mirror of
https://github.com/moparisthebest/curl
synced 2024-11-17 23:15:08 -05:00
063d3f3b96
... remove '== NULL' and '!= 0' Closes #6912
2061 lines
54 KiB
C
2061 lines
54 KiB
C
/***************************************************************************
|
|
* _ _ ____ _
|
|
* Project ___| | | | _ \| |
|
|
* / __| | | | |_) | |
|
|
* | (__| |_| | _ <| |___
|
|
* \___|\___/|_| \_\_____|
|
|
*
|
|
* Copyright (C) 1998 - 2021, 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.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 "curl_setup.h"
|
|
|
|
#include <curl/curl.h>
|
|
|
|
#include "mime.h"
|
|
#include "non-ascii.h"
|
|
#include "warnless.h"
|
|
#include "urldata.h"
|
|
#include "sendf.h"
|
|
|
|
#if (!defined(CURL_DISABLE_HTTP) && !defined(CURL_DISABLE_MIME)) || \
|
|
!defined(CURL_DISABLE_SMTP) || !defined(CURL_DISABLE_IMAP)
|
|
|
|
#if defined(HAVE_LIBGEN_H) && defined(HAVE_BASENAME)
|
|
#include <libgen.h>
|
|
#endif
|
|
|
|
#include "rand.h"
|
|
#include "slist.h"
|
|
#include "strcase.h"
|
|
/* The last 3 #include files should be in this order */
|
|
#include "curl_printf.h"
|
|
#include "curl_memory.h"
|
|
#include "memdebug.h"
|
|
|
|
#ifdef WIN32
|
|
# ifndef R_OK
|
|
# define R_OK 4
|
|
# endif
|
|
#endif
|
|
|
|
|
|
#define READ_ERROR ((size_t) -1)
|
|
#define STOP_FILLING ((size_t) -2)
|
|
|
|
static size_t mime_subparts_read(char *buffer, size_t size, size_t nitems,
|
|
void *instream, bool *hasread);
|
|
|
|
/* Encoders. */
|
|
static size_t encoder_nop_read(char *buffer, size_t size, bool ateof,
|
|
curl_mimepart *part);
|
|
static curl_off_t encoder_nop_size(curl_mimepart *part);
|
|
static size_t encoder_7bit_read(char *buffer, size_t size, bool ateof,
|
|
curl_mimepart *part);
|
|
static size_t encoder_base64_read(char *buffer, size_t size, bool ateof,
|
|
curl_mimepart *part);
|
|
static curl_off_t encoder_base64_size(curl_mimepart *part);
|
|
static size_t encoder_qp_read(char *buffer, size_t size, bool ateof,
|
|
curl_mimepart *part);
|
|
static curl_off_t encoder_qp_size(curl_mimepart *part);
|
|
|
|
static const struct mime_encoder encoders[] = {
|
|
{"binary", encoder_nop_read, encoder_nop_size},
|
|
{"8bit", encoder_nop_read, encoder_nop_size},
|
|
{"7bit", encoder_7bit_read, encoder_nop_size},
|
|
{"base64", encoder_base64_read, encoder_base64_size},
|
|
{"quoted-printable", encoder_qp_read, encoder_qp_size},
|
|
{ZERO_NULL, ZERO_NULL, ZERO_NULL}
|
|
};
|
|
|
|
/* Base64 encoding table */
|
|
static const char base64[] =
|
|
"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
|
|
|
|
/* Quoted-printable character class table.
|
|
*
|
|
* We cannot rely on ctype functions since quoted-printable input data
|
|
* is assumed to be ascii-compatible, even on non-ascii platforms. */
|
|
#define QP_OK 1 /* Can be represented by itself. */
|
|
#define QP_SP 2 /* Space or tab. */
|
|
#define QP_CR 3 /* Carriage return. */
|
|
#define QP_LF 4 /* Line-feed. */
|
|
static const unsigned char qp_class[] = {
|
|
0, 0, 0, 0, 0, 0, 0, 0, /* 00 - 07 */
|
|
0, QP_SP, QP_LF, 0, 0, QP_CR, 0, 0, /* 08 - 0F */
|
|
0, 0, 0, 0, 0, 0, 0, 0, /* 10 - 17 */
|
|
0, 0, 0, 0, 0, 0, 0, 0, /* 18 - 1F */
|
|
QP_SP, QP_OK, QP_OK, QP_OK, QP_OK, QP_OK, QP_OK, QP_OK, /* 20 - 27 */
|
|
QP_OK, QP_OK, QP_OK, QP_OK, QP_OK, QP_OK, QP_OK, QP_OK, /* 28 - 2F */
|
|
QP_OK, QP_OK, QP_OK, QP_OK, QP_OK, QP_OK, QP_OK, QP_OK, /* 30 - 37 */
|
|
QP_OK, QP_OK, QP_OK, QP_OK, QP_OK, 0 , QP_OK, QP_OK, /* 38 - 3F */
|
|
QP_OK, QP_OK, QP_OK, QP_OK, QP_OK, QP_OK, QP_OK, QP_OK, /* 40 - 47 */
|
|
QP_OK, QP_OK, QP_OK, QP_OK, QP_OK, QP_OK, QP_OK, QP_OK, /* 48 - 4F */
|
|
QP_OK, QP_OK, QP_OK, QP_OK, QP_OK, QP_OK, QP_OK, QP_OK, /* 50 - 57 */
|
|
QP_OK, QP_OK, QP_OK, QP_OK, QP_OK, QP_OK, QP_OK, QP_OK, /* 58 - 5F */
|
|
QP_OK, QP_OK, QP_OK, QP_OK, QP_OK, QP_OK, QP_OK, QP_OK, /* 60 - 67 */
|
|
QP_OK, QP_OK, QP_OK, QP_OK, QP_OK, QP_OK, QP_OK, QP_OK, /* 68 - 6F */
|
|
QP_OK, QP_OK, QP_OK, QP_OK, QP_OK, QP_OK, QP_OK, QP_OK, /* 70 - 77 */
|
|
QP_OK, QP_OK, QP_OK, QP_OK, QP_OK, QP_OK, QP_OK, 0, /* 78 - 7F */
|
|
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* 80 - 8F */
|
|
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* 90 - 9F */
|
|
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* A0 - AF */
|
|
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* B0 - BF */
|
|
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* C0 - CF */
|
|
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* D0 - DF */
|
|
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* E0 - EF */
|
|
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 /* F0 - FF */
|
|
};
|
|
|
|
|
|
/* Binary --> hexadecimal ASCII table. */
|
|
static const char aschex[] =
|
|
"\x30\x31\x32\x33\x34\x35\x36\x37\x38\x39\x41\x42\x43\x44\x45\x46";
|
|
|
|
|
|
|
|
#ifndef __VMS
|
|
#define filesize(name, stat_data) (stat_data.st_size)
|
|
#define fopen_read fopen
|
|
|
|
#else
|
|
|
|
#include <fabdef.h>
|
|
/*
|
|
* get_vms_file_size does what it takes to get the real size of the file
|
|
*
|
|
* For fixed files, find out the size of the EOF block and adjust.
|
|
*
|
|
* For all others, have to read the entire file in, discarding the contents.
|
|
* Most posted text files will be small, and binary files like zlib archives
|
|
* and CD/DVD images should be either a STREAM_LF format or a fixed format.
|
|
*
|
|
*/
|
|
curl_off_t VmsRealFileSize(const char *name,
|
|
const struct_stat *stat_buf)
|
|
{
|
|
char buffer[8192];
|
|
curl_off_t count;
|
|
int ret_stat;
|
|
FILE * file;
|
|
|
|
file = fopen(name, FOPEN_READTEXT); /* VMS */
|
|
if(!file)
|
|
return 0;
|
|
|
|
count = 0;
|
|
ret_stat = 1;
|
|
while(ret_stat > 0) {
|
|
ret_stat = fread(buffer, 1, sizeof(buffer), file);
|
|
if(ret_stat)
|
|
count += ret_stat;
|
|
}
|
|
fclose(file);
|
|
|
|
return count;
|
|
}
|
|
|
|
/*
|
|
*
|
|
* VmsSpecialSize checks to see if the stat st_size can be trusted and
|
|
* if not to call a routine to get the correct size.
|
|
*
|
|
*/
|
|
static curl_off_t VmsSpecialSize(const char *name,
|
|
const struct_stat *stat_buf)
|
|
{
|
|
switch(stat_buf->st_fab_rfm) {
|
|
case FAB$C_VAR:
|
|
case FAB$C_VFC:
|
|
return VmsRealFileSize(name, stat_buf);
|
|
break;
|
|
default:
|
|
return stat_buf->st_size;
|
|
}
|
|
}
|
|
|
|
#define filesize(name, stat_data) VmsSpecialSize(name, &stat_data)
|
|
|
|
/*
|
|
* vmsfopenread
|
|
*
|
|
* For upload to work as expected on VMS, different optional
|
|
* parameters must be added to the fopen command based on
|
|
* record format of the file.
|
|
*
|
|
*/
|
|
static FILE * vmsfopenread(const char *file, const char *mode)
|
|
{
|
|
struct_stat statbuf;
|
|
int result;
|
|
|
|
result = stat(file, &statbuf);
|
|
|
|
switch(statbuf.st_fab_rfm) {
|
|
case FAB$C_VAR:
|
|
case FAB$C_VFC:
|
|
case FAB$C_STMCR:
|
|
return fopen(file, FOPEN_READTEXT); /* VMS */
|
|
break;
|
|
default:
|
|
return fopen(file, FOPEN_READTEXT, "rfm=stmlf", "ctx=stm");
|
|
}
|
|
}
|
|
|
|
#define fopen_read vmsfopenread
|
|
#endif
|
|
|
|
|
|
#ifndef HAVE_BASENAME
|
|
/*
|
|
(Quote from The Open Group Base Specifications Issue 6 IEEE Std 1003.1, 2004
|
|
Edition)
|
|
|
|
The basename() function shall take the pathname pointed to by path and
|
|
return a pointer to the final component of the pathname, deleting any
|
|
trailing '/' characters.
|
|
|
|
If the string pointed to by path consists entirely of the '/' character,
|
|
basename() shall return a pointer to the string "/". If the string pointed
|
|
to by path is exactly "//", it is implementation-defined whether '/' or "//"
|
|
is returned.
|
|
|
|
If path is a null pointer or points to an empty string, basename() shall
|
|
return a pointer to the string ".".
|
|
|
|
The basename() function may modify the string pointed to by path, and may
|
|
return a pointer to static storage that may then be overwritten by a
|
|
subsequent call to basename().
|
|
|
|
The basename() function need not be reentrant. A function that is not
|
|
required to be reentrant is not required to be thread-safe.
|
|
|
|
*/
|
|
static char *Curl_basename(char *path)
|
|
{
|
|
/* Ignore all the details above for now and make a quick and simple
|
|
implementation here */
|
|
char *s1;
|
|
char *s2;
|
|
|
|
s1 = strrchr(path, '/');
|
|
s2 = strrchr(path, '\\');
|
|
|
|
if(s1 && s2) {
|
|
path = (s1 > s2? s1 : s2) + 1;
|
|
}
|
|
else if(s1)
|
|
path = s1 + 1;
|
|
else if(s2)
|
|
path = s2 + 1;
|
|
|
|
return path;
|
|
}
|
|
|
|
#define basename(x) Curl_basename((x))
|
|
#endif
|
|
|
|
|
|
/* Set readback state. */
|
|
static void mimesetstate(struct mime_state *state,
|
|
enum mimestate tok, void *ptr)
|
|
{
|
|
state->state = tok;
|
|
state->ptr = ptr;
|
|
state->offset = 0;
|
|
}
|
|
|
|
|
|
/* Escape header string into allocated memory. */
|
|
static char *escape_string(const char *src)
|
|
{
|
|
size_t bytecount = 0;
|
|
size_t i;
|
|
char *dst;
|
|
|
|
for(i = 0; src[i]; i++)
|
|
if(src[i] == '"' || src[i] == '\\')
|
|
bytecount++;
|
|
|
|
bytecount += i;
|
|
dst = malloc(bytecount + 1);
|
|
if(!dst)
|
|
return NULL;
|
|
|
|
for(i = 0; *src; src++) {
|
|
if(*src == '"' || *src == '\\')
|
|
dst[i++] = '\\';
|
|
dst[i++] = *src;
|
|
}
|
|
|
|
dst[i] = '\0';
|
|
return dst;
|
|
}
|
|
|
|
/* Check if header matches. */
|
|
static char *match_header(struct curl_slist *hdr, const char *lbl, size_t len)
|
|
{
|
|
char *value = NULL;
|
|
|
|
if(strncasecompare(hdr->data, lbl, len) && hdr->data[len] == ':')
|
|
for(value = hdr->data + len + 1; *value == ' '; value++)
|
|
;
|
|
return value;
|
|
}
|
|
|
|
/* Get a header from an slist. */
|
|
static char *search_header(struct curl_slist *hdrlist, const char *hdr)
|
|
{
|
|
size_t len = strlen(hdr);
|
|
char *value = NULL;
|
|
|
|
for(; !value && hdrlist; hdrlist = hdrlist->next)
|
|
value = match_header(hdrlist, hdr, len);
|
|
|
|
return value;
|
|
}
|
|
|
|
static char *strippath(const char *fullfile)
|
|
{
|
|
char *filename;
|
|
char *base;
|
|
filename = strdup(fullfile); /* duplicate since basename() may ruin the
|
|
buffer it works on */
|
|
if(!filename)
|
|
return NULL;
|
|
base = strdup(basename(filename));
|
|
|
|
free(filename); /* free temporary buffer */
|
|
|
|
return base; /* returns an allocated string or NULL ! */
|
|
}
|
|
|
|
/* Initialize data encoder state. */
|
|
static void cleanup_encoder_state(struct mime_encoder_state *p)
|
|
{
|
|
p->pos = 0;
|
|
p->bufbeg = 0;
|
|
p->bufend = 0;
|
|
}
|
|
|
|
|
|
/* Dummy encoder. This is used for 8bit and binary content encodings. */
|
|
static size_t encoder_nop_read(char *buffer, size_t size, bool ateof,
|
|
struct curl_mimepart *part)
|
|
{
|
|
struct mime_encoder_state *st = &part->encstate;
|
|
size_t insize = st->bufend - st->bufbeg;
|
|
|
|
(void) ateof;
|
|
|
|
if(!size)
|
|
return STOP_FILLING;
|
|
|
|
if(size > insize)
|
|
size = insize;
|
|
|
|
if(size)
|
|
memcpy(buffer, st->buf + st->bufbeg, size);
|
|
|
|
st->bufbeg += size;
|
|
return size;
|
|
}
|
|
|
|
static curl_off_t encoder_nop_size(curl_mimepart *part)
|
|
{
|
|
return part->datasize;
|
|
}
|
|
|
|
|
|
/* 7bit encoder: the encoder is just a data validity check. */
|
|
static size_t encoder_7bit_read(char *buffer, size_t size, bool ateof,
|
|
curl_mimepart *part)
|
|
{
|
|
struct mime_encoder_state *st = &part->encstate;
|
|
size_t cursize = st->bufend - st->bufbeg;
|
|
|
|
(void) ateof;
|
|
|
|
if(!size)
|
|
return STOP_FILLING;
|
|
|
|
if(size > cursize)
|
|
size = cursize;
|
|
|
|
for(cursize = 0; cursize < size; cursize++) {
|
|
*buffer = st->buf[st->bufbeg];
|
|
if(*buffer++ & 0x80)
|
|
return cursize? cursize: READ_ERROR;
|
|
st->bufbeg++;
|
|
}
|
|
|
|
return cursize;
|
|
}
|
|
|
|
|
|
/* Base64 content encoder. */
|
|
static size_t encoder_base64_read(char *buffer, size_t size, bool ateof,
|
|
curl_mimepart *part)
|
|
{
|
|
struct mime_encoder_state *st = &part->encstate;
|
|
size_t cursize = 0;
|
|
int i;
|
|
char *ptr = buffer;
|
|
|
|
while(st->bufbeg < st->bufend) {
|
|
/* Line full ? */
|
|
if(st->pos > MAX_ENCODED_LINE_LENGTH - 4) {
|
|
/* Yes, we need 2 characters for CRLF. */
|
|
if(size < 2) {
|
|
if(!cursize)
|
|
return STOP_FILLING;
|
|
break;
|
|
}
|
|
*ptr++ = '\r';
|
|
*ptr++ = '\n';
|
|
st->pos = 0;
|
|
cursize += 2;
|
|
size -= 2;
|
|
}
|
|
|
|
/* Be sure there is enough space and input data for a base64 group. */
|
|
if(size < 4) {
|
|
if(!cursize)
|
|
return STOP_FILLING;
|
|
break;
|
|
}
|
|
if(st->bufend - st->bufbeg < 3)
|
|
break;
|
|
|
|
/* Encode three bytes as four characters. */
|
|
i = st->buf[st->bufbeg++] & 0xFF;
|
|
i = (i << 8) | (st->buf[st->bufbeg++] & 0xFF);
|
|
i = (i << 8) | (st->buf[st->bufbeg++] & 0xFF);
|
|
*ptr++ = base64[(i >> 18) & 0x3F];
|
|
*ptr++ = base64[(i >> 12) & 0x3F];
|
|
*ptr++ = base64[(i >> 6) & 0x3F];
|
|
*ptr++ = base64[i & 0x3F];
|
|
cursize += 4;
|
|
st->pos += 4;
|
|
size -= 4;
|
|
}
|
|
|
|
/* If at eof, we have to flush the buffered data. */
|
|
if(ateof) {
|
|
if(size < 4) {
|
|
if(!cursize)
|
|
return STOP_FILLING;
|
|
}
|
|
else {
|
|
/* Buffered data size can only be 0, 1 or 2. */
|
|
ptr[2] = ptr[3] = '=';
|
|
i = 0;
|
|
switch(st->bufend - st->bufbeg) {
|
|
case 2:
|
|
i = (st->buf[st->bufbeg + 1] & 0xFF) << 8;
|
|
/* FALLTHROUGH */
|
|
case 1:
|
|
i |= (st->buf[st->bufbeg] & 0xFF) << 16;
|
|
ptr[0] = base64[(i >> 18) & 0x3F];
|
|
ptr[1] = base64[(i >> 12) & 0x3F];
|
|
if(++st->bufbeg != st->bufend) {
|
|
ptr[2] = base64[(i >> 6) & 0x3F];
|
|
st->bufbeg++;
|
|
}
|
|
cursize += 4;
|
|
st->pos += 4;
|
|
break;
|
|
}
|
|
}
|
|
}
|
|
|
|
#ifdef CURL_DOES_CONVERSIONS
|
|
/* This is now textual data, Convert character codes. */
|
|
if(part->easy && cursize) {
|
|
CURLcode result = Curl_convert_to_network(part->easy, buffer, cursize);
|
|
if(result)
|
|
return READ_ERROR;
|
|
}
|
|
#endif
|
|
|
|
return cursize;
|
|
}
|
|
|
|
static curl_off_t encoder_base64_size(curl_mimepart *part)
|
|
{
|
|
curl_off_t size = part->datasize;
|
|
|
|
if(size <= 0)
|
|
return size; /* Unknown size or no data. */
|
|
|
|
/* Compute base64 character count. */
|
|
size = 4 * (1 + (size - 1) / 3);
|
|
|
|
/* Effective character count must include CRLFs. */
|
|
return size + 2 * ((size - 1) / MAX_ENCODED_LINE_LENGTH);
|
|
}
|
|
|
|
|
|
/* Quoted-printable lookahead.
|
|
*
|
|
* Check if a CRLF or end of data is in input buffer at current position + n.
|
|
* Return -1 if more data needed, 1 if CRLF or end of data, else 0.
|
|
*/
|
|
static int qp_lookahead_eol(struct mime_encoder_state *st, int ateof, size_t n)
|
|
{
|
|
n += st->bufbeg;
|
|
if(n >= st->bufend && ateof)
|
|
return 1;
|
|
if(n + 2 > st->bufend)
|
|
return ateof? 0: -1;
|
|
if(qp_class[st->buf[n] & 0xFF] == QP_CR &&
|
|
qp_class[st->buf[n + 1] & 0xFF] == QP_LF)
|
|
return 1;
|
|
return 0;
|
|
}
|
|
|
|
/* Quoted-printable encoder. */
|
|
static size_t encoder_qp_read(char *buffer, size_t size, bool ateof,
|
|
curl_mimepart *part)
|
|
{
|
|
struct mime_encoder_state *st = &part->encstate;
|
|
char *ptr = buffer;
|
|
size_t cursize = 0;
|
|
int softlinebreak;
|
|
char buf[4];
|
|
|
|
/* On all platforms, input is supposed to be ASCII compatible: for this
|
|
reason, we use hexadecimal ASCII codes in this function rather than
|
|
character constants that can be interpreted as non-ascii on some
|
|
platforms. Preserve ASCII encoding on output too. */
|
|
while(st->bufbeg < st->bufend) {
|
|
size_t len = 1;
|
|
size_t consumed = 1;
|
|
int i = st->buf[st->bufbeg];
|
|
buf[0] = (char) i;
|
|
buf[1] = aschex[(i >> 4) & 0xF];
|
|
buf[2] = aschex[i & 0xF];
|
|
|
|
switch(qp_class[st->buf[st->bufbeg] & 0xFF]) {
|
|
case QP_OK: /* Not a special character. */
|
|
break;
|
|
case QP_SP: /* Space or tab. */
|
|
/* Spacing must be escaped if followed by CRLF. */
|
|
switch(qp_lookahead_eol(st, ateof, 1)) {
|
|
case -1: /* More input data needed. */
|
|
return cursize;
|
|
case 0: /* No encoding needed. */
|
|
break;
|
|
default: /* CRLF after space or tab. */
|
|
buf[0] = '\x3D'; /* '=' */
|
|
len = 3;
|
|
break;
|
|
}
|
|
break;
|
|
case QP_CR: /* Carriage return. */
|
|
/* If followed by a line-feed, output the CRLF pair.
|
|
Else escape it. */
|
|
switch(qp_lookahead_eol(st, ateof, 0)) {
|
|
case -1: /* Need more data. */
|
|
return cursize;
|
|
case 1: /* CRLF found. */
|
|
buf[len++] = '\x0A'; /* Append '\n'. */
|
|
consumed = 2;
|
|
break;
|
|
default: /* Not followed by LF: escape. */
|
|
buf[0] = '\x3D'; /* '=' */
|
|
len = 3;
|
|
break;
|
|
}
|
|
break;
|
|
default: /* Character must be escaped. */
|
|
buf[0] = '\x3D'; /* '=' */
|
|
len = 3;
|
|
break;
|
|
}
|
|
|
|
/* Be sure the encoded character fits within maximum line length. */
|
|
if(buf[len - 1] != '\x0A') { /* '\n' */
|
|
softlinebreak = st->pos + len > MAX_ENCODED_LINE_LENGTH;
|
|
if(!softlinebreak && st->pos + len == MAX_ENCODED_LINE_LENGTH) {
|
|
/* We may use the current line only if end of data or followed by
|
|
a CRLF. */
|
|
switch(qp_lookahead_eol(st, ateof, consumed)) {
|
|
case -1: /* Need more data. */
|
|
return cursize;
|
|
case 0: /* Not followed by a CRLF. */
|
|
softlinebreak = 1;
|
|
break;
|
|
}
|
|
}
|
|
if(softlinebreak) {
|
|
strcpy(buf, "\x3D\x0D\x0A"); /* "=\r\n" */
|
|
len = 3;
|
|
consumed = 0;
|
|
}
|
|
}
|
|
|
|
/* If the output buffer would overflow, do not store. */
|
|
if(len > size) {
|
|
if(!cursize)
|
|
return STOP_FILLING;
|
|
break;
|
|
}
|
|
|
|
/* Append to output buffer. */
|
|
memcpy(ptr, buf, len);
|
|
cursize += len;
|
|
ptr += len;
|
|
size -= len;
|
|
st->pos += len;
|
|
if(buf[len - 1] == '\x0A') /* '\n' */
|
|
st->pos = 0;
|
|
st->bufbeg += consumed;
|
|
}
|
|
|
|
return cursize;
|
|
}
|
|
|
|
static curl_off_t encoder_qp_size(curl_mimepart *part)
|
|
{
|
|
/* Determining the size can only be done by reading the data: unless the
|
|
data size is 0, we return it as unknown (-1). */
|
|
return part->datasize? -1: 0;
|
|
}
|
|
|
|
|
|
/* In-memory data callbacks. */
|
|
/* Argument is a pointer to the mime part. */
|
|
static size_t mime_mem_read(char *buffer, size_t size, size_t nitems,
|
|
void *instream)
|
|
{
|
|
curl_mimepart *part = (curl_mimepart *) instream;
|
|
size_t sz = curlx_sotouz(part->datasize - part->state.offset);
|
|
(void) size; /* Always 1.*/
|
|
|
|
if(!nitems)
|
|
return STOP_FILLING;
|
|
|
|
if(sz > nitems)
|
|
sz = nitems;
|
|
|
|
if(sz)
|
|
memcpy(buffer, part->data + curlx_sotouz(part->state.offset), sz);
|
|
|
|
return sz;
|
|
}
|
|
|
|
static int mime_mem_seek(void *instream, curl_off_t offset, int whence)
|
|
{
|
|
curl_mimepart *part = (curl_mimepart *) instream;
|
|
|
|
switch(whence) {
|
|
case SEEK_CUR:
|
|
offset += part->state.offset;
|
|
break;
|
|
case SEEK_END:
|
|
offset += part->datasize;
|
|
break;
|
|
}
|
|
|
|
if(offset < 0 || offset > part->datasize)
|
|
return CURL_SEEKFUNC_FAIL;
|
|
|
|
part->state.offset = offset;
|
|
return CURL_SEEKFUNC_OK;
|
|
}
|
|
|
|
static void mime_mem_free(void *ptr)
|
|
{
|
|
Curl_safefree(((curl_mimepart *) ptr)->data);
|
|
}
|
|
|
|
|
|
/* Named file callbacks. */
|
|
/* Argument is a pointer to the mime part. */
|
|
static int mime_open_file(curl_mimepart *part)
|
|
{
|
|
/* Open a MIMEKIND_FILE part. */
|
|
|
|
if(part->fp)
|
|
return 0;
|
|
part->fp = fopen_read(part->data, "rb");
|
|
return part->fp? 0: -1;
|
|
}
|
|
|
|
static size_t mime_file_read(char *buffer, size_t size, size_t nitems,
|
|
void *instream)
|
|
{
|
|
curl_mimepart *part = (curl_mimepart *) instream;
|
|
|
|
if(!nitems)
|
|
return STOP_FILLING;
|
|
|
|
if(mime_open_file(part))
|
|
return READ_ERROR;
|
|
|
|
return fread(buffer, size, nitems, part->fp);
|
|
}
|
|
|
|
static int mime_file_seek(void *instream, curl_off_t offset, int whence)
|
|
{
|
|
curl_mimepart *part = (curl_mimepart *) instream;
|
|
|
|
if(whence == SEEK_SET && !offset && !part->fp)
|
|
return CURL_SEEKFUNC_OK; /* Not open: implicitly already at BOF. */
|
|
|
|
if(mime_open_file(part))
|
|
return CURL_SEEKFUNC_FAIL;
|
|
|
|
return fseek(part->fp, (long) offset, whence)?
|
|
CURL_SEEKFUNC_CANTSEEK: CURL_SEEKFUNC_OK;
|
|
}
|
|
|
|
static void mime_file_free(void *ptr)
|
|
{
|
|
curl_mimepart *part = (curl_mimepart *) ptr;
|
|
|
|
if(part->fp) {
|
|
fclose(part->fp);
|
|
part->fp = NULL;
|
|
}
|
|
Curl_safefree(part->data);
|
|
part->data = NULL;
|
|
}
|
|
|
|
|
|
/* Subparts callbacks. */
|
|
/* Argument is a pointer to the mime structure. */
|
|
|
|
/* Readback a byte string segment. */
|
|
static size_t readback_bytes(struct mime_state *state,
|
|
char *buffer, size_t bufsize,
|
|
const char *bytes, size_t numbytes,
|
|
const char *trail)
|
|
{
|
|
size_t sz;
|
|
size_t offset = curlx_sotouz(state->offset);
|
|
|
|
if(numbytes > offset) {
|
|
sz = numbytes - offset;
|
|
bytes += offset;
|
|
}
|
|
else {
|
|
size_t tsz = strlen(trail);
|
|
|
|
sz = offset - numbytes;
|
|
if(sz >= tsz)
|
|
return 0;
|
|
bytes = trail + sz;
|
|
sz = tsz - sz;
|
|
}
|
|
|
|
if(sz > bufsize)
|
|
sz = bufsize;
|
|
|
|
memcpy(buffer, bytes, sz);
|
|
state->offset += sz;
|
|
return sz;
|
|
}
|
|
|
|
/* Read a non-encoded part content. */
|
|
static size_t read_part_content(curl_mimepart *part,
|
|
char *buffer, size_t bufsize, bool *hasread)
|
|
{
|
|
size_t sz = 0;
|
|
|
|
switch(part->lastreadstatus) {
|
|
case 0:
|
|
case CURL_READFUNC_ABORT:
|
|
case CURL_READFUNC_PAUSE:
|
|
case READ_ERROR:
|
|
return part->lastreadstatus;
|
|
default:
|
|
break;
|
|
}
|
|
|
|
/* If we can determine we are at end of part data, spare a read. */
|
|
if(part->datasize != (curl_off_t) -1 &&
|
|
part->state.offset >= part->datasize) {
|
|
/* sz is already zero. */
|
|
}
|
|
else {
|
|
switch(part->kind) {
|
|
case MIMEKIND_MULTIPART:
|
|
/*
|
|
* Cannot be processed as other kinds since read function requires
|
|
* an additional parameter and is highly recursive.
|
|
*/
|
|
sz = mime_subparts_read(buffer, 1, bufsize, part->arg, hasread);
|
|
break;
|
|
case MIMEKIND_FILE:
|
|
if(part->fp && feof(part->fp))
|
|
break; /* At EOF. */
|
|
/* FALLTHROUGH */
|
|
default:
|
|
if(part->readfunc) {
|
|
if(!(part->flags & MIME_FAST_READ)) {
|
|
if(*hasread)
|
|
return STOP_FILLING;
|
|
*hasread = TRUE;
|
|
}
|
|
sz = part->readfunc(buffer, 1, bufsize, part->arg);
|
|
}
|
|
break;
|
|
}
|
|
}
|
|
|
|
switch(sz) {
|
|
case STOP_FILLING:
|
|
break;
|
|
case 0:
|
|
case CURL_READFUNC_ABORT:
|
|
case CURL_READFUNC_PAUSE:
|
|
case READ_ERROR:
|
|
part->lastreadstatus = sz;
|
|
break;
|
|
default:
|
|
part->state.offset += sz;
|
|
part->lastreadstatus = sz;
|
|
break;
|
|
}
|
|
|
|
return sz;
|
|
}
|
|
|
|
/* Read and encode part content. */
|
|
static size_t read_encoded_part_content(curl_mimepart *part, char *buffer,
|
|
size_t bufsize, bool *hasread)
|
|
{
|
|
struct mime_encoder_state *st = &part->encstate;
|
|
size_t cursize = 0;
|
|
size_t sz;
|
|
bool ateof = FALSE;
|
|
|
|
for(;;) {
|
|
if(st->bufbeg < st->bufend || ateof) {
|
|
/* Encode buffered data. */
|
|
sz = part->encoder->encodefunc(buffer, bufsize, ateof, part);
|
|
switch(sz) {
|
|
case 0:
|
|
if(ateof)
|
|
return cursize;
|
|
break;
|
|
case READ_ERROR:
|
|
case STOP_FILLING:
|
|
return cursize? cursize: sz;
|
|
default:
|
|
cursize += sz;
|
|
buffer += sz;
|
|
bufsize -= sz;
|
|
continue;
|
|
}
|
|
}
|
|
|
|
/* We need more data in input buffer. */
|
|
if(st->bufbeg) {
|
|
size_t len = st->bufend - st->bufbeg;
|
|
|
|
if(len)
|
|
memmove(st->buf, st->buf + st->bufbeg, len);
|
|
st->bufbeg = 0;
|
|
st->bufend = len;
|
|
}
|
|
if(st->bufend >= sizeof(st->buf))
|
|
return cursize? cursize: READ_ERROR; /* Buffer full. */
|
|
sz = read_part_content(part, st->buf + st->bufend,
|
|
sizeof(st->buf) - st->bufend, hasread);
|
|
switch(sz) {
|
|
case 0:
|
|
ateof = TRUE;
|
|
break;
|
|
case CURL_READFUNC_ABORT:
|
|
case CURL_READFUNC_PAUSE:
|
|
case READ_ERROR:
|
|
case STOP_FILLING:
|
|
return cursize? cursize: sz;
|
|
default:
|
|
st->bufend += sz;
|
|
break;
|
|
}
|
|
}
|
|
|
|
/* NOTREACHED */
|
|
}
|
|
|
|
/* Readback a mime part. */
|
|
static size_t readback_part(curl_mimepart *part,
|
|
char *buffer, size_t bufsize, bool *hasread)
|
|
{
|
|
size_t cursize = 0;
|
|
#ifdef CURL_DOES_CONVERSIONS
|
|
char *convbuf = buffer;
|
|
#endif
|
|
|
|
/* Readback from part. */
|
|
|
|
while(bufsize) {
|
|
size_t sz = 0;
|
|
struct curl_slist *hdr = (struct curl_slist *) part->state.ptr;
|
|
switch(part->state.state) {
|
|
case MIMESTATE_BEGIN:
|
|
mimesetstate(&part->state,
|
|
(part->flags & MIME_BODY_ONLY)?
|
|
MIMESTATE_BODY: MIMESTATE_CURLHEADERS,
|
|
part->curlheaders);
|
|
break;
|
|
case MIMESTATE_USERHEADERS:
|
|
if(!hdr) {
|
|
mimesetstate(&part->state, MIMESTATE_EOH, NULL);
|
|
break;
|
|
}
|
|
if(match_header(hdr, "Content-Type", 12)) {
|
|
mimesetstate(&part->state, MIMESTATE_USERHEADERS, hdr->next);
|
|
break;
|
|
}
|
|
/* FALLTHROUGH */
|
|
case MIMESTATE_CURLHEADERS:
|
|
if(!hdr)
|
|
mimesetstate(&part->state, MIMESTATE_USERHEADERS, part->userheaders);
|
|
else {
|
|
sz = readback_bytes(&part->state, buffer, bufsize,
|
|
hdr->data, strlen(hdr->data), "\r\n");
|
|
if(!sz)
|
|
mimesetstate(&part->state, part->state.state, hdr->next);
|
|
}
|
|
break;
|
|
case MIMESTATE_EOH:
|
|
sz = readback_bytes(&part->state, buffer, bufsize, "\r\n", 2, "");
|
|
if(!sz)
|
|
mimesetstate(&part->state, MIMESTATE_BODY, NULL);
|
|
break;
|
|
case MIMESTATE_BODY:
|
|
#ifdef CURL_DOES_CONVERSIONS
|
|
if(part->easy && convbuf < buffer) {
|
|
CURLcode result = Curl_convert_to_network(part->easy, convbuf,
|
|
buffer - convbuf);
|
|
if(result)
|
|
return READ_ERROR;
|
|
convbuf = buffer;
|
|
}
|
|
#endif
|
|
cleanup_encoder_state(&part->encstate);
|
|
mimesetstate(&part->state, MIMESTATE_CONTENT, NULL);
|
|
break;
|
|
case MIMESTATE_CONTENT:
|
|
if(part->encoder)
|
|
sz = read_encoded_part_content(part, buffer, bufsize, hasread);
|
|
else
|
|
sz = read_part_content(part, buffer, bufsize, hasread);
|
|
switch(sz) {
|
|
case 0:
|
|
mimesetstate(&part->state, MIMESTATE_END, NULL);
|
|
/* Try sparing open file descriptors. */
|
|
if(part->kind == MIMEKIND_FILE && part->fp) {
|
|
fclose(part->fp);
|
|
part->fp = NULL;
|
|
}
|
|
/* FALLTHROUGH */
|
|
case CURL_READFUNC_ABORT:
|
|
case CURL_READFUNC_PAUSE:
|
|
case READ_ERROR:
|
|
case STOP_FILLING:
|
|
return cursize? cursize: sz;
|
|
}
|
|
break;
|
|
case MIMESTATE_END:
|
|
return cursize;
|
|
default:
|
|
break; /* Other values not in part state. */
|
|
}
|
|
|
|
/* Bump buffer and counters according to read size. */
|
|
cursize += sz;
|
|
buffer += sz;
|
|
bufsize -= sz;
|
|
}
|
|
|
|
#ifdef CURL_DOES_CONVERSIONS
|
|
if(part->easy && convbuf < buffer &&
|
|
part->state.state < MIMESTATE_BODY) {
|
|
CURLcode result = Curl_convert_to_network(part->easy, convbuf,
|
|
buffer - convbuf);
|
|
if(result)
|
|
return READ_ERROR;
|
|
}
|
|
#endif
|
|
|
|
return cursize;
|
|
}
|
|
|
|
/* Readback from mime. Warning: not a read callback function. */
|
|
static size_t mime_subparts_read(char *buffer, size_t size, size_t nitems,
|
|
void *instream, bool *hasread)
|
|
{
|
|
curl_mime *mime = (curl_mime *) instream;
|
|
size_t cursize = 0;
|
|
#ifdef CURL_DOES_CONVERSIONS
|
|
char *convbuf = buffer;
|
|
#endif
|
|
|
|
(void) size; /* Always 1. */
|
|
|
|
while(nitems) {
|
|
size_t sz = 0;
|
|
curl_mimepart *part = mime->state.ptr;
|
|
switch(mime->state.state) {
|
|
case MIMESTATE_BEGIN:
|
|
case MIMESTATE_BODY:
|
|
#ifdef CURL_DOES_CONVERSIONS
|
|
convbuf = buffer;
|
|
#endif
|
|
mimesetstate(&mime->state, MIMESTATE_BOUNDARY1, mime->firstpart);
|
|
/* The first boundary always follows the header termination empty line,
|
|
so is always preceded by a CRLF. We can then spare 2 characters
|
|
by skipping the leading CRLF in boundary. */
|
|
mime->state.offset += 2;
|
|
break;
|
|
case MIMESTATE_BOUNDARY1:
|
|
sz = readback_bytes(&mime->state, buffer, nitems, "\r\n--", 4, "");
|
|
if(!sz)
|
|
mimesetstate(&mime->state, MIMESTATE_BOUNDARY2, part);
|
|
break;
|
|
case MIMESTATE_BOUNDARY2:
|
|
sz = readback_bytes(&mime->state, buffer, nitems, mime->boundary,
|
|
strlen(mime->boundary), part? "\r\n": "--\r\n");
|
|
if(!sz) {
|
|
#ifdef CURL_DOES_CONVERSIONS
|
|
if(mime->easy && convbuf < buffer) {
|
|
CURLcode result = Curl_convert_to_network(mime->easy, convbuf,
|
|
buffer - convbuf);
|
|
if(result)
|
|
return READ_ERROR;
|
|
convbuf = buffer;
|
|
}
|
|
#endif
|
|
mimesetstate(&mime->state, MIMESTATE_CONTENT, part);
|
|
}
|
|
break;
|
|
case MIMESTATE_CONTENT:
|
|
if(!part) {
|
|
mimesetstate(&mime->state, MIMESTATE_END, NULL);
|
|
break;
|
|
}
|
|
sz = readback_part(part, buffer, nitems, hasread);
|
|
switch(sz) {
|
|
case CURL_READFUNC_ABORT:
|
|
case CURL_READFUNC_PAUSE:
|
|
case READ_ERROR:
|
|
case STOP_FILLING:
|
|
return cursize? cursize: sz;
|
|
case 0:
|
|
#ifdef CURL_DOES_CONVERSIONS
|
|
convbuf = buffer;
|
|
#endif
|
|
mimesetstate(&mime->state, MIMESTATE_BOUNDARY1, part->nextpart);
|
|
break;
|
|
}
|
|
break;
|
|
case MIMESTATE_END:
|
|
return cursize;
|
|
default:
|
|
break; /* other values not used in mime state. */
|
|
}
|
|
|
|
/* Bump buffer and counters according to read size. */
|
|
cursize += sz;
|
|
buffer += sz;
|
|
nitems -= sz;
|
|
}
|
|
|
|
#ifdef CURL_DOES_CONVERSIONS
|
|
if(mime->easy && convbuf < buffer &&
|
|
mime->state.state <= MIMESTATE_CONTENT) {
|
|
CURLcode result = Curl_convert_to_network(mime->easy, convbuf,
|
|
buffer - convbuf);
|
|
if(result)
|
|
return READ_ERROR;
|
|
}
|
|
#endif
|
|
|
|
return cursize;
|
|
}
|
|
|
|
static int mime_part_rewind(curl_mimepart *part)
|
|
{
|
|
int res = CURL_SEEKFUNC_OK;
|
|
enum mimestate targetstate = MIMESTATE_BEGIN;
|
|
|
|
if(part->flags & MIME_BODY_ONLY)
|
|
targetstate = MIMESTATE_BODY;
|
|
cleanup_encoder_state(&part->encstate);
|
|
if(part->state.state > targetstate) {
|
|
res = CURL_SEEKFUNC_CANTSEEK;
|
|
if(part->seekfunc) {
|
|
res = part->seekfunc(part->arg, (curl_off_t) 0, SEEK_SET);
|
|
switch(res) {
|
|
case CURL_SEEKFUNC_OK:
|
|
case CURL_SEEKFUNC_FAIL:
|
|
case CURL_SEEKFUNC_CANTSEEK:
|
|
break;
|
|
case -1: /* For fseek() error. */
|
|
res = CURL_SEEKFUNC_CANTSEEK;
|
|
break;
|
|
default:
|
|
res = CURL_SEEKFUNC_FAIL;
|
|
break;
|
|
}
|
|
}
|
|
}
|
|
|
|
if(res == CURL_SEEKFUNC_OK)
|
|
mimesetstate(&part->state, targetstate, NULL);
|
|
|
|
part->lastreadstatus = 1; /* Successful read status. */
|
|
return res;
|
|
}
|
|
|
|
static int mime_subparts_seek(void *instream, curl_off_t offset, int whence)
|
|
{
|
|
curl_mime *mime = (curl_mime *) instream;
|
|
curl_mimepart *part;
|
|
int result = CURL_SEEKFUNC_OK;
|
|
|
|
if(whence != SEEK_SET || offset)
|
|
return CURL_SEEKFUNC_CANTSEEK; /* Only support full rewind. */
|
|
|
|
if(mime->state.state == MIMESTATE_BEGIN)
|
|
return CURL_SEEKFUNC_OK; /* Already rewound. */
|
|
|
|
for(part = mime->firstpart; part; part = part->nextpart) {
|
|
int res = mime_part_rewind(part);
|
|
if(res != CURL_SEEKFUNC_OK)
|
|
result = res;
|
|
}
|
|
|
|
if(result == CURL_SEEKFUNC_OK)
|
|
mimesetstate(&mime->state, MIMESTATE_BEGIN, NULL);
|
|
|
|
return result;
|
|
}
|
|
|
|
/* Release part content. */
|
|
static void cleanup_part_content(curl_mimepart *part)
|
|
{
|
|
if(part->freefunc)
|
|
part->freefunc(part->arg);
|
|
|
|
part->readfunc = NULL;
|
|
part->seekfunc = NULL;
|
|
part->freefunc = NULL;
|
|
part->arg = (void *) part; /* Defaults to part itself. */
|
|
part->data = NULL;
|
|
part->fp = NULL;
|
|
part->datasize = (curl_off_t) 0; /* No size yet. */
|
|
cleanup_encoder_state(&part->encstate);
|
|
part->kind = MIMEKIND_NONE;
|
|
part->flags &= ~MIME_FAST_READ;
|
|
part->lastreadstatus = 1; /* Successful read status. */
|
|
part->state.state = MIMESTATE_BEGIN;
|
|
}
|
|
|
|
static void mime_subparts_free(void *ptr)
|
|
{
|
|
curl_mime *mime = (curl_mime *) ptr;
|
|
|
|
if(mime && mime->parent) {
|
|
mime->parent->freefunc = NULL; /* Be sure we won't be called again. */
|
|
cleanup_part_content(mime->parent); /* Avoid dangling pointer in part. */
|
|
}
|
|
curl_mime_free(mime);
|
|
}
|
|
|
|
/* Do not free subparts: unbind them. This is used for the top level only. */
|
|
static void mime_subparts_unbind(void *ptr)
|
|
{
|
|
curl_mime *mime = (curl_mime *) ptr;
|
|
|
|
if(mime && mime->parent) {
|
|
mime->parent->freefunc = NULL; /* Be sure we won't be called again. */
|
|
cleanup_part_content(mime->parent); /* Avoid dangling pointer in part. */
|
|
mime->parent = NULL;
|
|
}
|
|
}
|
|
|
|
|
|
void Curl_mime_cleanpart(curl_mimepart *part)
|
|
{
|
|
cleanup_part_content(part);
|
|
curl_slist_free_all(part->curlheaders);
|
|
if(part->flags & MIME_USERHEADERS_OWNER)
|
|
curl_slist_free_all(part->userheaders);
|
|
Curl_safefree(part->mimetype);
|
|
Curl_safefree(part->name);
|
|
Curl_safefree(part->filename);
|
|
Curl_mime_initpart(part, part->easy);
|
|
}
|
|
|
|
/* Recursively delete a mime handle and its parts. */
|
|
void curl_mime_free(curl_mime *mime)
|
|
{
|
|
curl_mimepart *part;
|
|
|
|
if(mime) {
|
|
mime_subparts_unbind(mime); /* Be sure it's not referenced anymore. */
|
|
while(mime->firstpart) {
|
|
part = mime->firstpart;
|
|
mime->firstpart = part->nextpart;
|
|
Curl_mime_cleanpart(part);
|
|
free(part);
|
|
}
|
|
free(mime);
|
|
}
|
|
}
|
|
|
|
CURLcode Curl_mime_duppart(curl_mimepart *dst, const curl_mimepart *src)
|
|
{
|
|
curl_mime *mime;
|
|
curl_mimepart *d;
|
|
const curl_mimepart *s;
|
|
CURLcode res = CURLE_OK;
|
|
|
|
DEBUGASSERT(dst);
|
|
|
|
/* Duplicate content. */
|
|
switch(src->kind) {
|
|
case MIMEKIND_NONE:
|
|
break;
|
|
case MIMEKIND_DATA:
|
|
res = curl_mime_data(dst, src->data, (size_t) src->datasize);
|
|
break;
|
|
case MIMEKIND_FILE:
|
|
res = curl_mime_filedata(dst, src->data);
|
|
/* Do not abort duplication if file is not readable. */
|
|
if(res == CURLE_READ_ERROR)
|
|
res = CURLE_OK;
|
|
break;
|
|
case MIMEKIND_CALLBACK:
|
|
res = curl_mime_data_cb(dst, src->datasize, src->readfunc,
|
|
src->seekfunc, src->freefunc, src->arg);
|
|
break;
|
|
case MIMEKIND_MULTIPART:
|
|
/* No one knows about the cloned subparts, thus always attach ownership
|
|
to the part. */
|
|
mime = curl_mime_init(dst->easy);
|
|
res = mime? curl_mime_subparts(dst, mime): CURLE_OUT_OF_MEMORY;
|
|
|
|
/* Duplicate subparts. */
|
|
for(s = ((curl_mime *) src->arg)->firstpart; !res && s; s = s->nextpart) {
|
|
d = curl_mime_addpart(mime);
|
|
res = d? Curl_mime_duppart(d, s): CURLE_OUT_OF_MEMORY;
|
|
}
|
|
break;
|
|
default: /* Invalid kind: should not occur. */
|
|
res = CURLE_BAD_FUNCTION_ARGUMENT; /* Internal error? */
|
|
break;
|
|
}
|
|
|
|
/* Duplicate headers. */
|
|
if(!res && src->userheaders) {
|
|
struct curl_slist *hdrs = Curl_slist_duplicate(src->userheaders);
|
|
|
|
if(!hdrs)
|
|
res = CURLE_OUT_OF_MEMORY;
|
|
else {
|
|
/* No one but this procedure knows about the new header list,
|
|
so always take ownership. */
|
|
res = curl_mime_headers(dst, hdrs, TRUE);
|
|
if(res)
|
|
curl_slist_free_all(hdrs);
|
|
}
|
|
}
|
|
|
|
if(!res) {
|
|
/* Duplicate other fields. */
|
|
dst->encoder = src->encoder;
|
|
res = curl_mime_type(dst, src->mimetype);
|
|
}
|
|
if(!res)
|
|
res = curl_mime_name(dst, src->name);
|
|
if(!res)
|
|
res = curl_mime_filename(dst, src->filename);
|
|
|
|
/* If an error occurred, rollback. */
|
|
if(res)
|
|
Curl_mime_cleanpart(dst);
|
|
|
|
return res;
|
|
}
|
|
|
|
/*
|
|
* Mime build functions.
|
|
*/
|
|
|
|
/* Create a mime handle. */
|
|
curl_mime *curl_mime_init(struct Curl_easy *easy)
|
|
{
|
|
curl_mime *mime;
|
|
|
|
mime = (curl_mime *) malloc(sizeof(*mime));
|
|
|
|
if(mime) {
|
|
mime->easy = easy;
|
|
mime->parent = NULL;
|
|
mime->firstpart = NULL;
|
|
mime->lastpart = NULL;
|
|
|
|
memset(mime->boundary, '-', 24);
|
|
if(Curl_rand_hex(easy, (unsigned char *) &mime->boundary[24],
|
|
MIME_RAND_BOUNDARY_CHARS + 1)) {
|
|
/* failed to get random separator, bail out */
|
|
free(mime);
|
|
return NULL;
|
|
}
|
|
mimesetstate(&mime->state, MIMESTATE_BEGIN, NULL);
|
|
}
|
|
|
|
return mime;
|
|
}
|
|
|
|
/* Initialize a mime part. */
|
|
void Curl_mime_initpart(curl_mimepart *part, struct Curl_easy *easy)
|
|
{
|
|
memset((char *) part, 0, sizeof(*part));
|
|
part->easy = easy;
|
|
part->lastreadstatus = 1; /* Successful read status. */
|
|
mimesetstate(&part->state, MIMESTATE_BEGIN, NULL);
|
|
}
|
|
|
|
/* Create a mime part and append it to a mime handle's part list. */
|
|
curl_mimepart *curl_mime_addpart(curl_mime *mime)
|
|
{
|
|
curl_mimepart *part;
|
|
|
|
if(!mime)
|
|
return NULL;
|
|
|
|
part = (curl_mimepart *) malloc(sizeof(*part));
|
|
|
|
if(part) {
|
|
Curl_mime_initpart(part, mime->easy);
|
|
part->parent = mime;
|
|
|
|
if(mime->lastpart)
|
|
mime->lastpart->nextpart = part;
|
|
else
|
|
mime->firstpart = part;
|
|
|
|
mime->lastpart = part;
|
|
}
|
|
|
|
return part;
|
|
}
|
|
|
|
/* Set mime part name. */
|
|
CURLcode curl_mime_name(curl_mimepart *part, const char *name)
|
|
{
|
|
if(!part)
|
|
return CURLE_BAD_FUNCTION_ARGUMENT;
|
|
|
|
Curl_safefree(part->name);
|
|
part->name = NULL;
|
|
|
|
if(name) {
|
|
part->name = strdup(name);
|
|
if(!part->name)
|
|
return CURLE_OUT_OF_MEMORY;
|
|
}
|
|
|
|
return CURLE_OK;
|
|
}
|
|
|
|
/* Set mime part remote file name. */
|
|
CURLcode curl_mime_filename(curl_mimepart *part, const char *filename)
|
|
{
|
|
if(!part)
|
|
return CURLE_BAD_FUNCTION_ARGUMENT;
|
|
|
|
Curl_safefree(part->filename);
|
|
part->filename = NULL;
|
|
|
|
if(filename) {
|
|
part->filename = strdup(filename);
|
|
if(!part->filename)
|
|
return CURLE_OUT_OF_MEMORY;
|
|
}
|
|
|
|
return CURLE_OK;
|
|
}
|
|
|
|
/* Set mime part content from memory data. */
|
|
CURLcode curl_mime_data(curl_mimepart *part,
|
|
const char *data, size_t datasize)
|
|
{
|
|
if(!part)
|
|
return CURLE_BAD_FUNCTION_ARGUMENT;
|
|
|
|
cleanup_part_content(part);
|
|
|
|
if(data) {
|
|
if(datasize == CURL_ZERO_TERMINATED)
|
|
datasize = strlen(data);
|
|
|
|
part->data = malloc(datasize + 1);
|
|
if(!part->data)
|
|
return CURLE_OUT_OF_MEMORY;
|
|
|
|
part->datasize = datasize;
|
|
|
|
if(datasize)
|
|
memcpy(part->data, data, datasize);
|
|
part->data[datasize] = '\0'; /* Set a null terminator as sentinel. */
|
|
|
|
part->readfunc = mime_mem_read;
|
|
part->seekfunc = mime_mem_seek;
|
|
part->freefunc = mime_mem_free;
|
|
part->flags |= MIME_FAST_READ;
|
|
part->kind = MIMEKIND_DATA;
|
|
}
|
|
|
|
return CURLE_OK;
|
|
}
|
|
|
|
/* Set mime part content from named local file. */
|
|
CURLcode curl_mime_filedata(curl_mimepart *part, const char *filename)
|
|
{
|
|
CURLcode result = CURLE_OK;
|
|
|
|
if(!part)
|
|
return CURLE_BAD_FUNCTION_ARGUMENT;
|
|
|
|
cleanup_part_content(part);
|
|
|
|
if(filename) {
|
|
char *base;
|
|
struct_stat sbuf;
|
|
|
|
if(stat(filename, &sbuf) || access(filename, R_OK))
|
|
result = CURLE_READ_ERROR;
|
|
|
|
part->data = strdup(filename);
|
|
if(!part->data)
|
|
result = CURLE_OUT_OF_MEMORY;
|
|
|
|
part->datasize = -1;
|
|
if(!result && S_ISREG(sbuf.st_mode)) {
|
|
part->datasize = filesize(filename, sbuf);
|
|
part->seekfunc = mime_file_seek;
|
|
}
|
|
|
|
part->readfunc = mime_file_read;
|
|
part->freefunc = mime_file_free;
|
|
part->kind = MIMEKIND_FILE;
|
|
|
|
/* As a side effect, set the filename to the current file's base name.
|
|
It is possible to withdraw this by explicitly calling
|
|
curl_mime_filename() with a NULL filename argument after the current
|
|
call. */
|
|
base = strippath(filename);
|
|
if(!base)
|
|
result = CURLE_OUT_OF_MEMORY;
|
|
else {
|
|
CURLcode res = curl_mime_filename(part, base);
|
|
|
|
if(res)
|
|
result = res;
|
|
free(base);
|
|
}
|
|
}
|
|
return result;
|
|
}
|
|
|
|
/* Set mime part type. */
|
|
CURLcode curl_mime_type(curl_mimepart *part, const char *mimetype)
|
|
{
|
|
if(!part)
|
|
return CURLE_BAD_FUNCTION_ARGUMENT;
|
|
|
|
Curl_safefree(part->mimetype);
|
|
part->mimetype = NULL;
|
|
|
|
if(mimetype) {
|
|
part->mimetype = strdup(mimetype);
|
|
if(!part->mimetype)
|
|
return CURLE_OUT_OF_MEMORY;
|
|
}
|
|
|
|
return CURLE_OK;
|
|
}
|
|
|
|
/* Set mime data transfer encoder. */
|
|
CURLcode curl_mime_encoder(curl_mimepart *part, const char *encoding)
|
|
{
|
|
CURLcode result = CURLE_BAD_FUNCTION_ARGUMENT;
|
|
const struct mime_encoder *mep;
|
|
|
|
if(!part)
|
|
return result;
|
|
|
|
part->encoder = NULL;
|
|
|
|
if(!encoding)
|
|
return CURLE_OK; /* Removing current encoder. */
|
|
|
|
for(mep = encoders; mep->name; mep++)
|
|
if(strcasecompare(encoding, mep->name)) {
|
|
part->encoder = mep;
|
|
result = CURLE_OK;
|
|
}
|
|
|
|
return result;
|
|
}
|
|
|
|
/* Set mime part headers. */
|
|
CURLcode curl_mime_headers(curl_mimepart *part,
|
|
struct curl_slist *headers, int take_ownership)
|
|
{
|
|
if(!part)
|
|
return CURLE_BAD_FUNCTION_ARGUMENT;
|
|
|
|
if(part->flags & MIME_USERHEADERS_OWNER) {
|
|
if(part->userheaders != headers) /* Allow setting twice the same list. */
|
|
curl_slist_free_all(part->userheaders);
|
|
part->flags &= ~MIME_USERHEADERS_OWNER;
|
|
}
|
|
part->userheaders = headers;
|
|
if(headers && take_ownership)
|
|
part->flags |= MIME_USERHEADERS_OWNER;
|
|
return CURLE_OK;
|
|
}
|
|
|
|
/* Set mime part content from callback. */
|
|
CURLcode curl_mime_data_cb(curl_mimepart *part, curl_off_t datasize,
|
|
curl_read_callback readfunc,
|
|
curl_seek_callback seekfunc,
|
|
curl_free_callback freefunc, void *arg)
|
|
{
|
|
if(!part)
|
|
return CURLE_BAD_FUNCTION_ARGUMENT;
|
|
|
|
cleanup_part_content(part);
|
|
|
|
if(readfunc) {
|
|
part->readfunc = readfunc;
|
|
part->seekfunc = seekfunc;
|
|
part->freefunc = freefunc;
|
|
part->arg = arg;
|
|
part->datasize = datasize;
|
|
part->kind = MIMEKIND_CALLBACK;
|
|
}
|
|
|
|
return CURLE_OK;
|
|
}
|
|
|
|
/* Set mime part content from subparts. */
|
|
CURLcode Curl_mime_set_subparts(curl_mimepart *part,
|
|
curl_mime *subparts, int take_ownership)
|
|
{
|
|
curl_mime *root;
|
|
|
|
if(!part)
|
|
return CURLE_BAD_FUNCTION_ARGUMENT;
|
|
|
|
/* Accept setting twice the same subparts. */
|
|
if(part->kind == MIMEKIND_MULTIPART && part->arg == subparts)
|
|
return CURLE_OK;
|
|
|
|
cleanup_part_content(part);
|
|
|
|
if(subparts) {
|
|
/* Must belong to the same data handle. */
|
|
if(part->easy && subparts->easy && part->easy != subparts->easy)
|
|
return CURLE_BAD_FUNCTION_ARGUMENT;
|
|
|
|
/* Should not have been attached already. */
|
|
if(subparts->parent)
|
|
return CURLE_BAD_FUNCTION_ARGUMENT;
|
|
|
|
/* Should not be the part's root. */
|
|
root = part->parent;
|
|
if(root) {
|
|
while(root->parent && root->parent->parent)
|
|
root = root->parent->parent;
|
|
if(subparts == root) {
|
|
if(part->easy)
|
|
failf(part->easy, "Can't add itself as a subpart!");
|
|
return CURLE_BAD_FUNCTION_ARGUMENT;
|
|
}
|
|
}
|
|
|
|
subparts->parent = part;
|
|
/* Subparts are processed internally: no read callback. */
|
|
part->seekfunc = mime_subparts_seek;
|
|
part->freefunc = take_ownership? mime_subparts_free: mime_subparts_unbind;
|
|
part->arg = subparts;
|
|
part->datasize = -1;
|
|
part->kind = MIMEKIND_MULTIPART;
|
|
}
|
|
|
|
return CURLE_OK;
|
|
}
|
|
|
|
CURLcode curl_mime_subparts(curl_mimepart *part, curl_mime *subparts)
|
|
{
|
|
return Curl_mime_set_subparts(part, subparts, TRUE);
|
|
}
|
|
|
|
|
|
/* Readback from top mime. */
|
|
/* Argument is the dummy top part. */
|
|
size_t Curl_mime_read(char *buffer, size_t size, size_t nitems, void *instream)
|
|
{
|
|
curl_mimepart *part = (curl_mimepart *) instream;
|
|
size_t ret;
|
|
bool hasread;
|
|
|
|
(void) size; /* Always 1. */
|
|
|
|
do {
|
|
hasread = FALSE;
|
|
ret = readback_part(part, buffer, nitems, &hasread);
|
|
/*
|
|
* If this is not possible to get some data without calling more than
|
|
* one read callback (probably because a content encoder is not able to
|
|
* deliver a new bunch for the few data accumulated so far), force another
|
|
* read until we get enough data or a special exit code.
|
|
*/
|
|
} while(ret == STOP_FILLING);
|
|
|
|
return ret;
|
|
}
|
|
|
|
/* Rewind mime stream. */
|
|
CURLcode Curl_mime_rewind(curl_mimepart *part)
|
|
{
|
|
return mime_part_rewind(part) == CURL_SEEKFUNC_OK?
|
|
CURLE_OK: CURLE_SEND_FAIL_REWIND;
|
|
}
|
|
|
|
/* Compute header list size. */
|
|
static size_t slist_size(struct curl_slist *s,
|
|
size_t overhead, const char *skip)
|
|
{
|
|
size_t size = 0;
|
|
size_t skiplen = skip? strlen(skip): 0;
|
|
|
|
for(; s; s = s->next)
|
|
if(!skip || !match_header(s, skip, skiplen))
|
|
size += strlen(s->data) + overhead;
|
|
return size;
|
|
}
|
|
|
|
/* Get/compute multipart size. */
|
|
static curl_off_t multipart_size(curl_mime *mime)
|
|
{
|
|
curl_off_t size;
|
|
size_t boundarysize;
|
|
curl_mimepart *part;
|
|
|
|
if(!mime)
|
|
return 0; /* Not present -> empty. */
|
|
|
|
boundarysize = 4 + strlen(mime->boundary) + 2;
|
|
size = boundarysize; /* Final boundary - CRLF after headers. */
|
|
|
|
for(part = mime->firstpart; part; part = part->nextpart) {
|
|
curl_off_t sz = Curl_mime_size(part);
|
|
|
|
if(sz < 0)
|
|
size = sz;
|
|
|
|
if(size >= 0)
|
|
size += boundarysize + sz;
|
|
}
|
|
|
|
return size;
|
|
}
|
|
|
|
/* Get/compute mime size. */
|
|
curl_off_t Curl_mime_size(curl_mimepart *part)
|
|
{
|
|
curl_off_t size;
|
|
|
|
if(part->kind == MIMEKIND_MULTIPART)
|
|
part->datasize = multipart_size(part->arg);
|
|
|
|
size = part->datasize;
|
|
|
|
if(part->encoder)
|
|
size = part->encoder->sizefunc(part);
|
|
|
|
if(size >= 0 && !(part->flags & MIME_BODY_ONLY)) {
|
|
/* Compute total part size. */
|
|
size += slist_size(part->curlheaders, 2, NULL);
|
|
size += slist_size(part->userheaders, 2, "Content-Type");
|
|
size += 2; /* CRLF after headers. */
|
|
}
|
|
return size;
|
|
}
|
|
|
|
/* Add a header. */
|
|
/* VARARGS2 */
|
|
CURLcode Curl_mime_add_header(struct curl_slist **slp, const char *fmt, ...)
|
|
{
|
|
struct curl_slist *hdr = NULL;
|
|
char *s = NULL;
|
|
va_list ap;
|
|
|
|
va_start(ap, fmt);
|
|
s = curl_mvaprintf(fmt, ap);
|
|
va_end(ap);
|
|
|
|
if(s) {
|
|
hdr = Curl_slist_append_nodup(*slp, s);
|
|
if(hdr)
|
|
*slp = hdr;
|
|
else
|
|
free(s);
|
|
}
|
|
|
|
return hdr? CURLE_OK: CURLE_OUT_OF_MEMORY;
|
|
}
|
|
|
|
/* Add a content type header. */
|
|
static CURLcode add_content_type(struct curl_slist **slp,
|
|
const char *type, const char *boundary)
|
|
{
|
|
return Curl_mime_add_header(slp, "Content-Type: %s%s%s", type,
|
|
boundary? "; boundary=": "",
|
|
boundary? boundary: "");
|
|
}
|
|
|
|
const char *Curl_mime_contenttype(const char *filename)
|
|
{
|
|
/*
|
|
* If no content type was specified, we scan through a few well-known
|
|
* extensions and pick the first we match!
|
|
*/
|
|
struct ContentType {
|
|
const char *extension;
|
|
const char *type;
|
|
};
|
|
static const struct ContentType ctts[] = {
|
|
{".gif", "image/gif"},
|
|
{".jpg", "image/jpeg"},
|
|
{".jpeg", "image/jpeg"},
|
|
{".png", "image/png"},
|
|
{".svg", "image/svg+xml"},
|
|
{".txt", "text/plain"},
|
|
{".htm", "text/html"},
|
|
{".html", "text/html"},
|
|
{".pdf", "application/pdf"},
|
|
{".xml", "application/xml"}
|
|
};
|
|
|
|
if(filename) {
|
|
size_t len1 = strlen(filename);
|
|
const char *nameend = filename + len1;
|
|
unsigned int i;
|
|
|
|
for(i = 0; i < sizeof(ctts) / sizeof(ctts[0]); i++) {
|
|
size_t len2 = strlen(ctts[i].extension);
|
|
|
|
if(len1 >= len2 && strcasecompare(nameend - len2, ctts[i].extension))
|
|
return ctts[i].type;
|
|
}
|
|
}
|
|
return NULL;
|
|
}
|
|
|
|
static bool content_type_match(const char *contenttype, const char *target)
|
|
{
|
|
size_t len = strlen(target);
|
|
|
|
if(contenttype && strncasecompare(contenttype, target, len))
|
|
switch(contenttype[len]) {
|
|
case '\0':
|
|
case '\t':
|
|
case '\r':
|
|
case '\n':
|
|
case ' ':
|
|
case ';':
|
|
return TRUE;
|
|
}
|
|
return FALSE;
|
|
}
|
|
|
|
CURLcode Curl_mime_prepare_headers(curl_mimepart *part,
|
|
const char *contenttype,
|
|
const char *disposition,
|
|
enum mimestrategy strategy)
|
|
{
|
|
curl_mime *mime = NULL;
|
|
const char *boundary = NULL;
|
|
char *customct;
|
|
const char *cte = NULL;
|
|
CURLcode ret = CURLE_OK;
|
|
|
|
/* Get rid of previously prepared headers. */
|
|
curl_slist_free_all(part->curlheaders);
|
|
part->curlheaders = NULL;
|
|
|
|
/* Be sure we won't access old headers later. */
|
|
if(part->state.state == MIMESTATE_CURLHEADERS)
|
|
mimesetstate(&part->state, MIMESTATE_CURLHEADERS, NULL);
|
|
|
|
/* Check if content type is specified. */
|
|
customct = part->mimetype;
|
|
if(!customct)
|
|
customct = search_header(part->userheaders, "Content-Type");
|
|
if(customct)
|
|
contenttype = customct;
|
|
|
|
/* If content type is not specified, try to determine it. */
|
|
if(!contenttype) {
|
|
switch(part->kind) {
|
|
case MIMEKIND_MULTIPART:
|
|
contenttype = MULTIPART_CONTENTTYPE_DEFAULT;
|
|
break;
|
|
case MIMEKIND_FILE:
|
|
contenttype = Curl_mime_contenttype(part->filename);
|
|
if(!contenttype)
|
|
contenttype = Curl_mime_contenttype(part->data);
|
|
if(!contenttype && part->filename)
|
|
contenttype = FILE_CONTENTTYPE_DEFAULT;
|
|
break;
|
|
default:
|
|
contenttype = Curl_mime_contenttype(part->filename);
|
|
break;
|
|
}
|
|
}
|
|
|
|
if(part->kind == MIMEKIND_MULTIPART) {
|
|
mime = (curl_mime *) part->arg;
|
|
if(mime)
|
|
boundary = mime->boundary;
|
|
}
|
|
else if(contenttype && !customct &&
|
|
content_type_match(contenttype, "text/plain"))
|
|
if(strategy == MIMESTRATEGY_MAIL || !part->filename)
|
|
contenttype = NULL;
|
|
|
|
/* Issue content-disposition header only if not already set by caller. */
|
|
if(!search_header(part->userheaders, "Content-Disposition")) {
|
|
if(!disposition)
|
|
if(part->filename || part->name ||
|
|
(contenttype && !strncasecompare(contenttype, "multipart/", 10)))
|
|
disposition = DISPOSITION_DEFAULT;
|
|
if(disposition && curl_strequal(disposition, "attachment") &&
|
|
!part->name && !part->filename)
|
|
disposition = NULL;
|
|
if(disposition) {
|
|
char *name = NULL;
|
|
char *filename = NULL;
|
|
|
|
if(part->name) {
|
|
name = escape_string(part->name);
|
|
if(!name)
|
|
ret = CURLE_OUT_OF_MEMORY;
|
|
}
|
|
if(!ret && part->filename) {
|
|
filename = escape_string(part->filename);
|
|
if(!filename)
|
|
ret = CURLE_OUT_OF_MEMORY;
|
|
}
|
|
if(!ret)
|
|
ret = Curl_mime_add_header(&part->curlheaders,
|
|
"Content-Disposition: %s%s%s%s%s%s%s",
|
|
disposition,
|
|
name? "; name=\"": "",
|
|
name? name: "",
|
|
name? "\"": "",
|
|
filename? "; filename=\"": "",
|
|
filename? filename: "",
|
|
filename? "\"": "");
|
|
Curl_safefree(name);
|
|
Curl_safefree(filename);
|
|
if(ret)
|
|
return ret;
|
|
}
|
|
}
|
|
|
|
/* Issue Content-Type header. */
|
|
if(contenttype) {
|
|
ret = add_content_type(&part->curlheaders, contenttype, boundary);
|
|
if(ret)
|
|
return ret;
|
|
}
|
|
|
|
/* Content-Transfer-Encoding header. */
|
|
if(!search_header(part->userheaders, "Content-Transfer-Encoding")) {
|
|
if(part->encoder)
|
|
cte = part->encoder->name;
|
|
else if(contenttype && strategy == MIMESTRATEGY_MAIL &&
|
|
part->kind != MIMEKIND_MULTIPART)
|
|
cte = "8bit";
|
|
if(cte) {
|
|
ret = Curl_mime_add_header(&part->curlheaders,
|
|
"Content-Transfer-Encoding: %s", cte);
|
|
if(ret)
|
|
return ret;
|
|
}
|
|
}
|
|
|
|
/* If we were reading curl-generated headers, restart with new ones (this
|
|
should not occur). */
|
|
if(part->state.state == MIMESTATE_CURLHEADERS)
|
|
mimesetstate(&part->state, MIMESTATE_CURLHEADERS, part->curlheaders);
|
|
|
|
/* Process subparts. */
|
|
if(part->kind == MIMEKIND_MULTIPART && mime) {
|
|
curl_mimepart *subpart;
|
|
|
|
disposition = NULL;
|
|
if(content_type_match(contenttype, "multipart/form-data"))
|
|
disposition = "form-data";
|
|
for(subpart = mime->firstpart; subpart; subpart = subpart->nextpart) {
|
|
ret = Curl_mime_prepare_headers(subpart, NULL, disposition, strategy);
|
|
if(ret)
|
|
return ret;
|
|
}
|
|
}
|
|
return ret;
|
|
}
|
|
|
|
/* Recursively reset paused status in the given part. */
|
|
void Curl_mime_unpause(curl_mimepart *part)
|
|
{
|
|
if(part) {
|
|
if(part->lastreadstatus == CURL_READFUNC_PAUSE)
|
|
part->lastreadstatus = 1; /* Successful read status. */
|
|
if(part->kind == MIMEKIND_MULTIPART) {
|
|
curl_mime *mime = (curl_mime *) part->arg;
|
|
|
|
if(mime) {
|
|
curl_mimepart *subpart;
|
|
|
|
for(subpart = mime->firstpart; subpart; subpart = subpart->nextpart)
|
|
Curl_mime_unpause(subpart);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
|
|
#else /* !CURL_DISABLE_HTTP || !CURL_DISABLE_SMTP || !CURL_DISABLE_IMAP */
|
|
|
|
/* Mime not compiled in: define stubs for externally-referenced functions. */
|
|
curl_mime *curl_mime_init(CURL *easy)
|
|
{
|
|
(void) easy;
|
|
return NULL;
|
|
}
|
|
|
|
void curl_mime_free(curl_mime *mime)
|
|
{
|
|
(void) mime;
|
|
}
|
|
|
|
curl_mimepart *curl_mime_addpart(curl_mime *mime)
|
|
{
|
|
(void) mime;
|
|
return NULL;
|
|
}
|
|
|
|
CURLcode curl_mime_name(curl_mimepart *part, const char *name)
|
|
{
|
|
(void) part;
|
|
(void) name;
|
|
return CURLE_NOT_BUILT_IN;
|
|
}
|
|
|
|
CURLcode curl_mime_filename(curl_mimepart *part, const char *filename)
|
|
{
|
|
(void) part;
|
|
(void) filename;
|
|
return CURLE_NOT_BUILT_IN;
|
|
}
|
|
|
|
CURLcode curl_mime_type(curl_mimepart *part, const char *mimetype)
|
|
{
|
|
(void) part;
|
|
(void) mimetype;
|
|
return CURLE_NOT_BUILT_IN;
|
|
}
|
|
|
|
CURLcode curl_mime_encoder(curl_mimepart *part, const char *encoding)
|
|
{
|
|
(void) part;
|
|
(void) encoding;
|
|
return CURLE_NOT_BUILT_IN;
|
|
}
|
|
|
|
CURLcode curl_mime_data(curl_mimepart *part,
|
|
const char *data, size_t datasize)
|
|
{
|
|
(void) part;
|
|
(void) data;
|
|
(void) datasize;
|
|
return CURLE_NOT_BUILT_IN;
|
|
}
|
|
|
|
CURLcode curl_mime_filedata(curl_mimepart *part, const char *filename)
|
|
{
|
|
(void) part;
|
|
(void) filename;
|
|
return CURLE_NOT_BUILT_IN;
|
|
}
|
|
|
|
CURLcode curl_mime_data_cb(curl_mimepart *part,
|
|
curl_off_t datasize,
|
|
curl_read_callback readfunc,
|
|
curl_seek_callback seekfunc,
|
|
curl_free_callback freefunc,
|
|
void *arg)
|
|
{
|
|
(void) part;
|
|
(void) datasize;
|
|
(void) readfunc;
|
|
(void) seekfunc;
|
|
(void) freefunc;
|
|
(void) arg;
|
|
return CURLE_NOT_BUILT_IN;
|
|
}
|
|
|
|
CURLcode curl_mime_subparts(curl_mimepart *part, curl_mime *subparts)
|
|
{
|
|
(void) part;
|
|
(void) subparts;
|
|
return CURLE_NOT_BUILT_IN;
|
|
}
|
|
|
|
CURLcode curl_mime_headers(curl_mimepart *part,
|
|
struct curl_slist *headers, int take_ownership)
|
|
{
|
|
(void) part;
|
|
(void) headers;
|
|
(void) take_ownership;
|
|
return CURLE_NOT_BUILT_IN;
|
|
}
|
|
|
|
CURLcode Curl_mime_add_header(struct curl_slist **slp, const char *fmt, ...)
|
|
{
|
|
(void)slp;
|
|
(void)fmt;
|
|
return CURLE_NOT_BUILT_IN;
|
|
}
|
|
|
|
#endif /* if disabled */
|