2002-09-03 15:52:59 +04:00
|
|
|
/***************************************************************************
|
2005-03-04 03:14:45 +03:00
|
|
|
* _ _ ____ _
|
|
|
|
* Project ___| | | | _ \| |
|
|
|
|
* / __| | | | |_) | |
|
|
|
|
* | (__| |_| | _ <| |___
|
2001-03-08 02:28:22 +03:00
|
|
|
* \___|\___/|_| \_\_____|
|
|
|
|
*
|
2018-01-29 02:58:59 +03:00
|
|
|
* Copyright (C) 1998 - 2018, Daniel Stenberg, <daniel@haxx.se>, et al.
|
2001-03-08 02:28:22 +03:00
|
|
|
*
|
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
|
2016-02-03 02:19:02 +03:00
|
|
|
* are also available at https://curl.haxx.se/docs/copyright.html.
|
2005-03-04 03:14:45 +03:00
|
|
|
*
|
2001-03-08 02:28:22 +03:00
|
|
|
* 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.
|
2001-03-08 02:28:22 +03:00
|
|
|
*
|
|
|
|
* 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
|
|
|
***************************************************************************/
|
2011-07-26 19:23:27 +04:00
|
|
|
|
2013-01-06 22:06:49 +04:00
|
|
|
#include "curl_setup.h"
|
2001-03-08 02:28:22 +03:00
|
|
|
|
2002-06-11 15:13:01 +04:00
|
|
|
#ifndef CURL_DISABLE_HTTP
|
2001-03-08 02:28:22 +03:00
|
|
|
|
2013-01-04 05:50:28 +04:00
|
|
|
#include "urldata.h" /* it includes http_chunks.h */
|
|
|
|
#include "sendf.h" /* for the client write stuff */
|
2001-03-08 02:28:22 +03:00
|
|
|
|
2013-01-04 05:50:28 +04:00
|
|
|
#include "content_encoding.h"
|
|
|
|
#include "http.h"
|
|
|
|
#include "non-ascii.h" /* for Curl_convert_to_network prototype */
|
2014-01-17 02:07:54 +04:00
|
|
|
#include "strtoofft.h"
|
2014-01-19 17:04:59 +04:00
|
|
|
#include "warnless.h"
|
2002-09-03 02:31:18 +04:00
|
|
|
|
2015-03-25 01:12:03 +03:00
|
|
|
/* The last #include files should be: */
|
|
|
|
#include "curl_memory.h"
|
2013-01-04 05:50:28 +04:00
|
|
|
#include "memdebug.h"
|
2001-03-08 02:28:22 +03:00
|
|
|
|
2005-03-04 03:14:45 +03:00
|
|
|
/*
|
2001-03-08 02:28:22 +03:00
|
|
|
* Chunk format (simplified):
|
|
|
|
*
|
|
|
|
* <HEX SIZE>[ chunk extension ] CRLF
|
2001-09-03 16:51:23 +04:00
|
|
|
* <DATA> CRLF
|
2001-03-08 02:28:22 +03:00
|
|
|
*
|
|
|
|
* Highlights from RFC2616 section 3.6 say:
|
|
|
|
|
|
|
|
The chunked encoding modifies the body of a message in order to
|
|
|
|
transfer it as a series of chunks, each with its own size indicator,
|
|
|
|
followed by an OPTIONAL trailer containing entity-header fields. This
|
|
|
|
allows dynamically produced content to be transferred along with the
|
|
|
|
information necessary for the recipient to verify that it has
|
|
|
|
received the full message.
|
|
|
|
|
|
|
|
Chunked-Body = *chunk
|
|
|
|
last-chunk
|
|
|
|
trailer
|
|
|
|
CRLF
|
|
|
|
|
|
|
|
chunk = chunk-size [ chunk-extension ] CRLF
|
|
|
|
chunk-data CRLF
|
|
|
|
chunk-size = 1*HEX
|
|
|
|
last-chunk = 1*("0") [ chunk-extension ] CRLF
|
|
|
|
|
|
|
|
chunk-extension= *( ";" chunk-ext-name [ "=" chunk-ext-val ] )
|
|
|
|
chunk-ext-name = token
|
|
|
|
chunk-ext-val = token | quoted-string
|
|
|
|
chunk-data = chunk-size(OCTET)
|
|
|
|
trailer = *(entity-header CRLF)
|
|
|
|
|
|
|
|
The chunk-size field is a string of hex digits indicating the size of
|
|
|
|
the chunk. The chunked encoding is ended by any chunk whose size is
|
|
|
|
zero, followed by the trailer, which is terminated by an empty line.
|
|
|
|
|
|
|
|
*/
|
|
|
|
|
2018-04-13 15:28:55 +03:00
|
|
|
#ifdef CURL_DOES_CONVERSIONS
|
|
|
|
/* Check for an ASCII hex digit.
|
|
|
|
We avoid the use of ISXDIGIT to accommodate non-ASCII hosts. */
|
|
|
|
static bool Curl_isxdigit_ascii(char digit)
|
|
|
|
{
|
|
|
|
return (digit >= 0x30 && digit <= 0x39) /* 0-9 */
|
|
|
|
|| (digit >= 0x41 && digit <= 0x46) /* A-F */
|
|
|
|
|| (digit >= 0x61 && digit <= 0x66); /* a-f */
|
|
|
|
}
|
|
|
|
#else
|
|
|
|
#define Curl_isxdigit_ascii(x) Curl_isxdigit(x)
|
|
|
|
#endif
|
|
|
|
|
2001-03-08 02:28:22 +03:00
|
|
|
void Curl_httpchunk_init(struct connectdata *conn)
|
|
|
|
{
|
2007-10-02 14:21:36 +04:00
|
|
|
struct Curl_chunker *chunk = &conn->chunk;
|
2017-09-10 00:09:06 +03:00
|
|
|
chunk->hexindex = 0; /* start at 0 */
|
|
|
|
chunk->dataleft = 0; /* no data left yet! */
|
2001-03-08 02:28:22 +03:00
|
|
|
chunk->state = CHUNK_HEX; /* we get hex first! */
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
2001-03-08 02:51:41 +03:00
|
|
|
* chunk_read() returns a OK for normal operations, or a positive return code
|
|
|
|
* for errors. STOP means this sequence of chunks is complete. The 'wrote'
|
|
|
|
* argument is set to tell the caller how many bytes we actually passed to the
|
|
|
|
* client (for byte-counting and whatever).
|
2001-03-08 02:28:22 +03:00
|
|
|
*
|
|
|
|
* The states and the state-machine is further explained in the header file.
|
2007-01-14 17:57:51 +03:00
|
|
|
*
|
|
|
|
* This function always uses ASCII hex values to accommodate non-ASCII hosts.
|
|
|
|
* For example, 0x0d and 0x0a are used instead of '\r' and '\n'.
|
2001-03-08 02:28:22 +03:00
|
|
|
*/
|
|
|
|
CHUNKcode Curl_httpchunk_read(struct connectdata *conn,
|
|
|
|
char *datap,
|
2004-03-04 18:25:06 +03:00
|
|
|
ssize_t datalen,
|
|
|
|
ssize_t *wrotep)
|
2001-03-08 02:28:22 +03:00
|
|
|
{
|
2017-09-10 00:09:06 +03:00
|
|
|
CURLcode result = CURLE_OK;
|
2016-06-21 16:47:12 +03:00
|
|
|
struct Curl_easy *data = conn->data;
|
2007-10-02 14:21:36 +04:00
|
|
|
struct Curl_chunker *ch = &conn->chunk;
|
2007-11-25 02:16:55 +03:00
|
|
|
struct SingleRequest *k = &data->req;
|
2004-03-04 18:25:06 +03:00
|
|
|
size_t piece;
|
2014-01-17 02:07:54 +04:00
|
|
|
curl_off_t length = (curl_off_t)datalen;
|
2004-03-04 18:25:06 +03:00
|
|
|
size_t *wrote = (size_t *)wrotep;
|
|
|
|
|
|
|
|
*wrote = 0; /* nothing's written yet */
|
2001-03-08 02:28:22 +03:00
|
|
|
|
2007-02-13 00:13:47 +03:00
|
|
|
/* the original data is written to the client, but we go on with the
|
|
|
|
chunk read process, to properly calculate the content length*/
|
2008-01-03 01:30:34 +03:00
|
|
|
if(data->set.http_te_skip && !k->ignorebody) {
|
|
|
|
result = Curl_client_write(conn, CLIENTWRITE_BODY, datap, datalen);
|
|
|
|
if(result)
|
|
|
|
return CHUNKE_WRITE_ERROR;
|
|
|
|
}
|
2007-02-13 00:13:47 +03:00
|
|
|
|
2001-03-08 02:28:22 +03:00
|
|
|
while(length) {
|
|
|
|
switch(ch->state) {
|
|
|
|
case CHUNK_HEX:
|
2018-04-13 15:28:55 +03:00
|
|
|
if(Curl_isxdigit_ascii(*datap)) {
|
2001-03-08 02:28:22 +03:00
|
|
|
if(ch->hexindex < MAXNUM_SIZE) {
|
|
|
|
ch->hexbuffer[ch->hexindex] = *datap;
|
|
|
|
datap++;
|
|
|
|
length--;
|
|
|
|
ch->hexindex++;
|
|
|
|
}
|
|
|
|
else {
|
2001-03-12 18:20:35 +03:00
|
|
|
return CHUNKE_TOO_LONG_HEX; /* longer hex than we support */
|
2001-03-08 02:28:22 +03:00
|
|
|
}
|
|
|
|
}
|
|
|
|
else {
|
2014-01-17 02:07:54 +04:00
|
|
|
char *endptr;
|
2014-01-17 03:34:36 +04:00
|
|
|
if(0 == ch->hexindex)
|
2001-03-12 18:20:35 +03:00
|
|
|
/* This is illegal data, we received junk where we expected
|
|
|
|
a hexadecimal digit. */
|
|
|
|
return CHUNKE_ILLEGAL_HEX;
|
2014-01-17 03:34:36 +04:00
|
|
|
|
2001-03-08 02:28:22 +03:00
|
|
|
/* length and datap are unmodified */
|
2017-09-10 00:09:06 +03:00
|
|
|
ch->hexbuffer[ch->hexindex] = 0;
|
2011-04-20 02:48:20 +04:00
|
|
|
|
2007-01-14 17:57:51 +03:00
|
|
|
/* convert to host encoding before calling strtoul */
|
2011-04-20 02:48:20 +04:00
|
|
|
result = Curl_convert_from_network(conn->data, ch->hexbuffer,
|
2007-01-14 17:57:51 +03:00
|
|
|
ch->hexindex);
|
2011-04-20 02:48:20 +04:00
|
|
|
if(result) {
|
2007-01-14 17:57:51 +03:00
|
|
|
/* Curl_convert_from_network calls failf if unsuccessful */
|
|
|
|
/* Treat it as a bad hex character */
|
2015-03-17 16:06:48 +03:00
|
|
|
return CHUNKE_ILLEGAL_HEX;
|
2007-01-14 17:57:51 +03:00
|
|
|
}
|
2011-04-20 02:48:20 +04:00
|
|
|
|
2017-08-15 00:33:23 +03:00
|
|
|
if(curlx_strtoofft(ch->hexbuffer, &endptr, 16, &ch->datasize))
|
2014-01-17 02:07:54 +04:00
|
|
|
return CHUNKE_ILLEGAL_HEX;
|
2014-01-17 03:34:36 +04:00
|
|
|
ch->state = CHUNK_LF; /* now wait for the CRLF */
|
2001-03-08 02:28:22 +03:00
|
|
|
}
|
|
|
|
break;
|
|
|
|
|
2014-01-17 03:34:36 +04:00
|
|
|
case CHUNK_LF:
|
|
|
|
/* waiting for the LF after a chunk size */
|
2007-01-14 17:57:51 +03:00
|
|
|
if(*datap == 0x0a) {
|
2001-03-08 02:28:22 +03:00
|
|
|
/* we're now expecting data to come, unless size was zero! */
|
|
|
|
if(0 == ch->datasize) {
|
2010-08-25 15:42:14 +04:00
|
|
|
ch->state = CHUNK_TRAILER; /* now check for trailers */
|
2017-09-10 00:09:06 +03:00
|
|
|
conn->trlPos = 0;
|
2001-03-08 02:28:22 +03:00
|
|
|
}
|
2014-01-17 03:34:36 +04:00
|
|
|
else
|
2001-03-08 02:28:22 +03:00
|
|
|
ch->state = CHUNK_DATA;
|
|
|
|
}
|
2014-01-17 03:34:36 +04:00
|
|
|
|
2001-03-08 02:28:22 +03:00
|
|
|
datap++;
|
|
|
|
length--;
|
|
|
|
break;
|
|
|
|
|
|
|
|
case CHUNK_DATA:
|
2014-01-17 03:34:36 +04:00
|
|
|
/* We expect 'datasize' of data. We have 'length' right now, it can be
|
|
|
|
more or less than 'datasize'. Get the smallest piece.
|
2001-03-08 02:28:22 +03:00
|
|
|
*/
|
2014-01-19 17:04:59 +04:00
|
|
|
piece = curlx_sotouz((ch->datasize >= length)?length:ch->datasize);
|
2001-03-08 02:28:22 +03:00
|
|
|
|
|
|
|
/* Write the data portion available */
|
2018-02-12 05:38:18 +03:00
|
|
|
if(!conn->data->set.http_te_skip && !k->ignorebody) {
|
|
|
|
if(!conn->data->set.http_ce_skip && k->writer_stack)
|
|
|
|
result = Curl_unencode_write(conn, k->writer_stack, datap, piece);
|
|
|
|
else
|
2017-11-05 17:09:48 +03:00
|
|
|
result = Curl_client_write(conn, CLIENTWRITE_BODY, datap, piece);
|
2002-09-03 02:31:18 +04:00
|
|
|
|
2018-02-12 05:38:18 +03:00
|
|
|
if(result)
|
|
|
|
return CHUNKE_WRITE_ERROR;
|
|
|
|
}
|
2004-03-04 18:25:06 +03:00
|
|
|
|
2001-03-08 02:28:22 +03:00
|
|
|
*wrote += piece;
|
|
|
|
ch->datasize -= piece; /* decrease amount left to expect */
|
|
|
|
datap += piece; /* move read pointer forward */
|
|
|
|
length -= piece; /* decrease space left in this round */
|
|
|
|
|
|
|
|
if(0 == ch->datasize)
|
2001-03-14 01:16:42 +03:00
|
|
|
/* end of data this round, we now expect a trailing CRLF */
|
|
|
|
ch->state = CHUNK_POSTLF;
|
|
|
|
break;
|
2001-03-08 02:28:22 +03:00
|
|
|
|
2001-03-14 01:16:42 +03:00
|
|
|
case CHUNK_POSTLF:
|
2007-01-14 17:57:51 +03:00
|
|
|
if(*datap == 0x0a) {
|
2014-01-17 03:34:36 +04:00
|
|
|
/* The last one before we go back to hex state and start all over. */
|
|
|
|
Curl_httpchunk_init(conn); /* sets state back to CHUNK_HEX */
|
2001-03-14 01:16:42 +03:00
|
|
|
}
|
2014-01-17 03:34:36 +04:00
|
|
|
else if(*datap != 0x0d)
|
2001-03-14 01:16:42 +03:00
|
|
|
return CHUNKE_BAD_CHUNK;
|
2014-01-17 03:34:36 +04:00
|
|
|
datap++;
|
|
|
|
length--;
|
2001-03-08 02:28:22 +03:00
|
|
|
break;
|
2001-03-14 01:16:42 +03:00
|
|
|
|
2005-07-12 22:15:34 +04:00
|
|
|
case CHUNK_TRAILER:
|
2014-01-17 03:34:36 +04:00
|
|
|
if((*datap == 0x0d) || (*datap == 0x0a)) {
|
2010-08-25 15:42:14 +04:00
|
|
|
/* this is the end of a trailer, but if the trailer was zero bytes
|
|
|
|
there was no trailer and we move on */
|
2007-02-22 00:59:40 +03:00
|
|
|
|
2010-08-25 15:42:14 +04:00
|
|
|
if(conn->trlPos) {
|
|
|
|
/* we allocate trailer with 3 bytes extra room to fit this */
|
2017-09-10 00:09:06 +03:00
|
|
|
conn->trailer[conn->trlPos++] = 0x0d;
|
|
|
|
conn->trailer[conn->trlPos++] = 0x0a;
|
|
|
|
conn->trailer[conn->trlPos] = 0;
|
2007-02-22 00:59:40 +03:00
|
|
|
|
2007-01-17 01:26:50 +03:00
|
|
|
/* Convert to host encoding before calling Curl_client_write */
|
2011-04-20 02:48:20 +04:00
|
|
|
result = Curl_convert_from_network(conn->data, conn->trailer,
|
2007-01-17 01:26:50 +03:00
|
|
|
conn->trlPos);
|
2011-04-20 02:48:20 +04:00
|
|
|
if(result)
|
2007-01-17 01:26:50 +03:00
|
|
|
/* Curl_convert_from_network calls failf if unsuccessful */
|
|
|
|
/* Treat it as a bad chunk */
|
2010-08-25 15:42:14 +04:00
|
|
|
return CHUNKE_BAD_CHUNK;
|
|
|
|
|
2008-01-03 01:30:34 +03:00
|
|
|
if(!data->set.http_te_skip) {
|
|
|
|
result = Curl_client_write(conn, CLIENTWRITE_HEADER,
|
|
|
|
conn->trailer, conn->trlPos);
|
|
|
|
if(result)
|
|
|
|
return CHUNKE_WRITE_ERROR;
|
|
|
|
}
|
2017-09-10 00:09:06 +03:00
|
|
|
conn->trlPos = 0;
|
2010-08-25 15:42:14 +04:00
|
|
|
ch->state = CHUNK_TRAILER_CR;
|
2014-01-17 03:34:36 +04:00
|
|
|
if(*datap == 0x0a)
|
|
|
|
/* already on the LF */
|
|
|
|
break;
|
2005-07-12 22:15:34 +04:00
|
|
|
}
|
2010-08-25 15:42:14 +04:00
|
|
|
else {
|
|
|
|
/* no trailer, we're on the final CRLF pair */
|
|
|
|
ch->state = CHUNK_TRAILER_POSTCR;
|
|
|
|
break; /* don't advance the pointer */
|
|
|
|
}
|
|
|
|
}
|
|
|
|
else {
|
2013-01-04 05:50:28 +04:00
|
|
|
/* conn->trailer is assumed to be freed in url.c on a
|
2010-08-25 15:42:14 +04:00
|
|
|
connection basis */
|
|
|
|
if(conn->trlPos >= conn->trlMax) {
|
|
|
|
/* we always allocate three extra bytes, just because when the full
|
|
|
|
header has been received we append CRLF\0 */
|
|
|
|
char *ptr;
|
|
|
|
if(conn->trlMax) {
|
|
|
|
conn->trlMax *= 2;
|
|
|
|
ptr = realloc(conn->trailer, conn->trlMax + 3);
|
|
|
|
}
|
|
|
|
else {
|
2017-09-10 00:09:06 +03:00
|
|
|
conn->trlMax = 128;
|
2010-08-25 15:42:14 +04:00
|
|
|
ptr = malloc(conn->trlMax + 3);
|
|
|
|
}
|
|
|
|
if(!ptr)
|
|
|
|
return CHUNKE_OUT_OF_MEMORY;
|
|
|
|
conn->trailer = ptr;
|
|
|
|
}
|
|
|
|
conn->trailer[conn->trlPos++]=*datap;
|
|
|
|
}
|
|
|
|
datap++;
|
|
|
|
length--;
|
|
|
|
break;
|
|
|
|
|
|
|
|
case CHUNK_TRAILER_CR:
|
|
|
|
if(*datap == 0x0a) {
|
|
|
|
ch->state = CHUNK_TRAILER_POSTCR;
|
2005-07-12 22:15:34 +04:00
|
|
|
datap++;
|
|
|
|
length--;
|
|
|
|
}
|
|
|
|
else
|
|
|
|
return CHUNKE_BAD_CHUNK;
|
|
|
|
break;
|
|
|
|
|
2010-08-25 15:42:14 +04:00
|
|
|
case CHUNK_TRAILER_POSTCR:
|
|
|
|
/* We enter this state when a CR should arrive so we expect to
|
|
|
|
have to first pass a CR before we wait for LF */
|
2014-01-17 03:34:36 +04:00
|
|
|
if((*datap != 0x0d) && (*datap != 0x0a)) {
|
2010-08-25 15:42:14 +04:00
|
|
|
/* not a CR then it must be another header in the trailer */
|
|
|
|
ch->state = CHUNK_TRAILER;
|
|
|
|
break;
|
|
|
|
}
|
2007-02-22 00:59:40 +03:00
|
|
|
if(*datap == 0x0d) {
|
2014-01-17 03:34:36 +04:00
|
|
|
/* skip if CR */
|
2007-02-22 00:59:40 +03:00
|
|
|
datap++;
|
|
|
|
length--;
|
|
|
|
}
|
2014-01-17 03:34:36 +04:00
|
|
|
/* now wait for the final LF */
|
|
|
|
ch->state = CHUNK_STOP;
|
2007-02-22 00:59:40 +03:00
|
|
|
break;
|
|
|
|
|
2001-03-08 02:28:22 +03:00
|
|
|
case CHUNK_STOP:
|
2007-11-07 12:21:35 +03:00
|
|
|
if(*datap == 0x0a) {
|
2007-02-22 00:59:40 +03:00
|
|
|
length--;
|
|
|
|
|
|
|
|
/* Record the length of any data left in the end of the buffer
|
|
|
|
even if there's no more chunks to read */
|
2014-01-19 17:04:59 +04:00
|
|
|
ch->dataleft = curlx_sotouz(length);
|
2007-02-22 00:59:40 +03:00
|
|
|
|
|
|
|
return CHUNKE_STOP; /* return stop */
|
|
|
|
}
|
2010-08-25 15:42:14 +04:00
|
|
|
else
|
2007-02-22 00:59:40 +03:00
|
|
|
return CHUNKE_BAD_CHUNK;
|
2001-03-08 02:28:22 +03:00
|
|
|
}
|
|
|
|
}
|
|
|
|
return CHUNKE_OK;
|
|
|
|
}
|
2014-03-14 18:44:18 +04:00
|
|
|
|
|
|
|
const char *Curl_chunked_strerror(CHUNKcode code)
|
|
|
|
{
|
2016-12-14 01:34:59 +03:00
|
|
|
switch(code) {
|
2014-03-14 18:44:18 +04:00
|
|
|
default:
|
|
|
|
return "OK";
|
|
|
|
case CHUNKE_TOO_LONG_HEX:
|
|
|
|
return "Too long hexadecimal number";
|
|
|
|
case CHUNKE_ILLEGAL_HEX:
|
|
|
|
return "Illegal or missing hexadecimal sequence";
|
|
|
|
case CHUNKE_BAD_CHUNK:
|
|
|
|
return "Malformed encoding found";
|
|
|
|
case CHUNKE_WRITE_ERROR:
|
|
|
|
return "Write error";
|
|
|
|
case CHUNKE_BAD_ENCODING:
|
|
|
|
return "Bad content-encoding found";
|
|
|
|
case CHUNKE_OUT_OF_MEMORY:
|
|
|
|
return "Out of memory";
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2002-06-11 15:13:01 +04:00
|
|
|
#endif /* CURL_DISABLE_HTTP */
|