mime: new MIME API.

Available in HTTP, SMTP and IMAP.
Deprecates the FORM API.
See CURLOPT_MIMEPOST.
Lib code and associated documentation.
This commit is contained in:
Patrick Monnerat 2017-09-02 17:47:10 +01:00
parent 5bae72734b
commit ce0881edee
38 changed files with 2946 additions and 994 deletions

View file

@ -32,6 +32,8 @@
#include "urldata.h" /* for struct Curl_easy */
#include "formdata.h"
#include "mime.h"
#include "non-ascii.h"
#include "vtls/vtls.h"
#include "strcase.h"
#include "sendf.h"
@ -42,13 +44,6 @@
#include "curl_memory.h"
#include "memdebug.h"
#ifndef HAVE_BASENAME
static char *Curl_basename(char *path);
#define basename(x) Curl_basename((x))
#endif
static size_t readfromfile(struct Form *form, char *buffer, size_t size);
static CURLcode formboundary(struct Curl_easy *data, char *buffer, size_t len);
/* What kind of Content-Type to use on un-specified files with unrecognized
extensions. */
@ -746,211 +741,6 @@ CURLFORMcode curl_formadd(struct curl_httppost **httppost,
return result;
}
#ifdef __VMS
#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 == NULL)
return 0;
count = 0;
ret_stat = 1;
while(ret_stat > 0) {
ret_stat = fread(buffer, 1, sizeof(buffer), file);
if(ret_stat != 0)
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;
}
}
#endif
#ifndef __VMS
#define filesize(name, stat_data) (stat_data.st_size)
#else
/* Getting the expected file size needs help on VMS */
#define filesize(name, stat_data) VmsSpecialSize(name, &stat_data)
#endif
/*
* AddFormData() adds a chunk of data to the FormData linked list.
*
* size is incremented by the chunk length, unless it is NULL
*/
static CURLcode AddFormData(struct FormData **formp,
enum formtype type,
const void *line,
curl_off_t length,
curl_off_t *size)
{
struct FormData *newform;
char *alloc2 = NULL;
CURLcode result = CURLE_OK;
if(length < 0 || (size && *size < 0))
return CURLE_BAD_FUNCTION_ARGUMENT;
newform = malloc(sizeof(struct FormData));
if(!newform)
return CURLE_OUT_OF_MEMORY;
newform->next = NULL;
if(type <= FORM_CONTENT) {
/* we make it easier for plain strings: */
if(!length)
length = strlen((char *)line);
#if (SIZEOF_SIZE_T < CURL_SIZEOF_CURL_OFF_T)
else if(length >= (curl_off_t)(size_t)-1) {
result = CURLE_BAD_FUNCTION_ARGUMENT;
goto error;
}
#endif
if(type != FORM_DATAMEM) {
newform->line = malloc((size_t)length+1);
if(!newform->line) {
result = CURLE_OUT_OF_MEMORY;
goto error;
}
alloc2 = newform->line;
memcpy(newform->line, line, (size_t)length);
/* zero terminate for easier debugging */
newform->line[(size_t)length]=0;
}
else {
newform->line = (char *)line;
type = FORM_DATA; /* in all other aspects this is just FORM_DATA */
}
newform->length = (size_t)length;
}
else
/* For callbacks and files we don't have any actual data so we just keep a
pointer to whatever this points to */
newform->line = (char *)line;
newform->type = type;
if(size) {
if(type != FORM_FILE)
/* for static content as well as callback data we add the size given
as input argument */
*size += length;
else {
/* Since this is a file to be uploaded here, add the size of the actual
file */
if(strcmp("-", newform->line)) {
struct_stat file;
if(!stat(newform->line, &file) && !S_ISDIR(file.st_mode))
*size += filesize(newform->line, file);
else {
result = CURLE_BAD_FUNCTION_ARGUMENT;
goto error;
}
}
}
}
if(*formp) {
(*formp)->next = newform;
*formp = newform;
}
else
*formp = newform;
return CURLE_OK;
error:
if(newform)
free(newform);
if(alloc2)
free(alloc2);
return result;
}
/*
* AddFormDataf() adds printf()-style formatted data to the formdata chain.
*/
static CURLcode AddFormDataf(struct FormData **formp,
curl_off_t *size,
const char *fmt, ...)
{
char *s;
CURLcode result;
va_list ap;
va_start(ap, fmt);
s = curl_mvaprintf(fmt, ap);
va_end(ap);
if(!s)
return CURLE_OUT_OF_MEMORY;
result = AddFormData(formp, FORM_DATAMEM, s, 0, size);
if(result)
free(s);
return result;
}
/*
* Curl_formclean() is used from http.c, this cleans a built FormData linked
* list
*/
void Curl_formclean(struct FormData **form_ptr)
{
struct FormData *next, *form;
form = *form_ptr;
if(!form)
return;
do {
next=form->next; /* the following form line */
if(form->type <= FORM_CONTENT)
free(form->line); /* free the line */
free(form); /* free the struct */
form = next;
} while(form); /* continue */
*form_ptr = NULL;
}
/*
* curl_formget()
* Serialize a curl_httppost struct.
@ -962,42 +752,34 @@ int curl_formget(struct curl_httppost *form, void *arg,
curl_formget_callback append)
{
CURLcode result;
curl_off_t size;
struct FormData *data, *ptr;
struct Curl_mimepart toppart;
result = Curl_getformdata(NULL, &data, form, NULL, &size);
if(result)
return (int)result;
Curl_mime_initpart(&toppart, NULL); /* default form is empty */
result = Curl_getformdata(NULL, &toppart, form, NULL);
if(!result)
result = Curl_mime_prepare_headers(&toppart, "multipart/form-data",
NULL, MIMESTRATEGY_FORM);
for(ptr = data; ptr; ptr = ptr->next) {
if((ptr->type == FORM_FILE) || (ptr->type == FORM_CALLBACK)) {
char buffer[8192];
size_t nread;
struct Form temp;
while(!result) {
char buffer[8192];
size_t nread = Curl_mime_read(buffer, 1, sizeof buffer, &toppart);
Curl_FormInit(&temp, ptr);
if(!nread)
break;
do {
nread = readfromfile(&temp, buffer, sizeof(buffer));
if((nread == (size_t) -1) ||
(nread > sizeof(buffer)) ||
(nread != append(arg, buffer, nread))) {
if(temp.fp)
fclose(temp.fp);
Curl_formclean(&data);
return -1;
}
} while(nread);
}
else {
if(ptr->length != append(arg, ptr->line, ptr->length)) {
Curl_formclean(&data);
return -1;
}
switch(nread) {
default:
if(append(arg, buffer, nread) != nread)
result = CURLE_READ_ERROR;
break;
case CURL_READFUNC_ABORT:
case CURL_READFUNC_PAUSE:
break;
}
}
Curl_formclean(&data);
return 0;
Curl_mime_cleanpart(&toppart);
return (int) result;
}
/*
@ -1031,118 +813,11 @@ void curl_formfree(struct curl_httppost *form)
} while(form); /* continue */
}
#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
implementaion 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;
}
#endif
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 ! */
}
static CURLcode formdata_add_filename(const struct curl_httppost *file,
struct FormData **form,
curl_off_t *size)
{
CURLcode result = CURLE_OK;
char *filename = file->showfilename;
char *filebasename = NULL;
char *filename_escaped = NULL;
if(!filename) {
filebasename = strippath(file->contents);
if(!filebasename)
return CURLE_OUT_OF_MEMORY;
filename = filebasename;
}
if(strchr(filename, '\\') || strchr(filename, '"')) {
char *p0, *p1;
/* filename need be escaped */
filename_escaped = malloc(strlen(filename)*2+1);
if(!filename_escaped) {
free(filebasename);
return CURLE_OUT_OF_MEMORY;
}
p0 = filename_escaped;
p1 = filename;
while(*p1) {
if(*p1 == '\\' || *p1 == '"')
*p0++ = '\\';
*p0++ = *p1++;
}
*p0 = '\0';
filename = filename_escaped;
}
result = AddFormDataf(form, size,
"; filename=\"%s\"",
filename);
free(filename_escaped);
free(filebasename);
return result;
}
/*
* Curl_getformdata() converts a linked list of "meta data" into a complete
* (possibly huge) multipart formdata. The input list is in 'post', while the
* output resulting linked lists gets stored in '*finalform'. *sizep will get
* the total size of the whole POST.
* A multipart/form_data content-type is built, unless a custom content-type
* is passed in 'custom_content_type'.
* Curl_getformdata() converts a linked list of "meta data" into a mime
* structure. The input list is in 'post', while the output is stored in
* mime part at '*finalform'.
*
* This function will not do a failf() for the potential memory failures but
* should for all other errors it spots. Just note that this function MAY get
@ -1150,422 +825,114 @@ static CURLcode formdata_add_filename(const struct curl_httppost *file,
*/
CURLcode Curl_getformdata(struct Curl_easy *data,
struct FormData **finalform,
struct Curl_mimepart *finalform,
struct curl_httppost *post,
const char *custom_content_type,
curl_off_t *sizep)
curl_read_callback fread_func)
{
struct FormData *form = NULL;
struct FormData *firstform;
struct curl_httppost *file;
CURLcode result = CURLE_OK;
curl_off_t size = 0; /* support potentially ENORMOUS formposts */
char fileboundary[42];
struct curl_slist *curList;
char boundary[42];
struct Curl_mime *form = NULL;
struct Curl_mime *multipart;
struct Curl_mimepart *part;
struct curl_httppost *file;
*finalform = NULL; /* default form is empty */
Curl_mime_cleanpart(finalform); /* default form is empty */
if(!post)
return result; /* no input => no output! */
result = formboundary(data, boundary, sizeof(boundary));
if(result)
return result;
form = curl_mime_init(data);
if(!form)
result = CURLE_OUT_OF_MEMORY;
/* Make the first line of the output */
result = AddFormDataf(&form, NULL,
"%s; boundary=%s\r\n",
custom_content_type?custom_content_type:
"Content-Type: multipart/form-data",
boundary);
if(result) {
return result;
}
/* we DO NOT include that line in the total size of the POST, since it'll be
part of the header! */
firstform = form;
do {
if(size) {
result = AddFormDataf(&form, &size, "\r\n");
if(result)
break;
}
/* boundary */
result = AddFormDataf(&form, &size, "--%s\r\n", boundary);
if(result)
break;
/* Maybe later this should be disabled when a custom_content_type is
passed, since Content-Disposition is not meaningful for all multipart
types.
*/
result = AddFormDataf(&form, &size,
"Content-Disposition: form-data; name=\"");
if(result)
break;
result = AddFormData(&form, FORM_DATA, post->name, post->namelength,
&size);
if(result)
break;
result = AddFormDataf(&form, &size, "\"");
if(result)
break;
if(post->more) {
/* If used, this is a link to more file names, we must then do
the magic to include several files with the same field name */
result = formboundary(data, fileboundary, sizeof(fileboundary));
if(result) {
break;
}
result = AddFormDataf(&form, &size,
"\r\nContent-Type: multipart/mixed;"
" boundary=%s\r\n",
fileboundary);
if(result)
break;
}
file = post;
do {
/* If 'showfilename' is set, that is a faked name passed on to us
to use to in the formpost. If that is not set, the actually used
local file name should be added. */
if(post->more) {
/* if multiple-file */
result = AddFormDataf(&form, &size,
"\r\n--%s\r\nContent-Disposition: "
"attachment",
fileboundary);
if(result)
break;
result = formdata_add_filename(file, &form, &size);
if(result)
break;
}
else if(post->flags & (HTTPPOST_FILENAME|HTTPPOST_BUFFER|
HTTPPOST_CALLBACK)) {
/* it should be noted that for the HTTPPOST_FILENAME and
HTTPPOST_CALLBACK cases the ->showfilename struct member is always
assigned at this point */
if(post->showfilename || (post->flags & HTTPPOST_FILENAME)) {
result = formdata_add_filename(post, &form, &size);
}
if(result)
break;
}
if(file->contenttype) {
/* we have a specified type */
result = AddFormDataf(&form, &size,
"\r\nContent-Type: %s",
file->contenttype);
if(result)
break;
}
curList = file->contentheader;
while(curList) {
/* Process the additional headers specified for this form */
result = AddFormDataf(&form, &size, "\r\n%s", curList->data);
if(result)
break;
curList = curList->next;
}
if(result)
break;
result = AddFormDataf(&form, &size, "\r\n\r\n");
if(result)
break;
if((post->flags & HTTPPOST_FILENAME) ||
(post->flags & HTTPPOST_READFILE)) {
/* we should include the contents from the specified file */
FILE *fileread;
fileread = !strcmp("-", file->contents)?
stdin:fopen(file->contents, "rb"); /* binary read for win32 */
/*
* VMS: This only allows for stream files on VMS. Stream files are
* OK, as are FIXED & VAR files WITHOUT implied CC For implied CC,
* every record needs to have a \n appended & 1 added to SIZE
*/
if(fileread) {
if(fileread != stdin) {
/* close the file */
fclose(fileread);
/* add the file name only - for later reading from this */
result = AddFormData(&form, FORM_FILE, file->contents, 0, &size);
}
else {
/* When uploading from stdin, we can't know the size of the file,
* thus must read the full file as before. We *could* use chunked
* transfer-encoding, but that only works for HTTP 1.1 and we
* can't be sure we work with such a server.
*/
size_t nread;
char buffer[512];
while((nread = fread(buffer, 1, sizeof(buffer), fileread)) != 0) {
result = AddFormData(&form, FORM_CONTENT, buffer, nread, &size);
if(result || feof(fileread) || ferror(fileread))
break;
}
}
}
else {
if(data)
failf(data, "couldn't open file \"%s\"", file->contents);
*finalform = NULL;
result = CURLE_READ_ERROR;
}
}
else if(post->flags & HTTPPOST_BUFFER)
/* include contents of buffer */
result = AddFormData(&form, FORM_CONTENT, post->buffer,
post->bufferlength, &size);
else if(post->flags & HTTPPOST_CALLBACK)
/* the contents should be read with the callback and the size is set
with the contentslength */
result = AddFormData(&form, FORM_CALLBACK, post->userp,
post->flags&CURL_HTTPPOST_LARGE?
post->contentlen:post->contentslength, &size);
else
/* include the contents we got */
result = AddFormData(&form, FORM_CONTENT, post->contents,
post->flags&CURL_HTTPPOST_LARGE?
post->contentlen:post->contentslength, &size);
file = file->more;
} while(file && !result); /* for each specified file for this field */
if(result)
break;
if(post->more) {
/* this was a multiple-file inclusion, make a termination file
boundary: */
result = AddFormDataf(&form, &size,
"\r\n--%s--",
fileboundary);
if(result)
break;
}
post = post->next;
} while(post); /* for each field */
/* end-boundary for everything */
if(!result)
result = AddFormDataf(&form, &size, "\r\n--%s--\r\n", boundary);
result = curl_mime_subparts(finalform, form);
if(result) {
Curl_formclean(&firstform);
return result;
/* Process each top part. */
for(; !result && post; post = post->next) {
/* If we have more than a file here, create a mime subpart and fill it. */
multipart = form;
if(post->more) {
part = curl_mime_addpart(form);
if(!part)
result = CURLE_OUT_OF_MEMORY;
if(!result)
result = curl_mime_name(part, post->name,
post->namelength? post->namelength: -1);
if(!result) {
multipart = curl_mime_init(data);
if(!multipart)
result = CURLE_OUT_OF_MEMORY;
}
if(!result)
result = curl_mime_subparts(part, multipart);
}
/* Generate all the part contents. */
for(file = post; !result && file; file = file->more) {
/* Create the part. */
part = curl_mime_addpart(multipart);
if(!part)
result = CURLE_OUT_OF_MEMORY;
/* Set the headers. */
if(!result)
result = curl_mime_headers(part, file->contentheader, 0);
/* Set the content type. */
if(!result &&file->contenttype)
result = curl_mime_type(part, file->contenttype);
/* Set field name. */
if(!result && !post->more)
result = curl_mime_name(part, post->name,
post->namelength? post->namelength: -1);
/* Process contents. */
if(!result) {
curl_off_t clen = post->contentslength;
if(post->flags & CURL_HTTPPOST_LARGE)
clen = post->contentlen;
if(!clen)
clen = -1;
if(post->flags & (HTTPPOST_FILENAME | HTTPPOST_READFILE)) {
result = curl_mime_filedata(part, file->contents);
if(!result && (post->flags & HTTPPOST_READFILE))
result = curl_mime_filename(part, NULL);
}
else if(post->flags & HTTPPOST_BUFFER)
result = curl_mime_data(part, post->buffer,
post->bufferlength? post->bufferlength: -1);
else if(post->flags & HTTPPOST_CALLBACK)
/* the contents should be read with the callback and the size is set
with the contentslength */
result = curl_mime_data_cb(part, clen,
fread_func, NULL, NULL, post->userp);
else {
result = curl_mime_data(part, post->contents, (ssize_t) clen);
#ifdef CURL_DOES_CONVERSIONS
/* Convert textual contents now. */
if(!result && data && part->datasize)
result = Curl_convert_to_network(data, part->data, part->datasize);
#endif
}
}
/* Set fake file name. */
if(!result && post->showfilename)
if(post->more || (post->flags & (HTTPPOST_FILENAME | HTTPPOST_BUFFER |
HTTPPOST_CALLBACK)))
result = curl_mime_filename(part, post->showfilename);
}
}
*sizep = size;
*finalform = firstform;
if(result)
Curl_mime_cleanpart(finalform);
return result;
}
/*
* Curl_FormInit() inits the struct 'form' points to with the 'formdata'
* and resets the 'sent' counter.
*/
int Curl_FormInit(struct Form *form, struct FormData *formdata)
{
if(!formdata)
return 1; /* error */
form->data = formdata;
form->sent = 0;
form->fp = NULL;
form->fread_func = ZERO_NULL;
return 0;
}
#ifndef __VMS
# define fopen_read fopen
#else
/*
* 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.
*
*/
# define fopen_read vmsfopenread
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");
}
}
#endif
/*
* readfromfile()
*
* The read callback that this function may use can return a value larger than
* 'size' (which then this function returns) that indicates a problem and it
* must be properly dealt with
*/
static size_t readfromfile(struct Form *form, char *buffer,
size_t size)
{
size_t nread;
bool callback = (form->data->type == FORM_CALLBACK)?TRUE:FALSE;
if(callback) {
if(form->fread_func == ZERO_NULL)
return 0;
nread = form->fread_func(buffer, 1, size, form->data->line);
}
else {
if(!form->fp) {
/* this file hasn't yet been opened */
form->fp = fopen_read(form->data->line, "rb"); /* b is for binary */
if(!form->fp)
return (size_t)-1; /* failure */
}
nread = fread(buffer, 1, size, form->fp);
}
if(!nread) {
/* this is the last chunk from the file, move on */
if(form->fp) {
fclose(form->fp);
form->fp = NULL;
}
form->data = form->data->next;
}
return nread;
}
/*
* Curl_FormReader() is the fread() emulation function that will be used to
* deliver the formdata to the transfer loop and then sent away to the peer.
*/
size_t Curl_FormReader(char *buffer,
size_t size,
size_t nitems,
FILE *mydata)
{
struct Form *form;
size_t wantedsize;
size_t gotsize = 0;
form=(struct Form *)mydata;
wantedsize = size * nitems;
if(!form->data)
return 0; /* nothing, error, empty */
if((form->data->type == FORM_FILE) ||
(form->data->type == FORM_CALLBACK)) {
gotsize = readfromfile(form, buffer, wantedsize);
if(gotsize)
/* If positive or -1, return. If zero, continue! */
return gotsize;
}
do {
if((form->data->length - form->sent) > wantedsize - gotsize) {
memcpy(buffer + gotsize, form->data->line + form->sent,
wantedsize - gotsize);
form->sent += wantedsize-gotsize;
return wantedsize;
}
memcpy(buffer+gotsize,
form->data->line + form->sent,
(form->data->length - form->sent) );
gotsize += form->data->length - form->sent;
form->sent = 0;
form->data = form->data->next; /* advance */
} while(form->data && (form->data->type < FORM_CALLBACK));
/* If we got an empty line and we have more data, we proceed to the next
line immediately to avoid returning zero before we've reached the end. */
return gotsize;
}
/*
* Curl_formpostheader() returns the first line of the formpost, the
* request-header part (which is not part of the request-body like the rest of
* the post).
*/
char *Curl_formpostheader(void *formp, size_t *len)
{
char *header;
struct Form *form=(struct Form *)formp;
if(!form->data)
return NULL; /* nothing, ERROR! */
header = form->data->line;
*len = form->data->length;
form->data = form->data->next; /* advance */
return header;
}
/*
* formboundary() creates a suitable boundary string and returns an allocated
* one.
*/
static CURLcode formboundary(struct Curl_easy *data,
char *buffer, size_t buflen)
{
/* 24 dashes and 16 hexadecimal digits makes 64 bit (18446744073709551615)
combinations */
if(buflen < 41)
return CURLE_BAD_FUNCTION_ARGUMENT;
memset(buffer, '-', 24);
Curl_rand_hex(data, (unsigned char *)&buffer[24], 17);
return CURLE_OK;
}
#else /* CURL_DISABLE_HTTP */
CURLFORMcode curl_formadd(struct curl_httppost **httppost,
struct curl_httppost **last_post,