uerr_t
retrieve_url (const char *origurl, char **file, char **newloc,
const char *refurl, int *dt, bool recursive)
{
uerr_t result;
char *url;
bool location_changed;
int dummy;
char *mynewloc, *proxy;
struct url *u, *proxy_url;
int up_error_code; /* url parse error code */
char *local_file;
int redirection_count = 0;
bool post_data_suspended = false;
char *saved_post_data = NULL;
char *saved_post_file_name = NULL;
/* If dt is NULL, use local storage. */
if (!dt)
{
dt = &dummy;
dummy = 0;
}
url = xstrdup (origurl);
if (newloc)
*newloc = NULL;
if (file)
*file = NULL;
u = url_parse (url, &up_error_code);
if (!u)
{
logprintf (LOG_NOTQUIET, "%s: %s.\n", url, url_error (up_error_code));
xfree (url);
return URLERROR;
}
if (!refurl)
refurl = opt.referer;
redirected:
result = NOCONERROR;
mynewloc = NULL;
local_file = NULL;
proxy_url = NULL;
proxy = getproxy (u);
if (proxy)
{
/* Parse the proxy URL. */
proxy_url = url_parse (proxy, &up_error_code);
if (!proxy_url)
{
logprintf (LOG_NOTQUIET, _("Error parsing proxy URL %s: %s.\n"),
proxy, url_error (up_error_code));
xfree (url);
RESTORE_POST_DATA;
return PROXERR;
}
if (proxy_url->scheme != SCHEME_HTTP && proxy_url->scheme != u->scheme)
{
logprintf (LOG_NOTQUIET, _("Error in proxy URL %s: Must be HTTP.\n"), proxy);
url_free (proxy_url);
xfree (url);
RESTORE_POST_DATA;
return PROXERR;
}
}
if (u->scheme == SCHEME_HTTP
#ifdef HAVE_SSL
|| u->scheme == SCHEME_HTTPS
#endif
|| (proxy_url && proxy_url->scheme == SCHEME_HTTP))
{
result = http_loop (u, &mynewloc, &local_file, refurl, dt, proxy_url);
}
else if (u->scheme == SCHEME_FTP)
{
/* If this is a redirection, temporarily turn off opt.ftp_glob
and opt.recursive, both being undesirable when following
redirects. */
bool oldrec = recursive, glob = opt.ftp_glob;
if (redirection_count)
oldrec = glob = false;
result = ftp_loop (u, dt, proxy_url, recursive, glob);
recursive = oldrec;
/* There is a possibility of having HTTP being redirected to
FTP. In these cases we must decide whether the text is HTML
according to the suffix. The HTML suffixes are `.html',
`.htm' and a few others, case-insensitive. */
if (redirection_count && local_file && u->scheme == SCHEME_FTP)
{
if (has_html_suffix_p (local_file))
*dt |= TEXTHTML;
}
}
if (proxy_url)
{
url_free (proxy_url);
proxy_url = NULL;
}
location_changed = (result == NEWLOCATION);
if (location_changed)
{
char *construced_newloc;
struct url *newloc_parsed;
assert (mynewloc != NULL);
if (local_file)
xfree (local_file);
/* The HTTP specs only allow absolute URLs to appear in
redirects, but a ton of boneheaded webservers and CGIs out
there break the rules and use relative URLs, and popular
browsers are lenient about this, so wget should be too. */
construced_newloc = uri_merge (url, mynewloc);
xfree (mynewloc);
mynewloc = construced_newloc;
/* Now, see if this new location makes sense. */
newloc_parsed = url_parse (mynewloc, &up_error_code);
if (!newloc_parsed)
{
logprintf (LOG_NOTQUIET, "%s: %s.\n", escnonprint_uri (mynewloc),
url_error (up_error_code));
url_free (u);
xfree (url);
xfree (mynewloc);
RESTORE_POST_DATA;
return result;
}
/* Now mynewloc will become newloc_parsed->url, because if the
Location contained relative paths like .././something, we
don't want that propagating as url. */
xfree (mynewloc);
mynewloc = xstrdup (newloc_parsed->url);
/* Check for max. number of redirections. */
if (++redirection_count > opt.max_redirect)
{
logprintf (LOG_NOTQUIET, _("%d redirections exceeded.\n"),
opt.max_redirect);
url_free (newloc_parsed);
url_free (u);
xfree (url);
xfree (mynewloc);
RESTORE_POST_DATA;
return WRONGCODE;
}
xfree (url);
url = mynewloc;
url_free (u);
u = newloc_parsed;
/* If we're being redirected from POST, we don't want to POST
again. Many requests answer POST with a redirection to an
index page; that redirection is clearly a GET. We "suspend"
POST data for the duration of the redirections, and restore
it when we're done. */
if (!post_data_suspended)
SUSPEND_POST_DATA;
goto redirected;
}
if (local_file)
{
if (*dt & RETROKF)
{
register_download (u->url, local_file);
if (redirection_count && 0 != strcmp (origurl, u->url))
register_redirection (origurl, u->url);
if (*dt & TEXTHTML)
register_html (u->url, local_file);
}
}
if (file)
*file = local_file ? local_file : NULL;
else
xfree_null (local_file);
url_free (u);
if (redirection_count)
{
if (newloc)
*newloc = url;
else
xfree (url);
}
else
{
if (newloc)
*newloc = NULL;
xfree (url);
}
RESTORE_POST_DATA;
return result;
}
uerr_t
retrieve_url (struct url * orig_parsed, const char *origurl, char **file,
char **newloc, const char *refurl, int *dt, bool recursive,
struct iri *iri, bool register_status)
{
uerr_t result;
char *url;
bool location_changed;
bool iri_fallbacked = 0;
int dummy;
char *mynewloc, *proxy;
struct url *u = orig_parsed, *proxy_url;
int up_error_code; /* url parse error code */
char *local_file;
int redirection_count = 0;
bool post_data_suspended = false;
char *saved_post_data = NULL;
char *saved_post_file_name = NULL;
/* If dt is NULL, use local storage. */
if (!dt)
{
dt = &dummy;
dummy = 0;
}
url = xstrdup (origurl);
if (newloc)
*newloc = NULL;
if (file)
*file = NULL;
if (!refurl)
refurl = opt.referer;
redirected:
/* (also for IRI fallbacking) */
result = NOCONERROR;
mynewloc = NULL;
local_file = NULL;
proxy_url = NULL;
proxy = getproxy (u);
if (proxy)
{
struct iri *pi = iri_new ();
set_uri_encoding (pi, opt.locale, true);
pi->utf8_encode = false;
/* Parse the proxy URL. */
proxy_url = url_parse (proxy, &up_error_code, NULL, true);
if (!proxy_url)
{
char *error = url_error (proxy, up_error_code);
logprintf (LOG_NOTQUIET, _("Error parsing proxy URL %s: %s.\n"),
proxy, error);
xfree (url);
xfree (error);
RESTORE_POST_DATA;
result = PROXERR;
goto bail;
}
if (proxy_url->scheme != SCHEME_HTTP && proxy_url->scheme != u->scheme)
{
logprintf (LOG_NOTQUIET, _("Error in proxy URL %s: Must be HTTP.\n"), proxy);
url_free (proxy_url);
xfree (url);
RESTORE_POST_DATA;
result = PROXERR;
goto bail;
}
}
if (u->scheme == SCHEME_HTTP
#ifdef HAVE_SSL
|| u->scheme == SCHEME_HTTPS
#endif
|| (proxy_url && proxy_url->scheme == SCHEME_HTTP))
{
result = http_loop (u, orig_parsed, &mynewloc, &local_file, refurl, dt,
proxy_url, iri);
}
else if (u->scheme == SCHEME_FTP)
{
/* If this is a redirection, temporarily turn off opt.ftp_glob
and opt.recursive, both being undesirable when following
redirects. */
bool oldrec = recursive, glob = opt.ftp_glob;
if (redirection_count)
oldrec = glob = false;
result = ftp_loop (u, &local_file, dt, proxy_url, recursive, glob);
recursive = oldrec;
/* There is a possibility of having HTTP being redirected to
FTP. In these cases we must decide whether the text is HTML
according to the suffix. The HTML suffixes are `.html',
`.htm' and a few others, case-insensitive. */
if (redirection_count && local_file && u->scheme == SCHEME_FTP)
{
if (has_html_suffix_p (local_file))
*dt |= TEXTHTML;
}
}
if (proxy_url)
{
url_free (proxy_url);
proxy_url = NULL;
}
location_changed = (result == NEWLOCATION || result == NEWLOCATION_KEEP_POST);
if (location_changed)
{
char *construced_newloc;
struct url *newloc_parsed;
assert (mynewloc != NULL);
if (local_file)
xfree (local_file);
/* The HTTP specs only allow absolute URLs to appear in
redirects, but a ton of boneheaded webservers and CGIs out
there break the rules and use relative URLs, and popular
browsers are lenient about this, so wget should be too. */
construced_newloc = uri_merge (url, mynewloc);
xfree (mynewloc);
mynewloc = construced_newloc;
/* Reset UTF-8 encoding state, keep the URI encoding and reset
the content encoding. */
iri->utf8_encode = opt.enable_iri;
set_content_encoding (iri, NULL);
xfree_null (iri->orig_url);
iri->orig_url = NULL;
/* Now, see if this new location makes sense. */
newloc_parsed = url_parse (mynewloc, &up_error_code, iri, true);
if (!newloc_parsed)
{
char *error = url_error (mynewloc, up_error_code);
logprintf (LOG_NOTQUIET, "%s: %s.\n", escnonprint_uri (mynewloc),
error);
if (orig_parsed != u)
{
url_free (u);
}
xfree (url);
xfree (mynewloc);
xfree (error);
RESTORE_POST_DATA;
goto bail;
}
/* Now mynewloc will become newloc_parsed->url, because if the
Location contained relative paths like .././something, we
don't want that propagating as url. */
xfree (mynewloc);
mynewloc = xstrdup (newloc_parsed->url);
/* Check for max. number of redirections. */
if (++redirection_count > opt.max_redirect)
{
logprintf (LOG_NOTQUIET, _("%d redirections exceeded.\n"),
opt.max_redirect);
url_free (newloc_parsed);
if (orig_parsed != u)
{
url_free (u);
}
xfree (url);
xfree (mynewloc);
RESTORE_POST_DATA;
result = WRONGCODE;
goto bail;
}
xfree (url);
url = mynewloc;
if (orig_parsed != u)
{
url_free (u);
}
u = newloc_parsed;
/* If we're being redirected from POST, and we received a
redirect code different than 307, we don't want to POST
again. Many requests answer POST with a redirection to an
index page; that redirection is clearly a GET. We "suspend"
POST data for the duration of the redirections, and restore
it when we're done.
RFC2616 HTTP/1.1 introduces code 307 Temporary Redirect
specifically to preserve the method of the request.
*/
if (result != NEWLOCATION_KEEP_POST && !post_data_suspended)
SUSPEND_POST_DATA;
goto redirected;
}
/* Try to not encode in UTF-8 if fetching failed */
if (!(*dt & RETROKF) && iri->utf8_encode)
{
iri->utf8_encode = false;
if (orig_parsed != u)
{
url_free (u);
}
u = url_parse (origurl, NULL, iri, true);
if (u)
{
DEBUGP (("[IRI fallbacking to non-utf8 for %s\n", quote (url)));
url = xstrdup (u->url);
iri_fallbacked = 1;
goto redirected;
}
else
DEBUGP (("[Couldn't fallback to non-utf8 for %s\n", quote (url)));
}
if (local_file && u && *dt & RETROKF)
{
register_download (u->url, local_file);
if (!opt.spider && redirection_count && 0 != strcmp (origurl, u->url))
register_redirection (origurl, u->url);
if (*dt & TEXTHTML)
register_html (local_file);
if (*dt & TEXTCSS)
register_css (local_file);
}
if (file)
*file = local_file ? local_file : NULL;
else
xfree_null (local_file);
if (orig_parsed != u)
{
url_free (u);
}
if (redirection_count || iri_fallbacked)
{
if (newloc)
*newloc = url;
else
xfree (url);
}
else
{
if (newloc)
*newloc = NULL;
xfree (url);
}
RESTORE_POST_DATA;
bail:
if (register_status)
inform_exit_status (result);
return result;
}
/* Retrieve the given URL. Decides which loop to call -- HTTP, FTP,
or simply copy it with file:// (#### the latter not yet
implemented!). */
uerr_t
retrieve_url (const char *origurl, char **file, char **newloc,
const char *refurl, int *dt)
{
uerr_t result;
char *url;
int location_changed, already_redirected, dummy;
int local_use_proxy;
char *mynewloc, *proxy;
struct urlinfo *u;
/* If dt is NULL, just ignore it. */
if (!dt)
dt = &dummy;
url = xstrdup (origurl);
if (newloc)
*newloc = NULL;
if (file)
*file = NULL;
already_redirected = 0;
again:
u = newurl ();
/* Parse the URL. RFC2068 requires `Location' to contain an
absoluteURI, but many sites break this requirement. #### We
should be liberal and accept a relative location, too. */
result = parseurl (url, u, already_redirected);
if (result != URLOK)
{
freeurl (u, 1);
logprintf (LOG_NOTQUIET, "%s: %s.\n", url, uerrmsg (result));
return result;
}
/* Set the referer. */
if (refurl)
u->referer = xstrdup (refurl);
else
u->referer = NULL;
local_use_proxy = USE_PROXY_P (u);
if (local_use_proxy)
{
struct urlinfo *pu = newurl ();
/* Copy the original URL to new location. */
memcpy (pu, u, sizeof (*u));
pu->proxy = NULL; /* A minor correction :) */
/* Initialize u to nil. */
memset (u, 0, sizeof (*u));
u->proxy = pu;
/* Get the appropriate proxy server, appropriate for the
current protocol. */
proxy = getproxy (pu->proto);
if (!proxy)
{
logputs (LOG_NOTQUIET, _("Could not find proxy host.\n"));
freeurl (u, 1);
return PROXERR;
}
/* Parse the proxy URL. */
result = parseurl (proxy, u, 0);
if (result != URLOK || u->proto != URLHTTP)
{
if (u->proto == URLHTTP)
logprintf (LOG_NOTQUIET, "Proxy %s: %s.\n", proxy, uerrmsg (result));
else
logprintf (LOG_NOTQUIET, _("Proxy %s: Must be HTTP.\n"), proxy);
freeurl (u, 1);
return PROXERR;
}
u->proto = URLHTTP;
}
assert (u->proto != URLFILE); /* #### Implement me! */
mynewloc = NULL;
if (u->proto == URLHTTP)
result = http_loop (u, &mynewloc, dt);
else if (u->proto == URLFTP)
{
/* If this is a redirection, we must not allow recursive FTP
retrieval, so we save recursion to oldrec, and restore it
later. */
int oldrec = opt.recursive;
if (already_redirected)
opt.recursive = 0;
result = ftp_loop (u, dt);
opt.recursive = oldrec;
/* There is a possibility of having HTTP being redirected to
FTP. In these cases we must decide whether the text is HTML
according to the suffix. The HTML suffixes are `.html' and
`.htm', case-insensitive.
#### All of this is, of course, crap. These types should be
determined through mailcap. */
if (already_redirected && u->local && (u->proto == URLFTP ))
{
char *suf = suffix (u->local);
if (suf && (!strcasecmp (suf, "html") || !strcasecmp (suf, "htm")))
*dt |= TEXTHTML;
FREE_MAYBE (suf);
}
}
location_changed = (result == NEWLOCATION);
if (location_changed)
{
/* Check for redirection to oneself. */
if (url_equal (url, mynewloc))
{
logprintf (LOG_NOTQUIET, _("%s: Redirection to itself.\n"),
mynewloc);
return WRONGCODE;
}
if (mynewloc)
{
free (url);
url = mynewloc;
}
freeurl (u, 1);
already_redirected = 1;
goto again;
}
if (file)
{
if (u->local)
*file = xstrdup (u->local);
else
*file = NULL;
}
freeurl (u, 1);
if (newloc)
*newloc = url;
else
free (url);
return result;
}
