Add project files.
This commit is contained in:
240
3rd-party/curl-7.52.1/src/tool_cb_hdr.c
vendored
Normal file
240
3rd-party/curl-7.52.1/src/tool_cb_hdr.c
vendored
Normal file
@@ -0,0 +1,240 @@
|
||||
/***************************************************************************
|
||||
* _ _ ____ _
|
||||
* Project ___| | | | _ \| |
|
||||
* / __| | | | |_) | |
|
||||
* | (__| |_| | _ <| |___
|
||||
* \___|\___/|_| \_\_____|
|
||||
*
|
||||
* Copyright (C) 1998 - 2016, Daniel Stenberg, <daniel@haxx.se>, et al.
|
||||
*
|
||||
* This software is licensed as described in the file COPYING, which
|
||||
* you should have received as part of this distribution. The terms
|
||||
* are also available at https://curl.haxx.se/docs/copyright.html.
|
||||
*
|
||||
* You may opt to use, copy, modify, merge, publish, distribute and/or sell
|
||||
* copies of the Software, and permit persons to whom the Software is
|
||||
* furnished to do so, under the terms of the COPYING file.
|
||||
*
|
||||
* This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY
|
||||
* KIND, either express or implied.
|
||||
*
|
||||
***************************************************************************/
|
||||
#include "tool_setup.h"
|
||||
|
||||
#include "strcase.h"
|
||||
|
||||
#define ENABLE_CURLX_PRINTF
|
||||
/* use our own printf() functions */
|
||||
#include "curlx.h"
|
||||
|
||||
#include "tool_cfgable.h"
|
||||
#include "tool_doswin.h"
|
||||
#include "tool_msgs.h"
|
||||
#include "tool_cb_hdr.h"
|
||||
|
||||
#include "memdebug.h" /* keep this as LAST include */
|
||||
|
||||
static char *parse_filename(const char *ptr, size_t len);
|
||||
|
||||
/*
|
||||
** callback for CURLOPT_HEADERFUNCTION
|
||||
*/
|
||||
|
||||
size_t tool_header_cb(void *ptr, size_t size, size_t nmemb, void *userdata)
|
||||
{
|
||||
struct HdrCbData *hdrcbdata = userdata;
|
||||
struct OutStruct *outs = hdrcbdata->outs;
|
||||
struct OutStruct *heads = hdrcbdata->heads;
|
||||
const char *str = ptr;
|
||||
const size_t cb = size * nmemb;
|
||||
const char *end = (char *)ptr + cb;
|
||||
char *url = NULL;
|
||||
|
||||
/*
|
||||
* Once that libcurl has called back tool_header_cb() the returned value
|
||||
* is checked against the amount that was intended to be written, if
|
||||
* it does not match then it fails with CURLE_WRITE_ERROR. So at this
|
||||
* point returning a value different from sz*nmemb indicates failure.
|
||||
*/
|
||||
size_t failure = (size * nmemb) ? 0 : 1;
|
||||
|
||||
if(!heads->config)
|
||||
return failure;
|
||||
|
||||
#ifdef DEBUGBUILD
|
||||
if(size * nmemb > (size_t)CURL_MAX_HTTP_HEADER) {
|
||||
warnf(heads->config->global, "Header data exceeds single call write "
|
||||
"limit!\n");
|
||||
return failure;
|
||||
}
|
||||
#endif
|
||||
|
||||
/*
|
||||
* Write header data when curl option --dump-header (-D) is given.
|
||||
*/
|
||||
|
||||
if(heads->config->headerfile && heads->stream) {
|
||||
size_t rc = fwrite(ptr, size, nmemb, heads->stream);
|
||||
if(rc != cb)
|
||||
return rc;
|
||||
/* flush the stream to send off what we got earlier */
|
||||
(void)fflush(heads->stream);
|
||||
}
|
||||
|
||||
/*
|
||||
* This callback sets the filename where output shall be written when
|
||||
* curl options --remote-name (-O) and --remote-header-name (-J) have
|
||||
* been simultaneously given and additionally server returns an HTTP
|
||||
* Content-Disposition header specifying a filename property.
|
||||
*/
|
||||
|
||||
if(hdrcbdata->honor_cd_filename &&
|
||||
(cb > 20) && checkprefix("Content-disposition:", str) &&
|
||||
!curl_easy_getinfo(outs->config->easy, CURLINFO_EFFECTIVE_URL, &url) &&
|
||||
url && (checkprefix("http://", url) || checkprefix("https://", url))) {
|
||||
const char *p = str + 20;
|
||||
|
||||
/* look for the 'filename=' parameter
|
||||
(encoded filenames (*=) are not supported) */
|
||||
for(;;) {
|
||||
char *filename;
|
||||
size_t len;
|
||||
|
||||
while(*p && (p < end) && !ISALPHA(*p))
|
||||
p++;
|
||||
if(p > end - 9)
|
||||
break;
|
||||
|
||||
if(memcmp(p, "filename=", 9)) {
|
||||
/* no match, find next parameter */
|
||||
while((p < end) && (*p != ';'))
|
||||
p++;
|
||||
continue;
|
||||
}
|
||||
p += 9;
|
||||
|
||||
/* this expression below typecasts 'cb' only to avoid
|
||||
warning: signed and unsigned type in conditional expression
|
||||
*/
|
||||
len = (ssize_t)cb - (p - str);
|
||||
filename = parse_filename(p, len);
|
||||
if(filename) {
|
||||
outs->filename = filename;
|
||||
outs->alloc_filename = TRUE;
|
||||
outs->is_cd_filename = TRUE;
|
||||
outs->s_isreg = TRUE;
|
||||
outs->fopened = FALSE;
|
||||
outs->stream = NULL;
|
||||
hdrcbdata->honor_cd_filename = FALSE;
|
||||
break;
|
||||
}
|
||||
else
|
||||
return failure;
|
||||
}
|
||||
}
|
||||
|
||||
return cb;
|
||||
}
|
||||
|
||||
/*
|
||||
* Copies a file name part and returns an ALLOCATED data buffer.
|
||||
*/
|
||||
static char *parse_filename(const char *ptr, size_t len)
|
||||
{
|
||||
char *copy;
|
||||
char *p;
|
||||
char *q;
|
||||
char stop = '\0';
|
||||
|
||||
/* simple implementation of strndup() */
|
||||
copy = malloc(len+1);
|
||||
if(!copy)
|
||||
return NULL;
|
||||
memcpy(copy, ptr, len);
|
||||
copy[len] = '\0';
|
||||
|
||||
p = copy;
|
||||
if(*p == '\'' || *p == '"') {
|
||||
/* store the starting quote */
|
||||
stop = *p;
|
||||
p++;
|
||||
}
|
||||
else
|
||||
stop = ';';
|
||||
|
||||
/* if the filename contains a path, only use filename portion */
|
||||
q = strrchr(copy, '/');
|
||||
if(q) {
|
||||
p = q + 1;
|
||||
if(!*p) {
|
||||
Curl_safefree(copy);
|
||||
return NULL;
|
||||
}
|
||||
}
|
||||
|
||||
/* If the filename contains a backslash, only use filename portion. The idea
|
||||
is that even systems that don't handle backslashes as path separators
|
||||
probably want the path removed for convenience. */
|
||||
q = strrchr(p, '\\');
|
||||
if(q) {
|
||||
p = q + 1;
|
||||
if(!*p) {
|
||||
Curl_safefree(copy);
|
||||
return NULL;
|
||||
}
|
||||
}
|
||||
|
||||
/* scan for the end letter and stop there */
|
||||
for(q = p; *q; ++q) {
|
||||
if(*q == stop) {
|
||||
*q = '\0';
|
||||
break;
|
||||
}
|
||||
}
|
||||
|
||||
/* make sure the file name doesn't end in \r or \n */
|
||||
q = strchr(p, '\r');
|
||||
if(q)
|
||||
*q = '\0';
|
||||
|
||||
q = strchr(p, '\n');
|
||||
if(q)
|
||||
*q = '\0';
|
||||
|
||||
if(copy != p)
|
||||
memmove(copy, p, strlen(p) + 1);
|
||||
|
||||
#if defined(MSDOS) || defined(WIN32)
|
||||
{
|
||||
char *sanitized;
|
||||
SANITIZEcode sc = sanitize_file_name(&sanitized, copy, 0);
|
||||
Curl_safefree(copy);
|
||||
if(sc)
|
||||
return NULL;
|
||||
copy = sanitized;
|
||||
}
|
||||
#endif /* MSDOS || WIN32 */
|
||||
|
||||
/* in case we built debug enabled, we allow an evironment variable
|
||||
* named CURL_TESTDIR to prefix the given file name to put it into a
|
||||
* specific directory
|
||||
*/
|
||||
#ifdef DEBUGBUILD
|
||||
{
|
||||
char *tdir = curlx_getenv("CURL_TESTDIR");
|
||||
if(tdir) {
|
||||
char buffer[512]; /* suitably large */
|
||||
snprintf(buffer, sizeof(buffer), "%s/%s", tdir, copy);
|
||||
Curl_safefree(copy);
|
||||
copy = strdup(buffer); /* clone the buffer, we don't use the libcurl
|
||||
aprintf() or similar since we want to use the
|
||||
same memory code as the "real" parse_filename
|
||||
function */
|
||||
curl_free(tdir);
|
||||
}
|
||||
}
|
||||
#endif
|
||||
|
||||
return copy;
|
||||
}
|
||||
|
Reference in New Issue
Block a user