps, and decides
   whether to encode, decode, or pass through each character in turn.
   The function still uses two passes, but their logic is the same --
   the first pass exists merely for the sake of allocation.  Another
   small difference is that we include `+' to URL_RESERVED.

   Anon's test case:

   "http://abc.xyz/%20%3F%%36%31%25aa% a?a=%61+a%2Ba&b=b%26c%3Dc"
   ->
   "http://abc.xyz/%20%3F%2561%25aa%25%20a?a=a+a%2Ba&b=b%26c%3Dc"

   Simpler test cases:

   "foo bar"         -> "foo%20bar"
   "foo%20bar"       -> "foo%20bar"
   "foo %20bar"      -> "foo%20%20bar"
   "foo%%20bar"      -> "foo%25%20bar"       (0x25 == '%')
   "foo%25%20bar"    -> "foo%25%20bar"
   "foo%2%20bar"     -> "foo%252%20bar"
   "foo+bar"         -> "foo+bar"            (plus is reserved!)
   "foo%2b+bar"      -> "foo%2b+bar"  */

char *
reencode_string (const char *s)
{
  const char *p1;
  char *newstr, *p2;
  int oldlen, newlen;

  int encode_count = 0;
  int decode_count = 0;

  /* First, pass through the string to see if there's anything to do,
     and to calculate the new length.  */
  for (p1 = s; *p1; p1++)
    {
      switch (decide_copy_method (p1))
        {
        case CM_ENCODE:
          ++encode_count;
          break;
        case CM_DECODE:
          ++decode_count;
          break;
        case CM_PASSTHROUGH:
          break;
        }
    }

  if (!encode_count && !decode_count)
    /* The string is good as it is. */
    return (char *)s;           /* C const model sucks. */

  oldlen = p1 - s;
  /* Each encoding adds two characters (hex digits), while each
     decoding removes two characters.  */
  newlen = oldlen + 2 * (encode_count - decode_count);
  newstr = xmalloc (newlen + 1);

  p1 = s;
  p2 = newstr;

  while (*p1)
    {
      switch (decide_copy_method (p1))
        {
        case CM_ENCODE:
          {
            unsigned char c = *p1++;
            *p2++ = '%';
            *p2++ = XDIGIT_TO_XCHAR (c >> 4);
            *p2++ = XDIGIT_TO_XCHAR (c & 0xf);
          }
          break;
        case CM_DECODE:
          *p2++ = ((XCHAR_TO_XDIGIT (*(p1 + 1)) << 4)
                   + (XCHAR_TO_XDIGIT (*(p1 + 2))));
          p1 += 3;              /* skip %xx */
          break;
        case CM_PASSTHROUGH:
          *p2++ = *p1++;
        }
    }
  *p2 = '\0';
  assert (p2 - newstr == newlen);
  return newstr;
}

/* Run PTR_VAR through reencode_string.  If a new string is consed,
   free PTR_VAR and make it point to the new storage.  Obviously,
   PTR_VAR needs to be an lvalue.  */

#define REENCODE(ptr_var) do {                  \
  char *rf_new = reencode_string (ptr_var);     \
  if (rf_new != ptr_var)                        \
    {                                           \
      xfree (ptr_var);                          \
      ptr_var = rf_new;                         \
    }                                           \
} while (0)
\f
/* Returns the scheme type if the scheme is supported, or
   SCHEME_INVALID if not.  */
enum url_scheme
url_scheme (const char *url)
{
  int i;

  for (i = 0; supported_schemes[i].leading_string; i++)
    if (!strncasecmp (url, supported_schemes[i].leading_string,
                      strlen (supported_schemes[i].leading_string)))
      return (enum url_scheme)i;
  return SCHEME_INVALID;
}

/* Return the number of characters needed to skip the scheme part of
   the URL, e.g. `http://'.  If no scheme is found, returns 0.  */
int
url_skip_scheme (const char *url)
{
  const char *p = url;

  /* Skip the scheme name.  We allow `-' and `+' because of `whois++',
     etc. */
  while (ISALNUM (*p) || *p == '-' || *p == '+')
    ++p;
  if (*p != ':')
    return 0;
  /* Skip ':'. */
  ++p;

  /* Skip "//" if found. */
  if (*p == '/' && *(p + 1) == '/')
    p += 2;

  return p - url;
}

/* Returns 1 if the URL begins with a scheme (supported or
   unsupported), 0 otherwise.  */
int
url_has_scheme (const char *url)
{
  const char *p = url;
  while (ISALNUM (*p) || *p == '-' || *p == '+')
    ++p;
  return *p == ':';
}

int
scheme_default_port (enum url_scheme scheme)
{
  return supported_schemes[scheme].default_port;
}

/* Skip the username and password, if present here.  The function
   should be called *not* with the complete URL, but with the part
   right after the scheme.

   If no username and password are found, return 0.  */
int
url_skip_uname (const char *url)
{
  const char *p;

  /* Look for '@' that comes before '/' or '?'. */
  p = (const char *)strpbrk (url, "/?@");
  if (!p || *p != '@')
    return 0;

  return p - url + 1;
}

static int
parse_uname (const char *str, int len, char **user, char **passwd)
{
  char *colon;

  if (len == 0)
    /* Empty user name not allowed. */
    return 0;

  colon = memchr (str, ':', len);
  if (colon == str)
    /* Empty user name again. */
    return 0;

  if (colon)
    {
      int pwlen = len - (colon + 1 - str);
      *passwd = xmalloc (pwlen + 1);
      memcpy (*passwd, colon + 1, pwlen);
      (*passwd)[pwlen] = '\0';
      len -= pwlen + 1;
    }
  else
    *passwd = NULL;

  *user = xmalloc (len + 1);
  memcpy (*user, str, len);
  (*user)[len] = '\0';

  return 1;
}

/* Used by main.c: detect URLs written using the "shorthand" URL forms
   popularized by Netscape and NcFTP.  HTTP shorthands look like this:

   www.foo.com[:port]/dir/file   -> http://www.foo.com[:port]/dir/file
   www.foo.com[:port]            -> http://www.foo.com[:port]

   FTP shorthands look like this:

   foo.bar.com:dir/file          -> ftp://foo.bar.com/dir/file
   foo.bar.com:/absdir/file      -> ftp://foo.bar.com//absdir/file

   If the URL needs not or cannot be rewritten, return NULL.  */
char *
rewrite_shorthand_url (const char *url)
{
  const char *p;

  if (url_has_scheme (url))
    return NULL;

  /* Look for a ':' or '/'.  The former signifies NcFTP syntax, the
     latter Netscape.  */
  for (p = url; *p && *p != ':' && *p != '/'; p++)
    ;

  if (p == url)
    return NULL;

  if (*p == ':')
    {
      const char *pp, *path;
      char *res;
      /* If the characters after the colon and before the next slash
         or end of string are all digits, it's HTTP.  */
      int digits = 0;
      for (pp = p + 1; ISDIGIT (*pp); pp++)
        ++digits;
      if (digits > 0
          && (*pp == '/' || *pp == '\0'))
        goto http;

      /* Prepend "ftp://" to the entire URL... */
      path = p + 1;
      res = xmalloc (6 + strlen (url) + 1);
      sprintf (res, "ftp://%s", url);
      /* ...and replace ':' with '/'. */
      res[6 + (p - url)] = '/';
      return res;
    }
  else
    {
      char *res;
    http:
      /* Just prepend "http://" to what we have. */
      res = xmalloc (7 + strlen (url) + 1);
      sprintf (res, "http://%s", url);
      return res;
    }
}
\f
static void parse_path PARAMS ((const char *, char **, char **));

static char *
strpbrk_or_eos (const char *s, const char *accept)
{
  char *p = strpbrk (s, accept);
  if (!p)
    p = (char *)s + strlen (s);
  return p;
}

/* Turn STR into lowercase; return non-zero if a character was
   actually changed. */

static int
lowercase_str (char *str)
{
  int change = 0;
  for (; *str; str++)
    if (!ISLOWER (*str))
      {
        change = 1;
        *str = TOLOWER (*str);
      }
  return change;
}

static char *parse_errors[] = {
#define PE_NO_ERROR            0
  "No error",
#define PE_UNRECOGNIZED_SCHEME 1
  "Unrecognized scheme",
#define PE_EMPTY_HOST          2
  "Empty host",
#define PE_BAD_PORT_NUMBER     3
  "Bad port number",
#define PE_INVALID_USER_NAME   4
  "Invalid user name"
};

#define SETERR(p, v) do {                       \
  if (p)                                        \
    *(p) = (v);                                 \
} while (0)

/* Parse a URL.

   Return a new struct url if successful, NULL on error.  In case of
   error, and if ERROR is not NULL, also set *ERROR to the appropriate
   error code. */
struct url *
url_parse (const char *url, int *error)
{
  struct url *u;
  const char *p;
  int path_modified, host_modified;

  enum url_scheme scheme;

  const char *uname_b,     *uname_e;
  const char *host_b,      *host_e;
  const char *path_b,      *path_e;
  const char *params_b,    *params_e;
  const char *query_b,     *query_e;
  const char *fragment_b,  *fragment_e;

  int port;
  char *user = NULL, *passwd = NULL;

  char *url_encoded;

  scheme = url_scheme (url);
  if (scheme == SCHEME_INVALID)
    {
      SETERR (error, PE_UNRECOGNIZED_SCHEME);
      return NULL;
    }

  url_encoded = reencode_string (url);
  p = url_encoded;

  p += strlen (supported_schemes[scheme].leading_string);
  uname_b = p;
  p += url_skip_uname (p);
  uname_e = p;

  /* scheme://user:pass@host[:port]... */
  /*                    ^              */

  /* We attempt to break down the URL into the components path,
     params, query, and fragment.  They are ordered like this:

       scheme://host[:port][/path][;params][?query][#fragment]  */

  params_b   = params_e   = NULL;
  query_b    = query_e    = NULL;
  fragment_b = fragment_e = NULL;

  host_b = p;
  p = strpbrk_or_eos (p, ":/;?#");
  host_e = p;

  if (host_b == host_e)
    {
      SETERR (error, PE_EMPTY_HOST);
      return NULL;
    }

  port = scheme_default_port (scheme);
  if (*p == ':')
    {
      const char *port_b, *port_e, *pp;

      /* scheme://host:port/tralala */
      /*              ^             */
      ++p;
      port_b = p;
      p = strpbrk_or_eos (p, "/;?#");
      port_e = p;

      if (port_b == port_e)
        {
          /* http://host:/whatever */
          /*             ^         */
          SETERR (error, PE_BAD_PORT_NUMBER);
          return NULL;
        }

      for (port = 0, pp = port_b; pp < port_e; pp++)
        {
          if (!ISDIGIT (*pp))
            {
              /* http://host:12randomgarbage/blah */
              /*               ^                  */
              SETERR (error, PE_BAD_PORT_NUMBER);
              return NULL;
            }
          port = 10 * port + (*pp - '0');
        }
    }

  if (*p == '/')
    {
      ++p;
      path_b = p;
      p = strpbrk_or_eos (p, ";?#");
      path_e = p;
    }
  else
    {
      /* Path is not allowed not to exist. */
      path_b = path_e = p;
    }

  if (*p == ';')
    {
      ++p;
      params_b = p;
      p = strpbrk_or_eos (p, "?#");
      params_e = p;
    }
  if (*p == '?')
    {
      ++p;
      query_b = p;
      p = strpbrk_or_eos (p, "#");
      query_e = p;
    }
  if (*p == '#')
    {
      ++p;
      fragment_b = p;
      p += strlen (p);
      fragment_e = p;
    }
  assert (*p == 0);

  if (uname_b != uname_e)
    {
      /* http://user:pass@host */
      /*        ^         ^    */
      /*     uname_b   uname_e */
      if (!parse_uname (uname_b, uname_e - uname_b - 1, &user, &passwd))
        {
          SETERR (error, PE_INVALID_USER_NAME);
          return NULL;
        }
    }

  u = (struct url *)xmalloc (sizeof (struct url));
  memset (u, 0, sizeof (*u));

  u->scheme = scheme;
  u->host   = strdupdelim (host_b, host_e);
  u->port   = port;
  u->user   = user;
  u->passwd = passwd;

  u->path = strdupdelim (path_b, path_e);
  path_modified = path_simplify (u->path);
  parse_path (u->path, &u->dir, &u->file);

  host_modified = lowercase_str (u->host);

  if (params_b)
    u->params = strdupdelim (params_b, params_e);
  if (query_b)
    u->query = strdupdelim (query_b, query_e);
  if (fragment_b)
    u->fragment = strdupdelim (fragment_b, fragment_e);


  if (path_modified || u->fragment || host_modified)
    {
      /* If path_simplify modified the path, or if a fragment is
         present, or if the original host name had caps in it, make
         sure that u->url is equivalent to what would be printed by
         url_string.  */
      u->url = url_string (u, 0);

      if (url_encoded != url)
        xfree ((char *) url_encoded);
    }
  else
    {
      if (url_encoded == url)
        u->url    = xstrdup (url);
      else
        u->url    = url_encoded;
    }
  url_encoded = NULL;

  return u;
}

const char *
url_error (int error_code)
{
  assert (error_code >= 0 && error_code < ARRAY_SIZE (parse_errors));
  return parse_errors[error_code];
}

static void
parse_path (const char *quoted_path, char **dir, char **file)
{
  char *path, *last_slash;

  STRDUP_ALLOCA (path, quoted_path);
  decode_string (path);

  last_slash = strrchr (path, '/');
  if (!last_slash)
    {
      *dir = xstrdup ("");
      *file = xstrdup (path);
    }
  else
    {
      *dir = strdupdelim (path, last_slash);
      *file = xstrdup (last_slash + 1);
    }
}

/* Note: URL's "full path" is the path with the query string and
   params appended.  The "fragment" (#foo) is intentionally ignored,
   but that might be changed.  For example, if the original URL was
   "http://host:port/foo/bar/baz;bullshit?querystring#uselessfragment",
   the full path will be "/foo/bar/baz;bullshit?querystring".  */

/* Return the length of the full path, without the terminating
   zero.  */

static int
full_path_length (const struct url *url)
{
  int len = 0;

#define FROB(el) if (url->el) len += 1 + strlen (url->el)

  FROB (path);
  FROB (params);
  FROB (query);

#undef FROB

  return len;
}

/* Write out the full path. */

static void
full_path_write (const struct url *url, char *where)
{
#define FROB(el, chr) do {                      \
  char *f_el = url->el;                         \
  if (f_el) {                                   \
    int l = strlen (f_el);                      \
    *where++ = chr;                             \
    memcpy (where, f_el, l);                    \
    where += l;                                 \
  }                                             \
} while (0)

  FROB (path, '/');
  FROB (params, ';');
  FROB (query, '?');

#undef FROB
}

/* Public function for getting the "full path". */
char *
url_full_path (const struct url *url)
{
  int length = full_path_length (url);
  char *full_path = (char *)xmalloc(length + 1);

  full_path_write (url, full_path);
  full_path[length] = '\0';

  return full_path;
}

/* Sync u->path and u->url with u->dir and u->file. */
static void
sync_path (struct url *url)
{
  char *newpath;

  xfree (url->path);

  if (!*url->dir)
    {
      newpath = xstrdup (url->file);
      REENCODE (newpath);
    }
  else
    {
      int dirlen = strlen (url->dir);
      int filelen = strlen (url->file);

      newpath = xmalloc (dirlen + 1 + filelen + 1);
      memcpy (newpath, url->dir, dirlen);
      newpath[dirlen] = '/';
      memcpy (newpath + dirlen + 1, url->file, filelen);
      newpath[dirlen + 1 + filelen] = '\0';
      REENCODE (newpath);
    }

  url->path = newpath;

  /* Synchronize u->url. */
  xfree (url->url);
  url->url = url_string (url, 0);
}

/* Mutators.  Code in ftp.c insists on changing u->dir and u->file.
   This way we can sync u->path and u->url when they get changed.  */

void
url_set_dir (struct url *url, const char *newdir)
{
  xfree (url->dir);
  url->dir = xstrdup (newdir);
  sync_path (url);
}

void
url_set_file (struct url *url, const char *newfile)
{
  xfree (url->file);
  url->file = xstrdup (newfile);
  sync_path (url);
}

void
url_free (struct url *url)
{
  xfree (url->host);
  xfree (url->path);
  xfree (url->url);

  FREE_MAYBE (url->params);
  FREE_MAYBE (url->query);
  FREE_MAYBE (url->fragment);
  FREE_MAYBE (url->user);
  FREE_MAYBE (url->passwd);

  xfree (url->dir);
  xfree (url->file);

  xfree (url);
}
\f
struct urlpos *
get_urls_file (const char *file)
{
  struct file_memory *fm;
  struct urlpos *head, *tail;
  const char *text, *text_end;

  /* Load the file.  */
  fm = read_file (file);
  if (!fm)
    {
      logprintf (LOG_NOTQUIET, "%s: %s\n", file, strerror (errno));
      return NULL;
    }
  DEBUGP (("Loaded %s (size %ld).\n", file, fm->length));
  head = tail = NULL;
  text = fm->content;
  text_end = fm->content + fm->length;
  while (text < text_end)
    {
      const char *line_beg = text;
      const char *line_end = memchr (text, '\n', text_end - text);
      if (!line_end)
        line_end = text_end;
      else
        ++line_end;
      text = line_end;
      while (line_beg < line_end
             && ISSPACE (*line_beg))
        ++line_beg;
      while (line_end > line_beg + 1
             && ISSPACE (*(line_end - 1)))
        --line_end;
      if (line_end > line_beg)
        {
          int up_error_code;
          char *url_text;
          struct urlpos *entry;
          struct url *url;

          /* We must copy the URL to a zero-terminated string.  *sigh*.  */
          url_text = strdupdelim (line_beg, line_end);
          url = url_parse (url_text, &up_error_code);
          if (!url)
            {
              logprintf (LOG_NOTQUIET, "%s: Invalid URL %s: %s\n",
                         file, url_text, url_error (up_error_code));
              xfree (url_text);
              continue;
            }
          xfree (url_text);

          entry = (struct urlpos *)xmalloc (sizeof (struct urlpos));
          memset (entry, 0, sizeof (*entry));
          entry->next = NULL;
          entry->url = url;

          if (!head)
            head = entry;
          else
            tail->next = entry;
          tail = entry;
        }
    }
  read_file_free (fm);
  return head;
}
\f
/* Free the linked list of urlpos.  */
void
free_urlpos (struct urlpos *l)
{
  while (l)
    {
      struct urlpos *next = l->next;
      if (l->url)
        url_free (l->url);
      FREE_MAYBE (l->local_name);
      xfree (l);
      l = next;
    }
}

/* Rotate FNAME opt.backups times */
void
rotate_backups(const char *fname)
{
  int maxlen = strlen (fname) + 1 + numdigit (opt.backups) + 1;
  char *from = (char *)alloca (maxlen);
  char *to = (char *)alloca (maxlen);
  struct stat sb;
  int i;

  if (stat (fname, &sb) == 0)
    if (S_ISREG (sb.st_mode) == 0)
      return;

  for (i = opt.backups; i > 1; i--)
    {
      sprintf (from, "%s.%d", fname, i - 1);
      sprintf (to, "%s.%d", fname, i);
      /* #### This will fail on machines without the rename() system
         call.  */
      rename (from, to);
    }

  sprintf (to, "%s.%d", fname, 1);
  rename(fname, to);
}

/* Create all the necessary directories for PATH (a file).  Calls
   mkdirhier() internally.  */
int
mkalldirs (const char *path)
{
  const char *p;
  char *t;
  struct stat st;
  int res;

  p = path + strlen (path);
  for (; *p != '/' && p != path; p--);
  /* Don't create if it's just a file.  */
  if ((p == path) && (*p != '/'))
    return 0;
  t = strdupdelim (path, p);
  /* Check whether the directory exists.  */
  if ((stat (t, &st) == 0))
    {
      if (S_ISDIR (st.st_mode))
        {
          xfree (t);
          return 0;
        }
      else
        {
          /* If the dir exists as a file name, remove it first.  This
             is *only* for Wget to work with buggy old CERN http
             servers.  Here is the scenario: When Wget tries to
             retrieve a directory without a slash, e.g.
             http://foo/bar (bar being a directory), CERN server will
             not redirect it too http://foo/bar/ -- it will generate a
             directory listing containing links to bar/file1,
             bar/file2, etc.  Wget will lose because it saves this
             HTML listing to a file `bar', so it cannot create the
             directory.  To work around this, if the file of the same
             name exists, we just remove it and create the directory
             anyway.  */
          DEBUGP (("Removing %s because of directory danger!\n", t));
          unlink (t);
        }
    }
  res = make_directory (t);
  if (res != 0)
    logprintf (LOG_NOTQUIET, "%s: %s", t, strerror (errno));
  xfree (t);
  return res;
}

static int
count_slashes (const char *s)
{
  int i = 0;
  while (*s)
    if (*s++ == '/')
      ++i;
  return i;
}

/* Return the path name of the URL-equivalent file name, with a
   remote-like structure of directories.  */
static char *
mkstruct (const struct url *u)
{
  char *dir, *dir_preencoding;
  char *file, *res, *dirpref;
  char *query = u->query && *u->query ? u->query : NULL;
  int l;

  if (opt.cut_dirs)
    {
      char *ptr = u->dir + (*u->dir == '/');
      int slash_count = 1 + count_slashes (ptr);
      int cut = MINVAL (opt.cut_dirs, slash_count);
      for (; cut && *ptr; ptr++)
        if (*ptr == '/')
          --cut;
      STRDUP_ALLOCA (dir, ptr);
    }
  else
    dir = u->dir + (*u->dir == '/');

  /* Check for the true name (or at least a consistent name for saving
     to directory) of HOST, reusing the hlist if possible.  */
  if (opt.add_hostdir)
    {
      /* Add dir_prefix and hostname (if required) to the beginning of
         dir.  */
      dirpref = (char *)alloca (strlen (opt.dir_prefix) + 1
                                + strlen (u->host)
                                + 1 + numdigit (u->port)
                                + 1);
      if (!DOTP (opt.dir_prefix))
        sprintf (dirpref, "%s/%s", opt.dir_prefix, u->host);
      else
        strcpy (dirpref, u->host);

      if (u->port != scheme_default_port (u->scheme))
        {
          int len = strlen (dirpref);
          dirpref[len] = ':';
          long_to_string (dirpref + len + 1, u->port);
        }
    }
  else                          /* not add_hostdir */
    {
      if (!DOTP (opt.dir_prefix))
        dirpref = opt.dir_prefix;
      else
        dirpref = "";
    }

  /* If there is a prefix, prepend it.  */
  if (*dirpref)
    {
      char *newdir = (char *)alloca (strlen (dirpref) + 1 + strlen (dir) + 2);
      sprintf (newdir, "%s%s%s", dirpref, *dir == '/' ? "" : "/", dir);
      dir = newdir;
    }

  dir_preencoding = dir;
  dir = reencode_string (dir_preencoding);

  l = strlen (dir);
  if (l && dir[l - 1] == '/')
    dir[l - 1] = '\0';

  if (!*u->file)
    file = "index.html";
  else
    file = u->file;

  /* Finally, construct the full name.  */
  res = (char *)xmalloc (strlen (dir) + 1 + strlen (file)
                         + (query ? (1 + strlen (query)) : 0)
                         + 1);
  sprintf (res, "%s%s%s", dir, *dir ? "/" : "", file);
  if (query)
    {
      strcat (res, "?");
      strcat (res, query);
    }
  if (dir != dir_preencoding)
    xfree (dir);
  return res;
}

/* Compose a file name out of BASE, an unescaped file name, and QUERY,
   an escaped query string.  The trick is to make sure that unsafe
   characters in BASE are escaped, and that slashes in QUERY are also
   escaped.  */

static char *
compose_file_name (char *base, char *query)
{
  char result[256];
  char *from;
  char *to = result;

  /* Copy BASE to RESULT and encode all unsafe characters.  */
  from = base;
  while (*from && to - result < sizeof (result))
    {
      if (UNSAFE_CHAR (*from))
        {
          unsigned char c = *from++;
          *to++ = '%';
          *to++ = XDIGIT_TO_XCHAR (c >> 4);
          *to++ = XDIGIT_TO_XCHAR (c & 0xf);
        }
      else
        *to++ = *from++;
    }

  if (query && to - result < sizeof (result))
    {
      *to++ = '?';

      /* Copy QUERY to RESULT and encode all '/' characters. */
      from = query;
      while (*from && to - result < sizeof (result))
        {
          if (*from == '/')
            {
              *to++ = '%';
              *to++ = '2';
              *to++ = 'F';
              ++from;
            }
          else
            *to++ = *from++;
        }
    }

  if (to - result < sizeof (result))
    *to = '\0';
  else
    /* Truncate input which is too long, presumably due to a huge
       query string.  */
    result[sizeof (result) - 1] = '\0';

  return xstrdup (result);
}

/* Create a unique filename, corresponding to a given URL.  Calls
   mkstruct if necessary.  Does *not* actually create any directories.  */
char *
url_filename (const struct url *u)
{
  char *file, *name;
  int have_prefix = 0;          /* whether we must prepend opt.dir_prefix */

  if (opt.dirstruct)
    {
      file = mkstruct (u);
      have_prefix = 1;
    }
  else
    {
      char *base = *u->file ? u->file : "index.html";
      char *query = u->query && *u->query ? u->query : NULL;
      file = compose_file_name (base, query);
    }

  if (!have_prefix)
    {
      /* Check whether the prefix directory is something other than "."
         before prepending it.  */
      if (!DOTP (opt.dir_prefix))
        {
          char *nfile = (char *)xmalloc (strlen (opt.dir_prefix)
                                         + 1 + strlen (file) + 1);
          sprintf (nfile, "%s/%s", opt.dir_prefix, file);
          xfree (file);
          file = nfile;
        }
    }
  /* DOS-ish file systems don't like `%' signs in them; we change it
     to `@'.  */
#ifdef WINDOWS
  {
    char *p = file;
    for (p = file; *p; p++)
      if (*p == '%')
        *p = '@';
  }
#endif /* WINDOWS */

  /* Check the cases in which the unique extensions are not used:
     1) Clobbering is turned off (-nc).
     2) Retrieval with regetting.
     3) Timestamping is used.
     4) Hierarchy is built.

     The exception is the case when file does exist and is a
     directory (actually support for bad httpd-s).  */
  if ((opt.noclobber || opt.always_rest || opt.timestamping || opt.dirstruct)
      && !(file_exists_p (file) && !file_non_directory_p (file)))
    return file;

  /* Find a unique name.  */
  name = unique_name (file);
  xfree (file);
  return name;
}

/* Like strlen(), but allow the URL to be ended with '?'.  */
static int
urlpath_length (const char *url)
{
  const char *q = strpbrk_or_eos (url, "?;#");
  return q - url;
}

/* Find the last occurrence of character C in the range [b, e), or
   NULL, if none are present.  This is almost completely equivalent to
   { *e = '\0'; return strrchr(b); }, except that it doesn't change
   the contents of the string.  */
static const char *
find_last_char (const char *b, const char *e, char c)
{
  for (; e > b; e--)
    if (*e == c)
      return e;
  return NULL;
}

/* Resolve the result of "linking" a base URI (BASE) to a
   link-specified URI (LINK).

   Either of the URIs may be absolute or relative, complete with the
   host name, or path only.  This tries to behave "reasonably" in all
   foreseeable cases.  It employs little specific knowledge about
   schemes or URL-specific stuff -- it just works on strings.

   The parameters LINKLENGTH is useful if LINK is not zero-terminated.
   See uri_merge for a gentler interface to this functionality.

   #### This function should handle `./' and `../' so that the evil
   path_simplify can go.  */
static char *
uri_merge_1 (const char *base, const char *link, int linklength, int no_scheme)
{
  char *constr;

  if (no_scheme)
    {
      const char *end = base + urlpath_length (base);

      if (!*link)
        {
          /* Empty LINK points back to BASE, query string and all. */
          constr = xstrdup (base);
        }
      else if (*link == '?')
        {
          /* LINK points to the same location, but changes the query
             string.  Examples: */
          /* uri_merge("path",         "?new") -> "path?new"     */
          /* uri_merge("path?foo",     "?new") -> "path?new"     */
          /* uri_merge("path?foo#bar", "?new") -> "path?new"     */
          /* uri_merge("path#foo",     "?new") -> "path?new"     */
          int baselength = end - base;
          constr = xmalloc (baselength + linklength + 1);
          memcpy (constr, base, baselength);
          memcpy (constr + baselength, link, linklength);
          constr[baselength + linklength] = '\0';
        }
      else if (*link == '#')
        {
          /* uri_merge("path",         "#new") -> "path#new"     */
          /* uri_merge("path#foo",     "#new") -> "path#new"     */
          /* uri_merge("path?foo",     "#new") -> "path?foo#new" */
          /* uri_merge("path?foo#bar", "#new") -> "path?foo#new" */
          int baselength;
          const char *end1 = strchr (base, '#');
          if (!end1)
            end1 = base + strlen (base);
          baselength = end1 - base;
          constr = xmalloc (baselength + linklength + 1);
          memcpy (constr, base, baselength);
          memcpy (constr + baselength, link, linklength);
          constr[baselength + linklength] = '\0';
        }
      else if (*link == '/')
        {
          /* LINK is an absolute path: we need to replace everything
             after (and including) the FIRST slash with LINK.

             So, if BASE is "http://host/whatever/foo/bar", and LINK is
             "/qux/xyzzy", our result should be
             "http://host/qux/xyzzy".  */
          int span;
          const char *slash;
          const char *start_insert = NULL; /* for gcc to shut up. */
          const char *pos = base;
          int seen_slash_slash = 0;
          /* We're looking for the first slash, but want to ignore
             double slash. */
        again:
          slash = memchr (pos, '/', end - pos);
          if (slash && !seen_slash_slash)
            if (*(slash + 1) == '/')
              {
                pos = slash + 2;
                seen_slash_slash = 1;
                goto again;
              }

          /* At this point, SLASH is the location of the first / after
             "//", or the first slash altogether.  START_INSERT is the
             pointer to the location where LINK will be inserted.  When
             examining the last two examples, keep in mind that LINK
             begins with '/'. */

          if (!slash && !seen_slash_slash)
            /* example: "foo" */
            /*           ^    */
            start_insert = base;
          else if (!slash && seen_slash_slash)
            /* example: "http://foo" */
            /*                     ^ */
            start_insert = end;
          else if (slash && !seen_slash_slash)
            /* example: "foo/bar" */
            /*           ^        */
            start_insert = base;
          else if (slash && seen_slash_slash)
            /* example: "http://something/" */
            /*                           ^  */
            start_insert = slash;

          span = start_insert - base;
          constr = (char *)xmalloc (span + linklength + 1);
          if (span)
            memcpy (constr, base, span);
          if (linklength)
            memcpy (constr + span, link, linklength);
          constr[span + linklength] = '\0';
        }
      else
        {
          /* LINK is a relative URL: we need to replace everything
             after last slash (possibly empty) with LINK.

             So, if BASE is "whatever/foo/bar", and LINK is "qux/xyzzy",
             our result should be "whatever/foo/qux/xyzzy".  */
          int need_explicit_slash = 0;
          int span;
          const char *start_insert;
          const char *last_slash = find_last_char (base, end, '/');
          if (!last_slash)
            {
              /* No slash found at all.  Append LINK to what we have,
                 but we'll need a slash as a separator.

                 Example: if base == "foo" and link == "qux/xyzzy", then
                 we cannot just append link to base, because we'd get
                 "fooqux/xyzzy", whereas what we want is
                 "foo/qux/xyzzy".

                 To make sure the / gets inserted, we set
                 need_explicit_slash to 1.  We also set start_insert
                 to end + 1, so that the length calculations work out
                 correctly for one more (slash) character.  Accessing
                 that character is fine, since it will be the
                 delimiter, '\0' or '?'.  */
              /* example: "foo?..." */
              /*               ^    ('?' gets changed to '/') */
              start_insert = end + 1;
              need_explicit_slash = 1;
            }
          else if (last_slash && last_slash != base && *(last_slash - 1) == '/')
            {
              /* example: http://host"  */
              /*                      ^ */
              start_insert = end + 1;
              need_explicit_slash = 1;
            }
          else
            {
              /* example: "whatever/foo/bar" */
              /*                        ^    */
              start_insert = last_slash + 1;
            }

          span = start_insert - base;
          constr = (char *)xmalloc (span + linklength + 1);
          if (span)
            memcpy (constr, base, span);
          if (need_explicit_slash)
            constr[span - 1] = '/';
          if (linklength)
            memcpy (constr + span, link, linklength);
          constr[span + linklength] = '\0';
        }
    }
  else /* !no_scheme */
    {
      constr = strdupdelim (link, link + linklength);
    }
  return constr;
}

/* Merge BASE with LINK and return the resulting URI.  This is an
   interface to uri_merge_1 that assumes that LINK is a
   zero-terminated string.  */
char *
uri_merge (const char *base, const char *link)
{
  return uri_merge_1 (base, link, strlen (link), !url_has_scheme (link));
}
\f
#define APPEND(p, s) do {                       \
  int len = strlen (s);                         \
  memcpy (p, s, len);                           \
  p += len;                                     \
} while (0)

/* Use this instead of password when the actual password is supposed
   to be hidden.  We intentionally use a generic string without giving
   away the number of characters in the password, like previous
   versions did.  */
#define HIDDEN_PASSWORD "*password*"

/* Recreate the URL string from the data in URL.

   If HIDE is non-zero (as it is when we're calling this on a URL we
   plan to print, but not when calling it to canonicalize a URL for
   use within the program), password will be hidden.  Unsafe
   characters in the URL will be quoted.  */

char *
url_string (const struct url *url, int hide_password)
{
  int size;
  char *result, *p;
  char *quoted_user = NULL, *quoted_passwd = NULL;

  int scheme_port  = supported_schemes[url->scheme].default_port;
  char *scheme_str = supported_schemes[url->scheme].leading_string;
  int fplen = full_path_length (url);

  assert (scheme_str != NULL);

  /* Make sure the user name and password are quoted. */
  if (url->user)
    {
      quoted_user = encode_string_maybe (url->user);
      if (url->passwd)
        {
          if (hide_password)
            quoted_passwd = HIDDEN_PASSWORD;
          else
            quoted_passwd = encode_string_maybe (url->passwd);
        }
    }

  size = (strlen (scheme_str)
          + strlen (url->host)
          + fplen
          + 1);
  if (url->port != scheme_port)
    size += 1 + numdigit (url->port);
  if (quoted_user)
    {
      size += 1 + strlen (quoted_user);
      if (quoted_passwd)
        size += 1 + strlen (quoted_passwd);
    }

  p = result = xmalloc (size);

  APPEND (p, scheme_str);
  if (quoted_user)
    {
      APPEND (p, quoted_user);
      if (quoted_passwd)
        {
          *p++ = ':';
          APPEND (p, quoted_passwd);
        }
      *p++ = '@';
    }

  APPEND (p, url->host);
  if (url->port != scheme_port)
    {
      *p++ = ':';
      long_to_string (p, url->port);
      p += strlen (p);
    }

  full_path_write (url, p);
  p += fplen;
  *p++ = '\0';

  assert (p - result == size);

  if (quoted_user && quoted_user != url->user)
    xfree (quoted_user);
  if (quoted_passwd && !hide_password
      && quoted_passwd != url->passwd)
    xfree (quoted_passwd);

  return result;
}
\f
/* Returns proxy host address, in accordance with SCHEME.  */
char *
getproxy (enum url_scheme scheme)
{
  char *proxy = NULL;
  char *rewritten_url;
  static char rewritten_storage[1024];

  switch (scheme)
    {
    case SCHEME_HTTP:
      proxy = opt.http_proxy ? opt.http_proxy : getenv ("http_proxy");
      break;
#ifdef HAVE_SSL
    case SCHEME_HTTPS:
      proxy = opt.https_proxy ? opt.https_proxy : getenv ("https_proxy");
      break;
#endif
    case SCHEME_FTP:
      proxy = opt.ftp_proxy ? opt.ftp_proxy : getenv ("ftp_proxy");
      break;
    case SCHEME_INVALID:
      break;
    }
  if (!proxy || !*proxy)
    return NULL;

  /* Handle shorthands. */
  rewritten_url = rewrite_shorthand_url (proxy);
  if (rewritten_url)
    {
      strncpy (rewritten_storage, rewritten_url, sizeof(rewritten_storage));
      rewritten_storage[sizeof (rewritten_storage) - 1] = '\0';
      proxy = rewritten_storage;
    }

  return proxy;
}

/* Should a host be accessed through proxy, concerning no_proxy?  */
int
no_proxy_match (const char *host, const char **no_proxy)
{
  if (!no_proxy)
    return 1;
  else
    return !sufmatch (no_proxy, host);
}
\f
static void write_backup_file PARAMS ((const char *, downloaded_file_t));
static void replace_attr PARAMS ((const char **, int, FILE *, const char *));
static char *local_quote_string PARAMS ((const char *));

/* Change the links in one HTML file.  LINKS is a list of links in the
   document, along with their positions and the desired direction of
   the conversion.  */
void
convert_links (const char *file, struct urlpos *links)
{
  struct file_memory *fm;
  FILE *fp;
  const char *p;
  downloaded_file_t downloaded_file_return;

  struct urlpos *link;
  int to_url_count = 0, to_file_count = 0;

  logprintf (LOG_VERBOSE, _("Converting %s... "), file);

  {
    /* First we do a "dry run": go through the list L and see whether
       any URL needs to be converted in the first place.  If not, just
       leave the file alone.  */
    int dry_count = 0;
    struct urlpos *dry = links;
    for (dry = links; dry; dry = dry->next)
      if (dry->convert != CO_NOCONVERT)
        ++dry_count;
    if (!dry_count)
      {
        logputs (LOG_VERBOSE, _("nothing to do.\n"));
        return;
      }
  }

  fm = read_file (file);
  if (!fm)
    {
      logprintf (LOG_NOTQUIET, _("Cannot convert links in %s: %s\n"),
                 file, strerror (errno));
      return;
    }

  downloaded_file_return = downloaded_file (CHECK_FOR_FILE, file);
  if (opt.backup_converted && downloaded_file_return)
    write_backup_file (file, downloaded_file_return);

  /* Before opening the file for writing, unlink the file.  This is
     important if the data in FM is mmaped.  In such case, nulling the
     file, which is what fopen() below does, would make us read all
     zeroes from the mmaped region.  */
  if (unlink (file) < 0 && errno != ENOENT)
    {
      logprintf (LOG_NOTQUIET, _("Unable to delete `%s': %s\n"),
                 file, strerror (errno));
      read_file_free (fm);
      return;
    }
  /* Now open the file for writing.  */
  fp = fopen (file, "wb");
  if (!fp)
    {
      logprintf (LOG_NOTQUIET, _("Cannot convert links in %s: %s\n"),
                 file, strerror (errno));
      read_file_free (fm);
      return;
    }
  /* Here we loop through all the URLs in file, replacing those of
     them that are downloaded with relative references.  */
  p = fm->content;
  for (link = links; link; link = link->next)
    {
      char *url_start = fm->content + link->pos;

      if (link->pos >= fm->length)
        {
          DEBUGP (("Something strange is going on.  Please investigate."));
          break;
        }
      /* If the URL is not to be converted, skip it.  */
      if (link->convert == CO_NOCONVERT)
        {
          DEBUGP (("Skipping %s at position %d.\n", link->url->url, link->pos));
          continue;
        }

      /* Echo the file contents, up to the offending URL's opening
         quote, to the outfile.  */
      fwrite (p, 1, url_start - p, fp);
      p = url_start;
      if (link->convert == CO_CONVERT_TO_RELATIVE)
        {
          /* Convert absolute URL to relative. */
          char *newname = construct_relative (file, link->local_name);
          char *quoted_newname = local_quote_string (newname);
          replace_attr (&p, link->size, fp, quoted_newname);
          DEBUGP (("TO_RELATIVE: %s to %s at position %d in %s.\n",
                   link->url->url, newname, link->pos, file));
          xfree (newname);
          xfree (quoted_newname);
          ++to_file_count;
        }
      else if (link->convert == CO_CONVERT_TO_COMPLETE)
        {
          /* Convert the link to absolute URL. */
          char *newlink = link->url->url;
          char *quoted_newlink = html_quote_string (newlink);
          replace_attr (&p, link->size, fp, quoted_newlink);
          DEBUGP (("TO_COMPLETE: <something> to %s at position %d in %s.\n",
                   newlink, link->pos, file));
          xfree (quoted_newlink);
          ++to_url_count;
        }
    }
  /* Output the rest of the file. */
  if (p - fm->content < fm->length)
    fwrite (p, 1, fm->length - (p - fm->content), fp);
  fclose (fp);
  read_file_free (fm);
  logprintf (LOG_VERBOSE,
             _("%d-%d\n"), to_file_count, to_url_count);
}

/* Construct and return a malloced copy of the relative link from two
   pieces of information: local name S1 of the referring file and
   local name S2 of the referred file.

   So, if S1 is "jagor.srce.hr/index.html" and S2 is
   "jagor.srce.hr/images/news.gif", the function will return
   "images/news.gif".

   Alternately, if S1 is "fly.cc.fer.hr/ioccc/index.html", and S2 is
   "fly.cc.fer.hr/images/fly.gif", the function will return
   "../images/fly.gif".

   Caveats: S1 should not begin with `/', unless S2 also begins with
   '/'.  S1 should not contain things like ".." and such --
   construct_relative ("fly/ioccc/../index.html",
   "fly/images/fly.gif") will fail.  (A workaround is to call
   something like path_simplify() on S1).  */
static char *
construct_relative (const char *s1, const char *s2)
{
  int i, cnt, sepdirs1;
  char *res;

  if (*s2 == '/')
    return xstrdup (s2);
  /* S1 should *not* be absolute, if S2 wasn't.  */
  assert (*s1 != '/');
  i = cnt = 0;
  /* Skip the directories common to both strings.  */
  while (1)
    {
      while (s1[i] && s2[i]
             && (s1[i] == s2[i])
             && (s1[i] != '/')
             && (s2[i] != '/'))
        ++i;
      if (s1[i] == '/' && s2[i] == '/')
        cnt = ++i;
      else
        break;
    }
  for (sepdirs1 = 0; s1[i]; i++)
    if (s1[i] == '/')
      ++sepdirs1;
  /* Now, construct the file as of:
     - ../ repeated sepdirs1 time
     - all the non-mutual directories of S2.  */
  res = (char *)xmalloc (3 * sepdirs1 + strlen (s2 + cnt) + 1);
  for (i = 0; i < sepdirs1; i++)
    memcpy (res + 3 * i, "../", 3);
  strcpy (res + 3 * i, s2 + cnt);
  return res;
}
\f
static void
write_backup_file (const char *file, downloaded_file_t downloaded_file_return)
{
  /* Rather than just writing over the original .html file with the
     converted version, save the former to *.orig.  Note we only do
     this for files we've _successfully_ downloaded, so we don't
     clobber .orig files sitting around from previous invocations. */

  /* Construct the backup filename as the original name plus ".orig". */
  size_t         filename_len = strlen(file);
  char*          filename_plus_orig_suffix;
  boolean        already_wrote_backup_file = FALSE;
  slist*         converted_file_ptr;
  static slist*  converted_files = NULL;

  if (downloaded_file_return == FILE_DOWNLOADED_AND_HTML_EXTENSION_ADDED)
    {
      /* Just write "orig" over "html".  We need to do it this way
         because when we're checking to see if we've downloaded the
         file before (to see if we can skip downloading it), we don't
         know if it's a text/html file.  Therefore we don't know yet
         at that stage that -E is going to cause us to tack on
         ".html", so we need to compare vs. the original URL plus
         ".orig", not the original URL plus ".html.orig". */
      filename_plus_orig_suffix = alloca (filename_len + 1);
      strcpy(filename_plus_orig_suffix, file);
      strcpy((filename_plus_orig_suffix + filename_len) - 4, "orig");
    }
  else /* downloaded_file_return == FILE_DOWNLOADED_NORMALLY */
    {
      /* Append ".orig" to the name. */
      filename_plus_orig_suffix = alloca (filename_len + sizeof(".orig"));
      strcpy(filename_plus_orig_suffix, file);
      strcpy(filename_plus_orig_suffix + filename_len, ".orig");
    }

  /* We can get called twice on the same URL thanks to the
     convert_all_links() call in main().  If we write the .orig file
     each time in such a case, it'll end up containing the first-pass
     conversion, not the original file.  So, see if we've already been
     called on this file. */
  converted_file_ptr = converted_files;
  while (converted_file_ptr != NULL)
    if (strcmp(converted_file_ptr->string, file) == 0)
      {
        already_wrote_backup_file = TRUE;
        break;
      }
    else
      converted_file_ptr = converted_file_ptr->next;

  if (!already_wrote_backup_file)
    {
      /* Rename <file> to <file>.orig before former gets written over. */
      if (rename(file, filename_plus_orig_suffix) != 0)
        logprintf (LOG_NOTQUIET, _("Cannot back up %s as %s: %s\n"),
                   file, filename_plus_orig_suffix, strerror (errno));

      /* Remember that we've already written a .orig backup for this file.
         Note that we never free this memory since we need it till the
         convert_all_links() call, which is one of the last things the
         program does before terminating.  BTW, I'm not sure if it would be
         safe to just set 'converted_file_ptr->string' to 'file' below,
         rather than making a copy of the string...  Another note is that I
         thought I could just add a field to the urlpos structure saying
         that we'd written a .orig file for this URL, but that didn't work,
         so I had to make this separate list.
         -- Dan Harkless <wget@harkless.org>

         This [adding a field to the urlpos structure] didn't work
         because convert_file() is called from convert_all_links at
         the end of the retrieval with a freshly built new urlpos
         list.
         -- Hrvoje Niksic <hniksic@arsdigita.com>
      */
      converted_file_ptr = xmalloc(sizeof(*converted_file_ptr));
      converted_file_ptr->string = xstrdup(file);  /* die on out-of-mem. */
      converted_file_ptr->next = converted_files;
      converted_files = converted_file_ptr;
    }
}

static int find_fragment PARAMS ((const char *, int, const char **,
                                  const char **));

static void
replace_attr (const char **pp, int raw_size, FILE *fp, const char *new_str)
{
  const char *p = *pp;
  int quote_flag = 0;
  int size = raw_size;
  char quote_char = '\"';
  const char *frag_beg, *frag_end;

  /* Structure of our string is:
       "...old-contents..."
       <---    size    --->  (with quotes)
     OR:
       ...old-contents...
       <---    size   -->    (no quotes)   */

  if (*p == '\"' || *p == '\'')
    {
      quote_char = *p;
      quote_flag = 1;
      ++p;
      size -= 2;                /* disregard opening and closing quote */
    }
  putc (quote_char, fp);
  fputs (new_str, fp);

  /* Look for fragment identifier, if any. */
  if (find_fragment (p, size, &frag_beg, &frag_end))
    fwrite (frag_beg, 1, frag_end - frag_beg, fp);
  p += size;
  if (quote_flag)
    ++p;
  putc (quote_char, fp);
  *pp = p;
}

/* Find the first occurrence of '#' in [BEG, BEG+SIZE) that is not
   preceded by '&'.  If the character is not found, return zero.  If
   the character is found, return 1 and set BP and EP to point to the
   beginning and end of the region.

   This is used for finding the fragment indentifiers in URLs.  */

static int
find_fragment (const char *beg, int size, const char **bp, const char **ep)
{
  const char *end = beg + size;
  int saw_amp = 0;
  for (; beg < end; beg++)
    {
      switch (*beg)
        {
        case '&':
          saw_amp = 1;
          break;
        case '#':
          if (!saw_amp)
            {
              *bp = beg;
              *ep = end;
              return 1;
            }
          /* fallthrough */
        default:
          saw_amp = 0;
        }
    }
  return 0;
}

/* The idea here was to quote ? as %3F to avoid passing part of the
   file name as the parameter when browsing the converted file through
   HTTP.  However, actually doing that breaks local browsing because
   "index.html%3Ffoo=bar" isn't even recognized as an HTML file!
   Perhaps this should be controlled by an option, but for now I'm
   leaving the question marks.

   This is the original docstring of this function:

   FILE should be a relative link to a local file.  It should be
   quoted as HTML because it will be used in HTML context.  However,
   we need to quote ? as %3F to avoid passing part of the file name as
   the parameter.  (This is not a problem when viewing locally, but is
   if the downloaded and converted tree is served by an HTTP
   server.)  */

/* Quote string as HTML. */

static char *
local_quote_string (const char *file)
{
  return html_quote_string (file);

#if 0
  const char *file_sans_qmark;
  int qm = count_char (file, '?');

  if (qm)
    {
      const char *from = file;
      char *to, *newname;

      /* qm * 2 because we replace each question mark with "%3F",
         i.e. replace one char with three, hence two more.  */
      int fsqlen = strlen (file) + qm * 2;

      to = newname = (char *)alloca (fsqlen + 1);
      for (; *from; from++)
        {
          if (*from != '?')
            *to++ = *from;
          else
            {
              *to++ = '%';
              *to++ = '3';
              *to++ = 'F';
            }
        }
      assert (to - newname == fsqlen);
      *to = '\0';

      file_sans_qmark = newname;
    }
  else
    file_sans_qmark = file;

  return html_quote_string (file_sans_qmark);
#endif
}

/* We're storing "modes" of type downloaded_file_t in the hash table.
   However, our hash tables only accept pointers for keys and values.
   So when we need a pointer, we use the address of a
   downloaded_file_t variable of static storage.  */
   
static downloaded_file_t *
downloaded_mode_to_ptr (downloaded_file_t mode)
{
  static downloaded_file_t
    v1 = FILE_NOT_ALREADY_DOWNLOADED,
    v2 = FILE_DOWNLOADED_NORMALLY,
    v3 = FILE_DOWNLOADED_AND_HTML_EXTENSION_ADDED,
    v4 = CHECK_FOR_FILE;

  switch (mode)
    {
    case FILE_NOT_ALREADY_DOWNLOADED:
      return &v1;
    case FILE_DOWNLOADED_NORMALLY:
      return &v2;
    case FILE_DOWNLOADED_AND_HTML_EXTENSION_ADDED:
      return &v3;
    case CHECK_FOR_FILE:
      return &v4;
    }
  return NULL;
}

/* This should really be merged with dl_file_url_map and
   downloaded_html_files in recur.c.  This was originally a list, but
   I changed it to a hash table beause it was actually taking a lot of
   time to find things in it.  */

static struct hash_table *downloaded_files_hash;

/* Remembers which files have been downloaded.  In the standard case, should be
   called with mode == FILE_DOWNLOADED_NORMALLY for each file we actually
   download successfully (i.e. not for ones we have failures on or that we skip
   due to -N).

   When we've downloaded a file and tacked on a ".html" extension due to -E,
   call this function with FILE_DOWNLOADED_AND_HTML_EXTENSION_ADDED rather than
   FILE_DOWNLOADED_NORMALLY.

   If you just want to check if a file has been previously added without adding
   it, call with mode == CHECK_FOR_FILE.  Please be sure to call this function
   with local filenames, not remote URLs. */
downloaded_file_t
downloaded_file (downloaded_file_t mode, const char *file)
{
  downloaded_file_t *ptr;

  if (mode == CHECK_FOR_FILE)
    {
      if (!downloaded_files_hash)
        return FILE_NOT_ALREADY_DOWNLOADED;
      ptr = hash_table_get (downloaded_files_hash, file);
      if (!ptr)
        return FILE_NOT_ALREADY_DOWNLOADED;
      return *ptr;
    }

  if (!downloaded_files_hash)
    downloaded_files_hash = make_string_hash_table (0);

  ptr = hash_table_get (downloaded_files_hash, file);
  if (ptr)
    return *ptr;

  ptr = downloaded_mode_to_ptr (mode);
  hash_table_put (downloaded_files_hash, xstrdup (file), &ptr);

  return FILE_NOT_ALREADY_DOWNLOADED;
}

static int
df_free_mapper (void *key, void *value, void *ignored)
{
  xfree (key);
  return 0;
}

void
downloaded_files_free (void)
{
  if (downloaded_files_hash)
    {
      hash_table_map (downloaded_files_hash, df_free_mapper, NULL);
      hash_table_destroy (downloaded_files_hash);
      downloaded_files_hash = NULL;
    }
}