curl/lib/content_encoding.c

425 строки
12 KiB
C
Исходник Обычный вид История

2002-09-03 15:52:59 +04:00
/***************************************************************************
* _ _ ____ _
* Project ___| | | | _ \| |
* / __| | | | |_) | |
* | (__| |_| | _ <| |___
* \___|\___/|_| \_\_____|
*
* Copyright (C) 1998 - 2010, Daniel Stenberg, <daniel@haxx.se>, et al.
*
2002-09-03 15:52:59 +04:00
* 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 http://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
2002-09-03 15:52:59 +04:00
* 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.
*
2002-09-03 15:52:59 +04:00
***************************************************************************/
#include "setup.h"
#ifdef HAVE_LIBZ
2003-04-11 12:49:20 +04:00
#include <stdlib.h>
#include <string.h>
#include "urldata.h"
#include <curl/curl.h>
#include "sendf.h"
#include "content_encoding.h"
#include "curl_memory.h"
#include "memdebug.h"
2004-10-28 01:46:11 +04:00
/* Comment this out if zlib is always going to be at least ver. 1.2.0.4
(doing so will reduce code size slightly). */
#define OLD_ZLIB_SUPPORT 1
#define DSIZ CURL_MAX_WRITE_SIZE /* buffer size for decompressed data */
2003-04-11 12:49:20 +04:00
#define GZIP_MAGIC_0 0x1f
#define GZIP_MAGIC_1 0x8b
/* gzip flag byte */
#define ASCII_FLAG 0x01 /* bit 0 set: file probably ascii text */
#define HEAD_CRC 0x02 /* bit 1 set: header CRC present */
#define EXTRA_FIELD 0x04 /* bit 2 set: extra field present */
#define ORIG_NAME 0x08 /* bit 3 set: original file name present */
#define COMMENT 0x10 /* bit 4 set: file comment present */
#define RESERVED 0xE0 /* bits 5..7: reserved */
static CURLcode
process_zlib_error(struct connectdata *conn, z_stream *z)
{
struct SessionHandle *data = conn->data;
if(z->msg)
2004-10-28 01:46:11 +04:00
failf (data, "Error while processing content unencoding: %s",
z->msg);
else
2004-10-28 01:46:11 +04:00
failf (data, "Error while processing content unencoding: "
"Unknown failure within decompression software.");
return CURLE_BAD_CONTENT_ENCODING;
}
static CURLcode
exit_zlib(z_stream *z, zlibInitState *zlib_init, CURLcode result)
{
inflateEnd(z);
2004-10-28 01:46:11 +04:00
*zlib_init = ZLIB_UNINIT;
return result;
}
2004-10-28 01:46:11 +04:00
static CURLcode
inflate_stream(struct connectdata *conn,
struct SingleRequest *k)
{
int allow_restart = 1;
2004-10-28 01:46:11 +04:00
z_stream *z = &k->z; /* zlib state structure */
uInt nread = z->avail_in;
Bytef *orig_in = z->next_in;
int status; /* zlib status */
CURLcode result = CURLE_OK; /* Curl_client_write status */
2004-10-28 01:46:11 +04:00
char *decomp; /* Put the decompressed data here. */
2004-10-28 01:46:11 +04:00
/* Dynamically allocate a buffer for decompression because it's uncommonly
large to hold on the stack */
decomp = malloc(DSIZ);
if(decomp == NULL) {
2004-10-28 01:46:11 +04:00
return exit_zlib(z, &k->zlib_init, CURLE_OUT_OF_MEMORY);
}
2004-10-28 01:46:11 +04:00
/* because the buffer size is fixed, iteratively decompress and transfer to
the client via client_write. */
for (;;) {
/* (re)set buffer for decompressed output for every iteration */
2004-10-28 01:46:11 +04:00
z->next_out = (Bytef *)decomp;
z->avail_out = DSIZ;
status = inflate(z, Z_SYNC_FLUSH);
if(status == Z_OK || status == Z_STREAM_END) {
allow_restart = 0;
if((DSIZ - z->avail_out) && (!k->ignorebody)) {
result = Curl_client_write(conn, CLIENTWRITE_BODY, decomp,
DSIZ - z->avail_out);
/* if !CURLE_OK, clean up, return */
if(result) {
2004-10-28 01:46:11 +04:00
free(decomp);
return exit_zlib(z, &k->zlib_init, result);
2004-10-28 01:46:11 +04:00
}
}
2004-10-28 01:46:11 +04:00
/* Done? clean up, return */
if(status == Z_STREAM_END) {
2004-10-28 01:46:11 +04:00
free(decomp);
if(inflateEnd(z) == Z_OK)
return exit_zlib(z, &k->zlib_init, result);
else
return exit_zlib(z, &k->zlib_init, process_zlib_error(conn, z));
}
/* Done with these bytes, exit */
if(status == Z_OK && z->avail_in == 0) {
2004-10-28 01:46:11 +04:00
free(decomp);
return result;
2004-10-28 01:46:11 +04:00
}
}
else if(allow_restart && status == Z_DATA_ERROR) {
/* some servers seem to not generate zlib headers, so this is an attempt
to fix and continue anyway */
(void) inflateEnd(z); /* don't care about the return code */
if(inflateInit2(z, -MAX_WBITS) != Z_OK) {
free(decomp);
return exit_zlib(z, &k->zlib_init, process_zlib_error(conn, z));
}
z->next_in = orig_in;
z->avail_in = nread;
allow_restart = 0;
continue;
}
else { /* Error; exit loop, handle below */
2004-10-28 01:46:11 +04:00
free(decomp);
return exit_zlib(z, &k->zlib_init, process_zlib_error(conn, z));
}
}
2004-10-28 01:46:11 +04:00
/* Will never get here */
}
2003-04-11 12:49:20 +04:00
2004-10-28 01:46:11 +04:00
CURLcode
Curl_unencode_deflate_write(struct connectdata *conn,
struct SingleRequest *k,
2004-10-28 01:46:11 +04:00
ssize_t nread)
{
z_stream *z = &k->z; /* zlib state structure */
/* Initialize zlib? */
if(k->zlib_init == ZLIB_UNINIT) {
2004-10-28 01:46:11 +04:00
z->zalloc = (alloc_func)Z_NULL;
z->zfree = (free_func)Z_NULL;
z->opaque = 0;
z->next_in = NULL;
z->avail_in = 0;
if(inflateInit(z) != Z_OK)
return process_zlib_error(conn, z);
2004-10-28 01:46:11 +04:00
k->zlib_init = ZLIB_INIT;
}
/* Set the compressed input when this function is called */
z->next_in = (Bytef *)k->str;
z->avail_in = (uInt)nread;
/* Now uncompress the data */
return inflate_stream(conn, k);
2004-10-28 01:46:11 +04:00
}
#ifdef OLD_ZLIB_SUPPORT
2003-04-11 12:49:20 +04:00
/* Skip over the gzip header */
static enum {
GZIP_OK,
GZIP_BAD,
GZIP_UNDERFLOW
} check_gzip_header(unsigned char const *data, ssize_t len, ssize_t *headerlen)
2003-04-11 12:49:20 +04:00
{
int method, flags;
const ssize_t totallen = len;
/* The shortest header is 10 bytes */
if(len < 10)
2003-04-11 12:49:20 +04:00
return GZIP_UNDERFLOW;
if((data[0] != GZIP_MAGIC_0) || (data[1] != GZIP_MAGIC_1))
2003-04-11 12:49:20 +04:00
return GZIP_BAD;
method = data[2];
flags = data[3];
if(method != Z_DEFLATED || (flags & RESERVED) != 0) {
2003-04-11 12:49:20 +04:00
/* Can't handle this compression method or unknown flag */
return GZIP_BAD;
}
/* Skip over time, xflags, OS code and all previous bytes */
len -= 10;
data += 10;
if(flags & EXTRA_FIELD) {
2003-04-11 12:49:20 +04:00
ssize_t extra_len;
if(len < 2)
2003-04-11 12:49:20 +04:00
return GZIP_UNDERFLOW;
extra_len = (data[1] << 8) | data[0];
if(len < (extra_len+2))
2003-04-11 12:49:20 +04:00
return GZIP_UNDERFLOW;
len -= (extra_len + 2);
data += (extra_len + 2);
2003-04-11 12:49:20 +04:00
}
if(flags & ORIG_NAME) {
2003-04-11 12:49:20 +04:00
/* Skip over NUL-terminated file name */
while(len && *data) {
2003-04-11 12:49:20 +04:00
--len;
++data;
}
if(!len || *data)
2003-04-11 12:49:20 +04:00
return GZIP_UNDERFLOW;
/* Skip over the NUL */
--len;
++data;
}
if(flags & COMMENT) {
2003-04-11 12:49:20 +04:00
/* Skip over NUL-terminated comment */
while(len && *data) {
2003-04-11 12:49:20 +04:00
--len;
++data;
}
if(!len || *data)
2003-04-11 12:49:20 +04:00
return GZIP_UNDERFLOW;
/* Skip over the NUL */
--len;
}
if(flags & HEAD_CRC) {
if(len < 2)
2003-04-11 12:49:20 +04:00
return GZIP_UNDERFLOW;
len -= 2;
}
*headerlen = totallen - len;
return GZIP_OK;
}
2004-10-28 01:46:11 +04:00
#endif
2003-04-11 12:49:20 +04:00
CURLcode
Curl_unencode_gzip_write(struct connectdata *conn,
struct SingleRequest *k,
2003-04-11 12:49:20 +04:00
ssize_t nread)
{
z_stream *z = &k->z; /* zlib state structure */
2003-04-11 12:49:20 +04:00
/* Initialize zlib? */
if(k->zlib_init == ZLIB_UNINIT) {
2003-04-11 12:49:20 +04:00
z->zalloc = (alloc_func)Z_NULL;
z->zfree = (free_func)Z_NULL;
2004-05-12 11:55:05 +04:00
z->opaque = 0;
2004-01-06 01:54:45 +03:00
z->next_in = NULL;
z->avail_in = 0;
2004-10-28 01:46:11 +04:00
if(strcmp(zlibVersion(), "1.2.0.4") >= 0) {
/* zlib ver. >= 1.2.0.4 supports transparent gzip decompressing */
if(inflateInit2(z, MAX_WBITS+32) != Z_OK) {
return process_zlib_error(conn, z);
}
k->zlib_init = ZLIB_INIT_GZIP; /* Transparent gzip decompress state */
}
else {
/* we must parse the gzip header ourselves */
if(inflateInit2(z, -MAX_WBITS) != Z_OK) {
return process_zlib_error(conn, z);
}
k->zlib_init = ZLIB_INIT; /* Initial call state */
2004-10-28 01:46:11 +04:00
}
}
if(k->zlib_init == ZLIB_INIT_GZIP) {
/* Let zlib handle the gzip decompression entirely */
z->next_in = (Bytef *)k->str;
z->avail_in = (uInt)nread;
/* Now uncompress the data */
return inflate_stream(conn, k);
2003-04-11 12:49:20 +04:00
}
2004-10-28 01:46:11 +04:00
#ifndef OLD_ZLIB_SUPPORT
/* Support for old zlib versions is compiled away and we are running with
an old version, so return an error. */
return exit_zlib(z, &k->zlib_init, CURLE_FUNCTION_NOT_FOUND);
#else
2003-04-11 12:49:20 +04:00
/* This next mess is to get around the potential case where there isn't
2004-05-12 11:55:05 +04:00
* enough data passed in to skip over the gzip header. If that happens, we
* malloc a block and copy what we have then wait for the next call. If
* there still isn't enough (this is definitely a worst-case scenario), we
* make the block bigger, copy the next part in and keep waiting.
2004-10-28 01:46:11 +04:00
*
* This is only required with zlib versions < 1.2.0.4 as newer versions
* can handle the gzip header themselves.
2004-05-12 11:55:05 +04:00
*/
2003-04-11 12:49:20 +04:00
2004-10-28 01:46:11 +04:00
switch (k->zlib_init) {
2003-04-11 12:49:20 +04:00
/* Skip over gzip header? */
2004-10-28 01:46:11 +04:00
case ZLIB_INIT:
{
2003-04-11 12:49:20 +04:00
/* Initial call state */
ssize_t hlen;
switch (check_gzip_header((unsigned char *)k->str, nread, &hlen)) {
case GZIP_OK:
z->next_in = (Bytef *)k->str + hlen;
z->avail_in = (uInt)(nread - hlen);
2004-10-28 01:46:11 +04:00
k->zlib_init = ZLIB_GZIP_INFLATING; /* Inflating stream state */
2003-04-11 12:49:20 +04:00
break;
case GZIP_UNDERFLOW:
2004-05-12 11:55:05 +04:00
/* We need more data so we can find the end of the gzip header. It's
* possible that the memory block we malloc here will never be freed if
* the transfer abruptly aborts after this point. Since it's unlikely
* that circumstances will be right for this code path to be followed in
* the first place, and it's even more unlikely for a transfer to fail
* immediately afterwards, it should seldom be a problem.
*/
z->avail_in = (uInt)nread;
2003-04-11 12:49:20 +04:00
z->next_in = malloc(z->avail_in);
if(z->next_in == NULL) {
2003-04-11 12:49:20 +04:00
return exit_zlib(z, &k->zlib_init, CURLE_OUT_OF_MEMORY);
}
memcpy(z->next_in, k->str, z->avail_in);
2004-10-28 01:46:11 +04:00
k->zlib_init = ZLIB_GZIP_HEADER; /* Need more gzip header data state */
2003-04-11 12:49:20 +04:00
/* We don't have any data to inflate yet */
return CURLE_OK;
case GZIP_BAD:
default:
return exit_zlib(z, &k->zlib_init, process_zlib_error(conn, z));
2003-04-11 12:49:20 +04:00
}
}
2004-10-28 01:46:11 +04:00
break;
case ZLIB_GZIP_HEADER:
{
2003-04-11 12:49:20 +04:00
/* Need more gzip header data state */
ssize_t hlen;
unsigned char *oldblock = z->next_in;
z->avail_in += (uInt)nread;
2003-04-11 12:49:20 +04:00
z->next_in = realloc(z->next_in, z->avail_in);
if(z->next_in == NULL) {
2003-04-11 12:49:20 +04:00
free(oldblock);
return exit_zlib(z, &k->zlib_init, CURLE_OUT_OF_MEMORY);
}
/* Append the new block of data to the previous one */
memcpy(z->next_in + z->avail_in - nread, k->str, nread);
switch (check_gzip_header(z->next_in, z->avail_in, &hlen)) {
case GZIP_OK:
/* This is the zlib stream data */
free(z->next_in);
/* Don't point into the malloced block since we just freed it */
z->next_in = (Bytef *)k->str + hlen + nread - z->avail_in;
z->avail_in = (uInt)(z->avail_in - hlen);
2004-10-28 01:46:11 +04:00
k->zlib_init = ZLIB_GZIP_INFLATING; /* Inflating stream state */
2003-04-11 12:49:20 +04:00
break;
case GZIP_UNDERFLOW:
/* We still don't have any data to inflate! */
return CURLE_OK;
case GZIP_BAD:
default:
free(z->next_in);
return exit_zlib(z, &k->zlib_init, process_zlib_error(conn, z));
2003-04-11 12:49:20 +04:00
}
}
2004-10-28 01:46:11 +04:00
break;
case ZLIB_GZIP_INFLATING:
default:
2003-04-11 12:49:20 +04:00
/* Inflating stream state */
z->next_in = (Bytef *)k->str;
z->avail_in = (uInt)nread;
2004-10-28 01:46:11 +04:00
break;
2003-04-11 12:49:20 +04:00
}
if(z->avail_in == 0) {
2003-04-11 12:49:20 +04:00
/* We don't have any data to inflate; wait until next time */
return CURLE_OK;
}
2004-10-28 01:46:11 +04:00
/* We've parsed the header, now uncompress the data */
return inflate_stream(conn, k);
2004-10-28 01:46:11 +04:00
#endif
2003-04-11 12:49:20 +04:00
}
void Curl_unencode_cleanup(struct connectdata *conn)
{
struct SessionHandle *data = conn->data;
struct SingleRequest *k = &data->req;
z_stream *z = &k->z;
if(k->zlib_init != ZLIB_UNINIT)
(void) exit_zlib(z, &k->zlib_init, CURLE_OK);
}
#endif /* HAVE_LIBZ */