[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[SCM] GNU Mailutils branch, master, updated. release-2.2-540-g3c5aed3
|
From: |
Sergey Poznyakoff |
|
Subject: |
[SCM] GNU Mailutils branch, master, updated. release-2.2-540-g3c5aed3 |
|
Date: |
Tue, 20 Dec 2011 23:26:41 +0000 |
This is an automated email from the git hooks/post-receive script. It was
generated because a ref change was pushed to the repository containing
the project "GNU Mailutils".
http://git.savannah.gnu.org/cgit/mailutils.git/commit/?id=3c5aed3965c70e6e1a894005b1adc7a407d0aae9
The branch, master has been updated
via 3c5aed3965c70e6e1a894005b1adc7a407d0aae9 (commit)
from c22d07a3d7bc380541bfb2603c6f5a2967ad63c9 (commit)
Those revisions listed above that are new to this repository have
not appeared on any other notification email; so we list those
revisions in full, below.
- Log -----------------------------------------------------------------
commit 3c5aed3965c70e6e1a894005b1adc7a407d0aae9
Author: Sergey Poznyakoff <address@hidden>
Date: Wed Dec 21 00:42:00 2011 +0200
Re-implement parser for RFC-2231 header fields.
The new implementation is able to return all parameters at once,
in an associative array. A subset of parameters can be requested.
The data are automatically converted to the output charset.
In addition, RFC-2047 parser is extended to support language
specifications (RFC-2231, chapter 5).
* include/mailutils/message.h (MU_MIMEHDR_CSINFO)
(MU_MIMEHDR_MULTILINE): Remove, not public anymore.
(mu_mimehdr_get_param,mu_mimehdr_aget_param)
(mu_message_aget_attachment_name)
(mu_message_get_attachment_name): Remove pflags agrument.
* include/mailutils/mime.h (mu_mime_param): New struct.
(mu_rfc2047_decode_param)
(mu_mime_header_parse,mu_mime_header_parse_subset): New proto.
* libmailutils/base/rfc2047.c (_rfc2047_decode_param): New
auxiliary function. Use memory stream to collect data.
(mu_rfc2047_decode): Rewrite as a wrapper around the above.
(mu_rfc2047_decode_param): New function.
* libmailutils/filter/decode.c (mu_decode_filter_args): Pass actual
(instead of maximal) number of arguments to mu_filter_chain_create.
* libmailutils/mime/mimehdr.c: Rewrite from scratch.
* libmailutils/tests/.gitignore: Add mimehdr.
* libmailutils/tests/Makefile.am (noinst_PROGRAMS): Add mimehdr.
(TESTSUITE_AT): Add mimehdr.at.
* libmailutils/tests/mimehdr.at: New test.
* libmailutils/tests/mimehdr.c: New test program.
* libmailutils/tests/testsuite.at: Include.
-----------------------------------------------------------------------
Summary of changes:
include/mailutils/message.h | 19 +-
include/mailutils/mime.h | 18 +-
libmailutils/base/rfc2047.c | 198 ++++---
libmailutils/filter/decode.c | 2 +-
libmailutils/mime/mimehdr.c | 1244 ++++++++++++++++++++++++---------------
libmailutils/tests/.gitignore | 1 +
libmailutils/tests/Makefile.am | 2 +
libmailutils/tests/mimehdr.at | 131 ++++
libmailutils/tests/mimehdr.c | 135 +++++
libmailutils/tests/testsuite.at | 4 +-
10 files changed, 1165 insertions(+), 589 deletions(-)
create mode 100644 libmailutils/tests/mimehdr.at
create mode 100644 libmailutils/tests/mimehdr.c
diff --git a/include/mailutils/message.h b/include/mailutils/message.h
index b448dda..ced7901 100644
--- a/include/mailutils/message.h
+++ b/include/mailutils/message.h
@@ -162,31 +162,20 @@ extern int mu_mime_io_buffer_aget_charset
(mu_mime_io_buffer_t info,
const char **charset);
-/* Bit values for *pflags in functions below */
-#define MU_MIMEHDR_MULTILINE 0x01 /* Parameter was multiline */
-#define MU_MIMEHDR_CSINFO 0x02 /* Parameter contains charset/language
- info */
-
extern int mu_mimehdr_get_disp (const char *str, char *buf, size_t bufsz,
size_t *retsz);
extern int mu_mimehdr_aget_disp (const char *str, char **pvalue);
extern int mu_mimehdr_get_param (const char *str, const char *param,
- char *buf, size_t bufsz, size_t *retsz,
- int *pflags);
+ char *buf, size_t bufsz, size_t *retsz);
extern int mu_mimehdr_aget_param (const char *str, const char *param,
- char **pval, int *pflags);
-extern int mu_mimehdr_decode_param (const char *value, int csinfo,
- const char *charset,
- char **pval, char **plang);
+ char **pval);
extern int mu_mimehdr_aget_decoded_param (const char *str, const char *param,
const char *charset,
char **pval, char **plang);
extern int mu_message_get_attachment_name (mu_message_t, char *name,
- size_t bufsz, size_t* sz,
- int *pflags);
-extern int mu_message_aget_attachment_name (mu_message_t, char **name,
- int *pflags);
+ size_t bufsz, size_t* sz);
+extern int mu_message_aget_attachment_name (mu_message_t, char **name);
extern int mu_message_aget_decoded_attachment_name (mu_message_t msg,
const char *charset,
char **name,
diff --git a/include/mailutils/mime.h b/include/mailutils/mime.h
index d41ed38..9c3d473 100644
--- a/include/mailutils/mime.h
+++ b/include/mailutils/mime.h
@@ -29,6 +29,13 @@
extern "C" {
#endif
+struct mu_mime_param
+{
+ char *lang;
+ char *cset;
+ char *value;
+};
+
int mu_mime_create (mu_mime_t *pmime, mu_message_t msg, int flags);
void mu_mime_destroy (mu_mime_t *pmime);
void mu_mime_ref (mu_mime_t mime);
@@ -49,13 +56,22 @@ int mu_rfc2047_decode (const char *tocode, const char
*fromstr,
int mu_rfc2047_encode (const char *charset, const char *encoding,
const char *text, char **result);
+int mu_rfc2047_decode_param (const char *tocode, const char *input,
+ struct mu_mime_param *param);
int mu_base64_encode (const unsigned char *input, size_t input_len,
unsigned char **output, size_t * output_len);
int mu_base64_decode (const unsigned char *input, size_t input_len,
unsigned char **output, size_t * output_len);
-
+
+
+int mu_mime_header_parse (const char *text, char *charset, char **pvalue,
+ mu_assoc_t *paramtab);
+int mu_mime_header_parse_subset (const char *text, const char *charset,
+ char **pvalue,
+ mu_assoc_t assoc);
+
#ifdef __cplusplus
}
#endif
diff --git a/libmailutils/base/rfc2047.c b/libmailutils/base/rfc2047.c
index 33f7fda..b5b1ee7 100644
--- a/libmailutils/base/rfc2047.c
+++ b/libmailutils/base/rfc2047.c
@@ -28,20 +28,9 @@
#include <mailutils/stream.h>
#include <mailutils/filter.h>
#include <mailutils/errno.h>
+#include <mailutils/mime.h>
#include <mailutils/util.h>
-static int
-realloc_buffer (char **bufp, size_t *bufsizep, size_t incr)
-{
- size_t newsize = *bufsizep + incr;
- char *newp = realloc (*bufp, newsize);
- if (newp == NULL)
- return 1;
- *bufp = newp;
- *bufsizep = newsize;
- return 0;
-}
-
int
getword (char **pret, const char **pstr, int delim)
{
@@ -65,52 +54,32 @@ getword (char **pret, const char **pstr, int delim)
return 0;
}
-int
-mu_rfc2047_decode (const char *tocode, const char *input, char **ptostr)
+static int
+_rfc2047_decode_param (const char *tocode, const char *input,
+ struct mu_mime_param *param)
{
int status = 0;
const char *fromstr;
- char *buffer;
- size_t bufsize;
- size_t bufpos;
size_t run_count = 0;
char *fromcode = NULL;
char *encoding_type = NULL;
char *encoded_text = NULL;
+ char *tocodetmp = NULL;
+ mu_stream_t str;
-#define BUFINC 128
-#define CHKBUF(count) do { \
- if (bufpos+count >= bufsize) \
- { \
- size_t s = bufpos + count - bufsize; \
- if (s < BUFINC) \
- s = BUFINC; \
- if (realloc_buffer (&buffer, &bufsize, s)) \
- { \
- free (buffer); \
- free (fromcode); \
- free (encoding_type); \
- free (encoded_text); \
- return ENOMEM; \
- } \
- } \
- } while (0)
-
- if (!input)
- return EINVAL;
- if (!ptostr)
- return MU_ERR_OUT_PTR_NULL;
+ memset (param, 0, sizeof (*param));
- fromstr = input;
+ status = mu_memory_stream_create (&str, MU_STREAM_RDWR);
+ if (status)
+ return status;
- /* Allocate the buffer. It is assumed that encoded string is always
- longer than it's decoded variant, so it's safe to use its length
- as the first estimate */
- bufsize = strlen (fromstr) + 1;
- buffer = malloc (bufsize);
- if (buffer == NULL)
- return ENOMEM;
- bufpos = 0;
+ if (tocode && (param->cset = strdup (tocode)) == NULL)
+ {
+ mu_stream_destroy (&str);
+ return ENOMEM;
+ }
+
+ fromstr = input;
while (*fromstr)
{
@@ -119,13 +88,39 @@ mu_rfc2047_decode (const char *tocode, const char *input,
char **ptostr)
mu_stream_t filter = NULL;
mu_stream_t in_stream = NULL;
const char *filter_type = NULL;
- size_t nbytes = 0, size;
+ size_t size;
const char *sp = fromstr + 2;
- char tmp[128];
+ char *lang;
status = getword (&fromcode, &sp, '?');
if (status)
break;
+ lang = strchr (fromcode, '*');
+ if (lang)
+ *lang++ = 0;
+ if (!param->cset)
+ {
+ param->cset = strdup (fromcode);
+ if (!param->cset)
+ {
+ status = ENOMEM;
+ break;
+ }
+ }
+ if (lang && !param->lang && (param->lang = strdup (lang)) == NULL)
+ {
+ status = ENOMEM;
+ break;
+ }
+ if (!tocode)
+ {
+ if ((tocodetmp = strdup (fromcode)) == NULL)
+ {
+ status = ENOMEM;
+ break;
+ }
+ tocode = tocodetmp;
+ }
status = getword (&encoding_type, &sp, '?');
if (status)
break;
@@ -162,22 +157,12 @@ mu_rfc2047_decode (const char *tocode, const char *input,
char **ptostr)
mu_static_memory_stream_create (&in_stream, encoded_text, size);
mu_stream_seek (in_stream, 0, MU_SEEK_SET, NULL);
- status = mu_decode_filter (&filter, in_stream, filter_type, fromcode,
- tocode);
+ status = mu_decode_filter (&filter, in_stream, filter_type,
+ fromcode, tocode);
mu_stream_unref (in_stream);
if (status != 0)
break;
-
- while ((status =
- mu_stream_read (filter, tmp, sizeof (tmp), &nbytes)) == 0
- && nbytes)
- {
- CHKBUF (nbytes);
- memcpy (buffer + bufpos, tmp, nbytes);
- bufpos += nbytes;
- }
-
- mu_stream_close (filter);
+ status = mu_stream_copy (str, filter, 0, NULL);
mu_stream_destroy (&filter);
if (status)
@@ -198,44 +183,89 @@ mu_rfc2047_decode (const char *tocode, const char *input,
char **ptostr)
{
if (--run_count)
{
- CHKBUF (run_count);
- memcpy (buffer + bufpos, fromstr - run_count, run_count);
- bufpos += run_count;
+ status = mu_stream_write (str, fromstr - run_count,
+ run_count, NULL);
+ if (status)
+ break;
run_count = 0;
}
- CHKBUF (1);
- buffer[bufpos++] = *fromstr++;
+ status = mu_stream_write (str, fromstr, 1, NULL);
+ if (status)
+ break;
+ fromstr++;
}
}
else
{
- CHKBUF (1);
- buffer[bufpos++] = *fromstr++;
+ status = mu_stream_write (str, fromstr, 1, NULL);
+ if (status)
+ break;
+ fromstr++;
}
}
- if (*fromstr)
- {
- size_t len = strlen (fromstr);
- CHKBUF (len);
- memcpy (buffer + bufpos, fromstr, len);
- bufpos += len;
- }
+ if (status == 0 && *fromstr)
+ status = mu_stream_write (str, fromstr, strlen (fromstr), NULL);
- CHKBUF (1);
- buffer[bufpos++] = 0;
-
free (fromcode);
free (encoding_type);
free (encoded_text);
+ free (tocodetmp);
+
+ if (status == 0)
+ {
+ mu_off_t size;
- if (status)
- free (buffer);
- else
- *ptostr = realloc (buffer, bufpos);
+ mu_stream_size (str, &size);
+ param->value = malloc (size + 1);
+ if (!param->value)
+ status = ENOMEM;
+ else
+ {
+ mu_stream_seek (str, 0, MU_SEEK_SET, NULL);
+ status = mu_stream_read (str, param->value, size, NULL);
+ param->value[size] = 0;
+ }
+ }
+
+ mu_stream_destroy (&str);
return status;
}
+int
+mu_rfc2047_decode_param (const char *tocode, const char *input,
+ struct mu_mime_param *param)
+{
+ int rc;
+ struct mu_mime_param tmp;
+
+ if (!input)
+ return EINVAL;
+ if (!param)
+ return MU_ERR_OUT_PTR_NULL;
+ rc = _rfc2047_decode_param (tocode, input, &tmp);
+ if (rc == 0)
+ *param = tmp;
+ return rc;
+}
+
+int
+mu_rfc2047_decode (const char *tocode, const char *input, char **ptostr)
+{
+ int rc;
+ struct mu_mime_param param;
+
+ if (!input)
+ return EINVAL;
+ if (!ptostr)
+ return MU_ERR_OUT_PTR_NULL;
+ rc = _rfc2047_decode_param (tocode, input, ¶m);
+ free (param.cset);
+ free (param.lang);
+ if (rc == 0)
+ *ptostr = param.value;
+ return rc;
+}
/**
Encode a header according to RFC 2047
diff --git a/libmailutils/filter/decode.c b/libmailutils/filter/decode.c
index 6944474..3910d7d 100644
--- a/libmailutils/filter/decode.c
+++ b/libmailutils/filter/decode.c
@@ -73,7 +73,7 @@ mu_decode_filter_args (mu_stream_t *pfilter, mu_stream_t
input,
rc = mu_filter_chain_create (pfilter, input,
MU_FILTER_DECODE, MU_STREAM_READ,
- xargc, xargv);
+ i, xargv);
free (xargv);
return rc;
}
diff --git a/libmailutils/mime/mimehdr.c b/libmailutils/mime/mimehdr.c
index 7460221..aedfac6 100644
--- a/libmailutils/mime/mimehdr.c
+++ b/libmailutils/mime/mimehdr.c
@@ -1,4 +1,5 @@
-/* GNU Mailutils -- a suite of utilities for electronic mail
+/* Operations on RFC-2231-compliant mail headers fields.
+ GNU Mailutils -- a suite of utilities for electronic mail
Copyright (C) 1999, 2000, 2001, 2004, 2005, 2007, 2009, 2010, 2011
Free Software Foundation, Inc.
@@ -15,6 +16,7 @@
You should have received a copy of the GNU Lesser General
Public License along with this library. If not,
see <http://www.gnu.org/licenses/>. */
+
#ifdef HAVE_CONFIG_H
#include <config.h>
#endif
@@ -32,386 +34,693 @@
#include <mailutils/mime.h>
#include <mailutils/filter.h>
#include <mailutils/util.h>
+#include <mailutils/wordsplit.h>
+#include <mailutils/assoc.h>
+#include <mailutils/iterator.h>
+#include <mailutils/diag.h>
+#include <mailutils/nls.h>
+
+#define MU_MIMEHDR_MULTILINE 0x01 /* Parameter was multiline */
+#define MU_MIMEHDR_CSINFO 0x02 /* Parameter contains charset/language
+ info */
+
+/* Free members of struct mu_mime_param, but do not free it itself. */
+static void
+_mu_mime_param_free (struct mu_mime_param *p)
+{
+ free (p->lang);
+ free (p->cset);
+ free (p->value);
+}
-/* See RFC 2045, 5.1. Syntax of the Content-Type Header Field */
-#define _ISSPECIAL(c) !!strchr ("()<>@,;:\\\"/[]?=", c)
+/* Treat ITEM as a pointer to struct mu_mime_param and reclaim all
+ memory associated with it.
-/* _header_get_param - an auxiliary function to extract values from
- Content-Type, Content-Disposition and similar headers.
+ This is intended for use as a destroy_item method of assoc tables and
+ lists. */
+static void
+_mu_mime_param_free_item (void *item)
+{
+ _mu_mime_param_free (item);
+ free (item);
+}
- Arguments:
+/* Recode a string between two charsets.
- FIELD_BODY Header value, complying to RFCs 2045, 2183, 2231.3;
- DISP Disposition. Unless it is NULL, the disposition part
- of FIELD_BODY is compared with it. If they differ,
- the function returns MU_ERR_NOENT.
- PARAM Name of the parameter to extract from FIELD_BODY;
- BUF Where to extract the value to;
- BUFSZ Size of BUF;
- PRET Pointer to the memory location for the return buffer (see
- below).
- PLEN Pointer to the return size.
- PFLAGS On return, flags describing the parameter are stored there.
- The MU_MIMEHDR_MULTILINE bit is set if the parameter value
- was multiline (RFC 2231.3). The MU_MIMEHDR_CSINFO bit is set
- if the parameter value includes charset/language
- information (RFC 2231.4).
-
- The function parses FIELD_BODY and extracts the value of the parameter
- PARAM.
-
- If BUF is not NULL and BUFSZ is not 0, the extracted value is stored into
- BUF. At most BUFSZ-1 bytes are copied.
-
- Otherwise, if PRET is not NULL, the function allocates enough memory to
- hold the extracted value, copies there the result, and stores the
- pointer to the allocated memory into the location pointed to by PRET.
-
- If PLEN is not NULL, the size of the extracted value (without terminating
- NUL character) is stored there.
-
- If BUF==NULL *and* PRET==NULL, no memory is allocated, but PLEN is
- honored anyway, i.e. unless it is NULL it receives size of the result.
- This can be used to estimate the needed buffer size.
-
- Return values:
- 0 on success.
- MU_ERR_NOENT, requested parameter not found, or disposition does
- not match DISP.
- MU_ERR_PARSE, if FIELD_BODY does not comply to any of the abovemntioned
- RFCs.
- ENOMEM , if unable to allocate memory.
+ Input:
+ TEXT - A string.
+ ICS - Charset of TEXT.
+ OCS - Charset to convert TEXT to.
+ Output:
+ PRESULT - On success, the pointer to the resulting string is stored here.
*/
+static int
+_recode_string (char *text, const char *ics, const char *ocs, char **presult)
+{
+ mu_stream_t istr, ostr, cvt;
+ mu_off_t size;
+ char *decoded;
+ int rc;
+
+ rc = mu_static_memory_stream_create (&istr, text, strlen (text));
+ if (rc)
+ return rc;
+ rc = mu_memory_stream_create (&ostr, 0);
+ if (rc)
+ return rc;
+ rc = mu_decode_filter (&cvt, istr, NULL, ics, ocs);
+ mu_stream_unref (istr);
+ if (rc)
+ {
+ mu_stream_unref (ostr);
+ return rc;
+ }
+ rc = mu_stream_copy (ostr, cvt, 0, &size);
+ mu_stream_unref (cvt);
+ if (rc)
+ {
+ mu_stream_unref (ostr);
+ return rc;
+ }
-/* Internal flag used by _header_get_param to delay increasing
- estimated continuation index. */
-#define _MU_MIMEHDR_INCR_CIND 0x8000
+ decoded = malloc (size + 1);
+ if (!decoded)
+ {
+ mu_stream_unref (ostr);
+ return ENOMEM;
+ }
-int
-_header_get_param (const char *field_body,
- const char *disp,
- const char *param,
- char *buf, size_t bufsz,
- char **pret, size_t *plen,
- int *pflags)
+ mu_stream_seek (ostr, 0, MU_SEEK_SET, NULL);
+ rc = mu_stream_read (ostr, decoded, size, NULL);
+ mu_stream_unref (ostr);
+ if (rc)
+ free (decoded);
+ else
+ {
+ decoded[size] = 0;
+ *presult = decoded;
+ }
+ return rc;
+}
+
+/* Structure for composing continued parameters.
+ See RFC 2231, Section 3, "Parameter Value Continuations" */
+struct param_continuation
{
- int res = MU_ERR_NOENT; /* Return value, pessimistic default */
- size_t param_len = strlen (param);
- char *p;
- size_t size;
- char *mem = NULL; /* Allocated memory storage */
- size_t retlen = 0; /* Total number of bytes copied */
- unsigned long cind = 0; /* Expected continued parameter index.
- See RFC 2231, Section 3,
- "Parameter Value Continuations" */
- int flags = 0;
-
- if (field_body == NULL)
- return EINVAL;
+ char *param_name; /* Parameter name */
+ size_t param_length; /* Length of param_name */
+ mu_stream_t param_value; /* Its value (memory stream) */
+ int param_cind; /* Expected continued parameter index. */
+ /* Language/character set information */
+ const char *param_lang;
+ const char *param_cset;
+};
- if (bufsz == 0) /* Make sure buf value is meaningful */
- buf = NULL;
+static int
+free_param_continuation (struct param_continuation *p)
+{
+ free (p->param_name);
+ mu_stream_destroy (&p->param_value);
+ /* param_lang and param_cset are handled separately */
+ memset (p, 0, sizeof (*p));
+}
+
+/* Auxiliary function to store the data collected in CONT into ASSOC.
+ If SUBSET is True, ASSOC is populated with empty mu_mime_param
+ structures. In this case data will be stored only if CONT->param_name
+ is already in ASSOC. If OUTCHARSET is not NULL, the value from
+ CONT->param_value will be recoded to that charset before storing it. */
+static int
+flush_param (struct param_continuation *cont, mu_assoc_t assoc, int subset,
+ const char *outcharset)
+{
+ int rc;
+ struct mu_mime_param param, *param_slot = NULL;
+ mu_off_t size;
- p = strchr (field_body, ';');
- if (!p)
- return MU_ERR_NOENT;
- /* Allow for possible whitespace before the semicolon */
- for (size = p - field_body;
- size > 0 && mu_isblank (field_body[size-1]); size--)
- ;
- /* Remove surrounding quotes.
- FIXME: unescape the quoted contents. */
- if (field_body[0] == '"' && field_body[size-1] == '"')
+ if (subset)
{
- field_body++;
- size -= 2;
+ param_slot = mu_assoc_ref (assoc, cont->param_name);
+ if (!param_slot)
+ return 0;
}
- if (disp && mu_c_strncasecmp (field_body, disp, size))
- return MU_ERR_NOENT;
-
- while (p && *p)
+
+ if (cont->param_lang)
{
- char *v, *e, *ep, *cp;
- size_t len, escaped_chars = 0;
-
- if (*p != ';')
+ param.lang = strdup (cont->param_lang);
+ if (!param.lang)
+ return ENOMEM;
+ }
+ else
+ param.lang = NULL;
+
+ if (outcharset || cont->param_cset)
+ {
+ param.cset = strdup (outcharset ? outcharset : cont->param_cset);
+ if (!param.cset)
{
- res = MU_ERR_PARSE;
- break;
+ free (param.lang);
+ return ENOMEM;
}
-
- /* walk upto start of param */
- p = mu_str_skip_class (p + 1, MU_CTYPE_SPACE);
+ }
+
+ rc = mu_stream_size (cont->param_value, &size);
+ if (rc == 0)
+ {
+ param.value = malloc (size + 1);
+ if (!param.value)
+ rc = ENOMEM;
+ }
- /* Reportedly, some MUAs insert several semicolons */
- if (*p == ';')
- continue;
+ if (rc == 0)
+ {
+ rc = mu_stream_seek (cont->param_value, 0, MU_SEEK_SET, NULL);
+ if (rc == 0)
+ rc = mu_stream_read (cont->param_value, param.value, size, NULL);
+ param.value[size] = 0;
+ }
+
+ if (rc)
+ {
+ free (param.lang);
+ free (param.cset);
+ return rc;
+ }
- /* Ignore stray characters */
- if (_ISSPECIAL (*p))
+ if (cont->param_cset && outcharset &&
+ mu_c_strcasecmp (cont->param_cset, outcharset))
+ {
+ char *tmp;
+ rc = _recode_string (param.value, cont->param_cset, outcharset, &tmp);
+ free (param.value);
+ if (rc)
{
- p = strchr (p, ';');
- continue;
+ free (param.lang);
+ free (param.cset);
+ return rc;
}
+ param.value = tmp;
+ }
- if ((ep = strchr (p, '=')) == NULL)
- break;
- /* Allow for optional whitespace after '=' */
- v = mu_str_skip_class (ep + 1, MU_CTYPE_SPACE);
- /* Find end of the parameter */
- if (*v == '"')
+ if (param_slot)
+ {
+ *param_slot = param;
+ }
+ else
+ {
+ rc = mu_assoc_install (assoc, cont->param_name, ¶m);
+ if (rc)
+ _mu_mime_param_free (¶m);
+ }
+
+ return rc;
+}
+
+/* Create and initialize an empty associative array for parameters. */
+int
+mu_mime_param_assoc_create (mu_assoc_t *paramtab)
+{
+ mu_assoc_t assoc;
+ int rc = mu_assoc_create (&assoc, sizeof (struct mu_mime_param),
+ MU_ASSOC_ICASE);
+ if (rc == 0)
+ mu_assoc_set_free (assoc, _mu_mime_param_free_item);
+ *paramtab = assoc;
+ return rc;
+}
+
+/* Add an empty structure for the slot NAME in ASSOC. */
+int
+mu_mime_param_assoc_add (mu_assoc_t assoc, const char *name)
+{
+ struct mu_mime_param param;
+
+ memset (¶m, 0, sizeof param);
+ return mu_assoc_install (assoc, name, ¶m);
+}
+
+/* See FIXME near the end of _mime_header_parse, below. */
+static int
+_remove_entry (void *item, void *data)
+{
+ struct mu_mime_param *p = item;
+ mu_assoc_t assoc = data;
+ mu_assoc_remove_ref (assoc, p);
+ return 0;
+}
+
+/* A working horse of this module. Parses input string, which should
+ be a header field value complying to RFCs 2045, 2183, 2231.3.
+
+ Input:
+ TEXT - The string.
+ ASSOC - Associative array of parameters indexed by their names.
+ SUBSET - If true, store only those parameters that are already
+ in ASSOC.
+ Output:
+ PVALUE - Unless NULL, a pointer to the field value is stored here on
+ success.
+ ASSOC - Unless NULL, parameters are stored here.
+
+ Either PVALUE or ASSOC (but not both) can be NULL, meaning that the
+ corresponding data are of no interest to the caller.
+*/
+static int
+_mime_header_parse (const char *text, char **pvalue,
+ mu_assoc_t assoc, const char *outcharset, int subset)
+{
+ int rc;
+ struct mu_wordsplit ws;
+ struct param_continuation cont;
+ size_t i;
+
+ ws.ws_delim = " \t\r\n;";
+ ws.ws_escape = "\\\"";
+ if (mu_wordsplit (text, &ws,
+ MU_WRDSF_DELIM | MU_WRDSF_ESCAPE |
+ MU_WRDSF_NOVAR | MU_WRDSF_NOCMD |
+ MU_WRDSF_DQUOTE | MU_WRDSF_SQUEEZE_DELIMS |
+ MU_WRDSF_RETURN_DELIMS | MU_WRDSF_WS))
+ {
+ mu_debug (MU_DEBCAT_MIME, MU_DEBUG_ERROR,
+ (_("wordsplit: %s"), mu_wordsplit_strerror (&ws)));
+ return MU_ERR_PARSE;
+ }
+
+ if (!assoc)
+ {
+ if (!pvalue)
+ return MU_ERR_OUT_PTR_NULL;
+ *pvalue = strdup (ws.ws_wordv[i]);
+ mu_wordsplit_free (&ws);
+ if (!*pvalue)
+ return ENOMEM;
+ return 0;
+ }
+
+ memset (&cont, 0, sizeof (cont));
+ for (i = 1; i < ws.ws_wordc; i++)
+ {
+ size_t klen;
+ char *key;
+ char *val;
+ const char *lang = NULL;
+ const char *cset = NULL;
+ char *langp = NULL;
+ char *csetp = NULL;
+ char *p;
+ char *decoded;
+ int flags = 0;
+ struct mu_mime_param param;
+
+ key = ws.ws_wordv[i];
+ if (key[0] == ';')
+ /* Reportedly, some MUAs insert several semicolons */
+ continue;
+ p = strchr (key, '=');
+ if (!p)
+ val = "";
+ else
{
- /* Quoted string */
- for (e = ++v; *e != '"'; e++)
+ *p++ = 0;
+ val = p;
+ }
+
+ klen = strlen (key);
+ if (klen == 0)
+ continue;
+
+ p = strchr (key, '*');
+ if (p)
+ {
+ /* It is a parameter value continuation (RFC 2231, Section 3)
+ or parameter value character set and language information
+ (ibid., Section 4). */
+ klen = p - key;
+ if (p[1])
{
- if (*e == 0) /* Malformed header */
+ if (mu_isdigit (p[1]))
{
- res = MU_ERR_PARSE;
- break;
- }
- if (*e == '\\')
- {
- if (*++e == 0)
+ char *q;
+ unsigned long n = strtoul (p + 1, &q, 10);
+
+ if (*q && *q != '*')
{
- res = MU_ERR_PARSE;
- break;
+ mu_debug (MU_DEBCAT_MIME, MU_DEBUG_TRACE0,
+ (_("malformed parameter name %s: skipping"),
+ key));
+ continue;
}
- escaped_chars++;
+
+ if (n != cont.param_cind)
+ {
+ mu_debug (MU_DEBCAT_MIME, MU_DEBUG_TRACE0,
+ (_("continuation index out of sequence in %s: "
+ "skipping"),
+ key));
+ /* Ignore this parameter. Another possibility would be
+ to drop the continuation assembled so far. That makes
+ little difference, because the string is malformed
+ anyway.
+
+ We try to continue just to gather as many information
+ as possible from this mess.
+ */
+ continue;
+ }
+
+ if (n == 0)
+ {
+ cont.param_name = malloc (klen + 1);
+ if (!cont.param_name)
+ {
+ rc = ENOMEM;
+ break;
+ }
+ cont.param_length = klen;
+ memcpy (cont.param_name, key, klen);
+ cont.param_name[klen] = 0;
+
+ rc = mu_memory_stream_create (&cont.param_value,
+ MU_STREAM_RDWR);
+ if (rc)
+ break;
+ }
+ else if (cont.param_length != klen ||
+ memcmp (cont.param_name, key, klen))
+ {
+ mu_debug (MU_DEBCAT_MIME, MU_DEBUG_TRACE0,
+ (_("continuation name mismatch: %s: "
+ "skipping"),
+ key));
+ continue;
+ }
+
+ if (*q == '*')
+ flags |= MU_MIMEHDR_CSINFO;
+
+ cont.param_cind++;
+ flags |= MU_MIMEHDR_MULTILINE;
}
}
- if (res == MU_ERR_PARSE)
- break;
- len = e - v;
- e++;
+ else
+ {
+ flags |= MU_MIMEHDR_CSINFO;
+ *p = 0;
+ }
}
- else
+ else if (cont.param_name)
{
- for (e = v + 1; *e && !(*e == ';' || mu_isspace (*e)); e++)
- ;
- len = e - v;
+ rc = flush_param (&cont, assoc, subset, outcharset);
+ free_param_continuation (&cont);
+ if (rc)
+ break;
}
+
+ if (flags & MU_MIMEHDR_CSINFO)
+ {
+ p = strchr (val, '\'');
+ if (p)
+ {
+ char *q = strchr (p + 1, '\'');
+ if (q)
+ {
+ cset = val;
+ *p++ = 0;
+ lang = p;
+ *q++ = 0;
+ val = q;
+ }
+ }
- /* Is it our parameter? */
- if (mu_c_strncasecmp (p, param, param_len))
- { /* nope, jump to next */
- p = strchr (e, ';');
- continue;
+ if ((flags & MU_MIMEHDR_MULTILINE) && cont.param_cind == 1)
+ {
+ cont.param_lang = lang;
+ cont.param_cset = cset;
+ }
}
- cp = p + param_len;
-
- if (*cp == '*')
+ if (flags & MU_MIMEHDR_CSINFO)
{
- cp++;
- /* It is a parameter value continuation (RFC 2231, Section 3)
- or parameter value character set and language information
- (ibid., Section 4). */
- if (mu_isdigit (*cp))
- {
- /* See if the index is OK */
+ char *tmp;
- char *end;
- unsigned long n = strtoul (cp, &end, 10);
-
- if (*end == '*')
- {
- flags |= MU_MIMEHDR_CSINFO;
- end++;
- }
- if (n != cind)
+ rc = mu_str_url_decode (&tmp, val);
+ if (rc)
+ break;
+ if (!(flags & MU_MIMEHDR_MULTILINE))
+ {
+ if (!outcharset || mu_c_strcasecmp (cset, outcharset) == 0)
+ decoded = tmp;
+ else
{
- res = MU_ERR_PARSE;
- break;
+ rc = _recode_string (tmp, cset, outcharset, &decoded);
+ free (tmp);
}
- /* Everything OK, mark this as a multiline (continued)
- parameter. We also need to increment the estimation,
- but it cannot be done right now because its value is
- used below to decide whether to do flag cleanup
- on error. So we set _MU_MIMEHDR_INCR_CIND flag instead
- and increment cind later. */
- flags |= (MU_MIMEHDR_MULTILINE|_MU_MIMEHDR_INCR_CIND);
- /* And point cp to the last character: there are more
- checks ahead. */
- cp = end;
+ if (rc)
+ break;
}
else
- flags |= MU_MIMEHDR_CSINFO;
+ decoded = tmp;
}
- /* Allow for optional whitespace before '=' */
- cp = mu_str_skip_class (cp, MU_CTYPE_SPACE);
- /* cp must now point to the equals sign */
- if (cp != ep)
+ else
{
- /* Clean up everything, unless we're in the middle of a
- parameter continuation. */
- if (cind == 0)
- flags = 0;
-
- /* Try next parameter */
- p = strchr (e, ';');
- continue;
+ rc = mu_rfc2047_decode_param (outcharset, val, ¶m);
+ if (rc)
+ return rc;
+ cset = csetp = param.cset;
+ lang = langp = param.lang;
+ decoded = param.value;
}
-
- if (flags & _MU_MIMEHDR_INCR_CIND)
+ val = decoded;
+
+ if (flags & MU_MIMEHDR_MULTILINE)
{
- /* Increase the estimation. */
- flags &= ~_MU_MIMEHDR_INCR_CIND;
- cind++;
+ rc = mu_stream_write (cont.param_value, val, strlen (val),
+ NULL);
+ free (decoded);
+ free (csetp);
+ free (langp);
+ if (rc)
+ break;
+ continue;
}
-
- res = 0; /* Indicate success */
-
- /* Prepare P for the next iteration */
- p = e;
- /* Escape characters that appear in quoted-pairs are
- semantically "invisible" (RFC 2822, Section 3.2.2,
- "Quoted characters") */
- len -= escaped_chars;
-
- /* Adjust len if nearing end of the buffer */
- if (bufsz && len >= bufsz)
- len = bufsz - 1;
-
- if (pret)
+ memset (¶m, 0, sizeof (param));
+ if (lang)
{
- /* The caller wants us to allocate the memory */
- if (!buf && !mem)
- {
- mem = malloc (len + 1);
- if (!mem)
- {
- res = ENOMEM;
- break;
- }
- buf = mem;
- }
- else if (mem)
+ param.lang = strdup (lang);
+ if (!param.lang)
+ rc = ENOMEM;
+ else
{
- /* If we got here, it means we are iterating over
- a parameter value continuation, and cind=0 has
- already been passed. Reallocate the memory to
- accomodate next chunk of data. */
- char *newmem = realloc (mem, retlen + len + 1);
- if (!newmem)
+ param.cset = strdup (cset);
+ if (!param.cset)
{
- res = ENOMEM;
- break;
+ free (param.lang);
+ rc = ENOMEM;
}
- buf = mem = newmem;
}
}
- if (buf)
+ free (csetp);
+ free (langp);
+
+ if (rc)
+ {
+ free (decoded);
+ break;
+ }
+
+ param.value = strdup (val);
+ free (decoded);
+ if (!param.value)
{
- /* Actually copy the data. Buf is not NULL either because
- the user passed it as an argument, or because we allocated
- memory for it. */
- if (escaped_chars)
+ _mu_mime_param_free (¶m);
+ rc = ENOMEM;
+ break;
+ }
+
+ if (subset)
+ {
+ struct mu_mime_param *p = mu_assoc_ref (assoc, key);
+ if (p)
+ *p = param;
+ }
+ else
+ {
+ rc = mu_assoc_install (assoc, key, ¶m);
+ if (rc)
{
- int i;
- for (i = 0; i < len; i++)
- {
- if (*v == '\\')
- ++v;
- buf[retlen + i] = *v++;
- }
+ _mu_mime_param_free (¶m);
+ break;
}
- else
- memcpy (buf + retlen, v, len);
}
- /* Adjust total result size ... */
- retlen += len;
- /* ... and remaining buffer size, if necessary */
- if (bufsz)
+ }
+
+ if (rc == 0 && cont.param_name)
+ rc = flush_param (&cont, assoc, subset, outcharset);
+ free_param_continuation (&cont);
+ if (rc == 0)
+ {
+ if (pvalue)
{
- bufsz -= len;
- if (bufsz == 0)
- break;
+ *pvalue = strdup (ws.ws_wordv[0]);
+ if (!*pvalue)
+ rc = ENOMEM;
}
}
- if (res == 0)
+ mu_wordsplit_free (&ws);
+
+ if (subset)
{
- /* Everything OK, prepare the returned data. */
- if (buf)
- buf[retlen] = 0;
- if (plen)
- *plen = retlen;
- if (pret)
- *pret = mem;
- if (pflags)
- *pflags = flags;
+ /* Eliminate empty elements.
+
+ FIXME: What I wanted to do initially is commented out, because
+ unfortunately iterator_ctl is not defined for assoc tables...
+ */
+#if 0
+ mu_iterator_t itr;
+
+ rc = mu_assoc_get_iterator (assoc, &itr);
+ if (rc == 0)
+ {
+ for (mu_iterator_first (itr); !mu_iterator_is_done (itr);
+ mu_iterator_next (itr))
+ {
+ const char *name;
+ struct mu_mime_param *p;
+
+ mu_iterator_current_kv (itr, (const void **)&name, (void**)&p);
+ if (!p->value)
+ mu_iterator_ctl (itr, mu_itrctl_delete, NULL);
+ }
+ mu_iterator_destroy (&itr);
+ }
+#else
+ /* ... Instead, the following kludgy approach is taken: */
+ mu_iterator_t itr;
+ mu_list_t elist;
+
+ rc = mu_list_create (&elist);
+ if (rc)
+ return rc;
+ rc = mu_assoc_get_iterator (assoc, &itr);
+ if (rc == 0)
+ {
+ for (mu_iterator_first (itr); rc == 0 && !mu_iterator_is_done (itr);
+ mu_iterator_next (itr))
+ {
+ const char *name;
+ struct mu_mime_param *p;
+
+ mu_iterator_current_kv (itr, (const void **)&name, (void**)&p);
+ if (!p->value)
+ rc = mu_list_append (elist, p);
+ }
+ mu_iterator_destroy (&itr);
+ }
+ if (rc == 0)
+ mu_list_foreach (elist, _remove_entry, assoc);
+ mu_list_destroy (&elist);
+#endif
}
- else if (mem)
- free (mem);
- return res;
+
+ return rc;
}
-static size_t
-disp_segment_len (const char *str)
+/* Parse header value from TEXT and return its value and a subset of
+ parameters.
+
+ Input:
+ TEXT - Header value.
+ CSET - Output charset. Can be NULL, in which case no conversions
+ take place.
+ ASSOC - Parameter array initialized with empty slots for those
+ parameters, which are wanted on output. It should be
+ created using mu_mime_param_assoc_create and populated
+ using mu_mime_param_assoc_add.
+ Output:
+ PVALUE - A pointer to the field value is stored here on success.
+ ASSOC - Receives available parameters matching the input subset.
+
+ Either PVALUE or ASSOC (but not both) can be NULL, meaning that the
+ corresponding data are of no interest to the caller.
+*/
+int
+mu_mime_header_parse_subset (const char *text, const char *cset,
+ char **pvalue, mu_assoc_t assoc)
{
- char *p = strchr (str, ';');
- size_t size;
-
- if (!p)
- size = strlen (str);
- else
- size = p - str;
- while (size > 0 && mu_isblank (str[size-1]))
- size--;
- return size;
+ return _mime_header_parse (text, pvalue, assoc, cset, 1);
}
-/* STR is a value of a structured MIME header, e.g. Content-Type.
+/* Parse header value from TEXT and return its value and parameters.
+
+ Input:
+ TEXT - Header value.
+ CSET - Output charset. Can be NULL, in which case no conversions
+ take place.
+ Output:
+ PVALUE - A pointer to the field value is stored here on success.
+ PASSOC - Receives an associative array of parameters.
+
+ Either PVALUE or PASSOC (but not both) can be NULL, meaning that the
+ corresponding data are of no interest to the caller.
+*/
+int
+mu_mime_header_parse (const char *text, char *cset, char **pvalue,
+ mu_assoc_t *passoc)
+{
+ int rc;
+ mu_assoc_t assoc;
+
+ rc = mu_mime_param_assoc_create (&assoc);
+ if (rc == 0)
+ {
+ rc = _mime_header_parse (text, pvalue, assoc, cset, 0);
+ if (rc || !passoc)
+ mu_assoc_destroy (&assoc);
+ else
+ *passoc = assoc;
+ }
+
+ return rc;
+}
+
+/* TEXT is a value of a structured MIME header, e.g. Content-Type.
This function returns the `disposition part' of it. In other
- words, it returns disposition, if STR is a Content-Disposition
+ words, it returns disposition, if TEXT is a Content-Disposition
value, and `type/subtype' part, if it is a Content-Type value.
*/
int
-mu_mimehdr_get_disp (const char *str, char *buf, size_t bufsz, size_t *retsz)
+mu_mimehdr_get_disp (const char *text, char *buf, size_t bufsz, size_t *retsz)
{
- size_t size;
-
- str = mu_str_skip_class (str, MU_CTYPE_BLANK);
- size = disp_segment_len (str);
- if (size > 2 && str[0] == '"' && str[size-1] == '"')
+ int rc;
+ char *value;
+
+ rc = mu_mime_header_parse (text, NULL, &value, NULL);
+ if (rc == 0)
{
- str++;
- size -= 2;
+ size_t size = strlen (value);
+ if (size > bufsz)
+ size = bufsz;
+ if (buf)
+ size = mu_cpystr (buf, value, size);
+ if (retsz)
+ *retsz = size;
}
- if (buf)
- size = mu_cpystr (buf, str, size);
- if (retsz)
- *retsz = size;
+ free (value);
return 0;
}
/* Same as mu_mimehdr_get_disp, but allocates memory */
int
-mu_mimehdr_aget_disp (const char *str, char **pvalue)
+mu_mimehdr_aget_disp (const char *text, char **pvalue)
{
- char *p;
- size_t size;
-
- str = mu_str_skip_class (str, MU_CTYPE_BLANK);
- size = disp_segment_len (str);
- if (size > 2 && str[0] == '"' && str[size-1] == '"')
- {
- str++;
- size -= 2;
- }
-
- p = malloc (size + 1);
- if (!p)
- return ENOMEM;
- memcpy (p, str, size);
- p[size] = 0;
- *pvalue = p;
- return 0;
+ return mu_mime_header_parse (text, NULL, pvalue, NULL);
}
-/* Get the value of the parameter PARAM from STR, which must be
+/* Get the value of parameter NAME from STR, which must be
a value of a structured MIME header.
At most BUFSZ-1 of data are stored in BUF. A terminating NUL
character is appended to it.
@@ -419,202 +728,100 @@ mu_mimehdr_aget_disp (const char *str, char **pvalue)
Unless NULL, RETSZ is filled with the actual length of the
returned data (not including the NUL terminator).
- Unless PFLAGS is null it will contain, on return, the flags describing
- the parameter. The MU_MIMEHDR_MULTILINE bit is set if the parameter value
- was multiline (RFC 2231.3). The MU_MIMEHDR_CSINFO bit is set if the
- parameter value includes charset/language information (RFC 2231.4).
-
BUF may be NULL, in which case the function will only fill
- RETSZ and PFLAGS, as described above. */
+ RETSZ, as described above. */
int
-mu_mimehdr_get_param (const char *str, const char *param,
- char *buf, size_t bufsz, size_t *retsz,
- int *pflags)
+mu_mimehdr_get_param (const char *str, const char *name,
+ char *buf, size_t bufsz, size_t *retsz)
{
- return _header_get_param (str, NULL, param, buf, bufsz, NULL, retsz,
- pflags);
+ int rc;
+ char *value;
+
+ rc = mu_mimehdr_aget_param (str, name, &value);
+ if (rc == 0)
+ {
+ size_t size = strlen (value);
+ if (size > bufsz)
+ size = bufsz;
+ if (buf)
+ size = mu_cpystr (buf, value, size);
+ if (retsz)
+ *retsz = size;
+ }
+ free (value);
+ return rc;
}
/* Same as mu_mimehdr_get_param, but allocates memory. */
int
-mu_mimehdr_aget_param (const char *str, const char *param,
- char **pval, int *pflags)
+mu_mimehdr_aget_param (const char *str, const char *name, char **pval)
{
- return _header_get_param (str, NULL, param, NULL, 0, pval, NULL, pflags);
+ return mu_mimehdr_aget_decoded_param (str, name, NULL, pval, NULL);
}
-/* Decode a parameter value. Arguments:
-
- Input:
- VALUE Parameter value.
- FLAGS Flags obtained from a previous call to one of the functions
- above.
- CHARSET Output charset.
-
- Output:
- PVAL A pointer to the decoded value is stored there.
- The memory is allocated using malloc.
- PLANG If language information was present in VALUE, its
- malloc'ed copy is stored in the memory location pointed
- to by this variable. If there was no language information,
- *PLANG is set to NULL.
-
- Both PVAL and PLANG may be NULL if that particular piece of information
- is not needed. */
+
+/* Similar to mu_mimehdr_aget_param, but the returned value is decoded
+ according to the CHARSET. Unless PLANG is NULL, it receives malloc'ed
+ language name from STR. If there was no language name, *PLANG is set
+ to NULL.
+*/
int
-mu_mimehdr_decode_param (const char *value, int flags,
- const char *charset, char **pval, char **plang)
+mu_mimehdr_aget_decoded_param (const char *str, const char *name,
+ const char *charset,
+ char **pval, char **plang)
{
- char *decoded;
+ mu_assoc_t assoc;
int rc;
- char *lang = NULL;
- char *data;
- if (flags == 0)
- {
- rc = mu_rfc2047_decode (charset, value, &decoded);
- if (rc)
- return rc;
- }
- else
+ rc = mu_mime_param_assoc_create (&assoc);
+ if (rc == 0)
{
- rc = mu_str_url_decode (&decoded, value);
- if (rc)
- return rc;
-
- if ((flags & MU_MIMEHDR_CSINFO)
- && (lang = strchr (decoded, '\''))
- && (data = strchr (lang + 1, '\'')))
+ rc = mu_mime_param_assoc_add (assoc, name);
+ if (rc == 0)
{
- char *source_cs = decoded;
-
- *lang++ = 0;
- *data++ = 0;
-
- lang = lang[0] ? strdup (lang) : NULL;
-
- if (source_cs[0] && charset && mu_c_strcasecmp (source_cs, charset))
+ rc = mu_mime_header_parse_subset (str, charset, NULL, assoc);
+ if (rc == 0)
{
- char *outval = NULL;
- mu_stream_t instr = NULL;
- mu_stream_t outstr = NULL;
- mu_stream_t cvt = NULL;
- char iobuf[512];
-
- do
+ struct mu_mime_param *param = mu_assoc_ref (assoc, name);
+ if (!param)
+ rc = MU_ERR_NOENT;
+ else
{
- size_t total = 0, pos;
- size_t nbytes;
-
- rc = mu_static_memory_stream_create (&instr, data,
- strlen (data));
- if (rc)
- break;
-
- rc = mu_memory_stream_create (&outstr, 0);
- if (rc)
- break;
-
- rc = mu_decode_filter (&cvt, instr, NULL,
- source_cs, charset);
- if (rc)
- break;
-
- while (mu_stream_read (cvt, iobuf, sizeof (iobuf),
- &nbytes) == 0
- && nbytes)
- {
- rc = mu_stream_write (outstr, iobuf, nbytes, NULL);
- if (rc)
- break;
- total += nbytes;
- }
-
- if (rc)
- break;
-
- outval = malloc (total + 1);
- if (!outval)
+ *pval = param->value;
+ if (plang)
{
- rc = ENOMEM;
- break;
+ *plang = param->lang;
+ param->lang = NULL;
}
-
- mu_stream_seek (outstr, 0, MU_SEEK_SET, NULL);
- pos = 0;
- while (mu_stream_read (outstr, outval + pos,
- total - pos, &nbytes) == 0
- && nbytes)
- pos += nbytes;
- outval[pos] = 0;
+ param->value = NULL;
}
- while (0);
-
- mu_stream_close (cvt);
- mu_stream_destroy (&cvt);
- mu_stream_close (instr);
- mu_stream_destroy (&instr);
- mu_stream_close (outstr);
- mu_stream_destroy (&outstr);
-
- free (decoded);
-
- if (rc)
- {
- /* Cleanup after an error. */
- free (lang);
- free (outval);
- return rc;
- }
- decoded = outval;
}
- else
- memmove (decoded, data, strlen (data) + 1);
}
- }
-
- if (pval)
- *pval = decoded;
- else
- free (decoded);
-
- if (plang)
- *plang = lang;
- return 0;
-}
-
-/* Similar to mu_mimehdr_aget_param, but the returned value is decoded
- according to the CHARSET. Unless PLANG is NULL, it receives malloc'ed
- language name from STR. If there was no language name, *PLANG is set
- to NULL.
-*/
-int
-mu_mimehdr_aget_decoded_param (const char *str, const char *param,
- const char *charset,
- char **pval, char **plang)
-{
- char *value;
- int rc;
- int flags;
-
- rc = mu_mimehdr_aget_param (str, param, &value, &flags);
- if (rc == 0)
- {
- rc = mu_mimehdr_decode_param (value, flags, charset, pval, plang);
- free (value);
+ mu_assoc_destroy (&assoc);
}
return rc;
}
-/* Get the attachment name from MSG. See _header_get_param, for a
- description of the rest of arguments. */
+/* Get the attachment name from a message.
+
+ Input:
+ MSG - The input message.
+ CHARSET - Character set to recode output values to. Can be NULL.
+ Output:
+ PBUF - Output value.
+ PSZ - Its size in bytes, not counting the terminating zero.
+ PLANG - Language the name is written in, if provided in the header.
+
+ Either PSZ or PLAN (or both) can be NULL.
+*/
static int
-_get_attachment_name (mu_message_t msg, char *buf, size_t bufsz,
- char **pbuf, size_t *sz, int *pflags)
+_get_attachment_name (mu_message_t msg, const char *charset,
+ char **pbuf, size_t *psz, char **plang)
{
int ret = EINVAL;
mu_header_t hdr;
char *value = NULL;
+ mu_assoc_t assoc;
if (!msg)
return ret;
@@ -622,7 +829,8 @@ _get_attachment_name (mu_message_t msg, char *buf, size_t
bufsz,
if ((ret = mu_message_get_header (msg, &hdr)) != 0)
return ret;
- ret = mu_header_aget_value_unfold (hdr, "Content-Disposition", &value);
+ ret = mu_header_aget_value_unfold (hdr, MU_HEADER_CONTENT_DISPOSITION,
+ &value);
/* If the header wasn't there, we'll fall back to Content-Type, but
other errors are fatal. */
@@ -631,33 +839,88 @@ _get_attachment_name (mu_message_t msg, char *buf, size_t
bufsz,
if (ret == 0 && value != NULL)
{
- ret = _header_get_param (value, "attachment",
- "filename", buf, bufsz, pbuf, sz, pflags);
- free (value);
- value = NULL;
- if (ret == 0 || ret != MU_ERR_NOENT)
+ ret = mu_mime_param_assoc_create (&assoc);
+ if (ret)
return ret;
+ ret = mu_mime_param_assoc_add (assoc, "filename");
+ if (ret == 0)
+ {
+ char *disp;
+
+ ret = mu_mime_header_parse_subset (value, charset, &disp, assoc);
+ if (ret == 0)
+ {
+ struct mu_mime_param *param;
+ if (mu_c_strcasecmp (disp, "attachment") == 0 &&
+ (param = mu_assoc_ref (assoc, "filename")))
+ {
+ *pbuf = param->value;
+ if (psz)
+ *psz = strlen (*pbuf);
+ param->value = NULL;
+ if (plang)
+ {
+ *plang = param->lang;
+ param->lang = NULL;
+ }
+ }
+ else
+ ret = MU_ERR_NOENT;
+ free (disp);
+ mu_assoc_destroy (&assoc);
+ }
+ }
}
+ free (value);
+
+ if (ret == 0)
+ return ret;
+
/* If we didn't get the name, we fall back on the Content-Type name
parameter. */
- free (value);
- ret = mu_header_aget_value_unfold (hdr, "Content-Type", &value);
+ ret = mu_header_aget_value_unfold (hdr, MU_HEADER_CONTENT_TYPE, &value);
if (ret == 0)
- ret = _header_get_param (value, NULL, "name", buf, bufsz, pbuf, sz,
- pflags);
- free (value);
+ {
+ ret = mu_mime_param_assoc_create (&assoc);
+ if (ret)
+ return ret;
+ ret = mu_mime_param_assoc_add (assoc, "name");
+ if (ret == 0)
+ {
+ ret = mu_mime_header_parse_subset (value, charset, NULL, assoc);
+ if (ret == 0)
+ {
+ struct mu_mime_param *param;
+ if ((param = mu_assoc_ref (assoc, "name")))
+ {
+ *pbuf = param->value;
+ if (psz)
+ *psz = strlen (*pbuf);
+ param->value = NULL;
+ if (plang)
+ {
+ *plang = param->lang;
+ param->lang = NULL;
+ }
+ }
+ else
+ ret = MU_ERR_NOENT;
+ }
+ }
+ free (value);
+ }
return ret;
}
int
-mu_message_aget_attachment_name (mu_message_t msg, char **name, int *pflags)
+mu_message_aget_attachment_name (mu_message_t msg, char **name)
{
if (name == NULL)
return MU_ERR_OUT_PTR_NULL;
- return _get_attachment_name (msg, NULL, 0, name, NULL, pflags);
+ return _get_attachment_name (msg, NULL, name, NULL, NULL);
}
int
@@ -666,21 +929,28 @@ mu_message_aget_decoded_attachment_name (mu_message_t msg,
char **pval,
char **plang)
{
- char *value;
- int flags;
- int rc = mu_message_aget_attachment_name (msg, &value, &flags);
- if (rc == 0)
- {
- rc = mu_mimehdr_decode_param (value, flags, charset, pval, plang);
- free (value);
- }
- return rc;
+ if (pval == NULL)
+ return MU_ERR_OUT_PTR_NULL;
+ return _get_attachment_name (msg, charset, pval, NULL, plang);
}
int
mu_message_get_attachment_name (mu_message_t msg, char *buf, size_t bufsz,
- size_t *sz, int *pflags)
+ size_t *sz)
{
- return _get_attachment_name (msg, buf, bufsz, NULL, sz, pflags);
+ char *tmp;
+ size_t size;
+ int rc = _get_attachment_name (msg, NULL, &tmp, &size, NULL);
+ if (rc == 0)
+ {
+ if (size > bufsz)
+ size = bufsz;
+ if (buf)
+ size = mu_cpystr (buf, tmp, size);
+ if (sz)
+ *sz = size;
+ }
+ free (tmp);
+ return rc;
}
diff --git a/libmailutils/tests/.gitignore b/libmailutils/tests/.gitignore
index 6a1188d..955a4e3 100644
--- a/libmailutils/tests/.gitignore
+++ b/libmailutils/tests/.gitignore
@@ -16,6 +16,7 @@ fsfolder
imapio
listop
mailcap
+mimehdr
prop
scantime
strftime
diff --git a/libmailutils/tests/Makefile.am b/libmailutils/tests/Makefile.am
index aca151a..8b97c86 100644
--- a/libmailutils/tests/Makefile.am
+++ b/libmailutils/tests/Makefile.am
@@ -51,6 +51,7 @@ noinst_PROGRAMS = \
imapio\
listop\
mailcap\
+ mimehdr\
prop\
scantime\
strftime\
@@ -88,6 +89,7 @@ TESTSUITE_AT = \
linecon.at\
list.at\
mailcap.at\
+ mimehdr.at\
prop.at\
scantime.at\
strftime.at\
diff --git a/libmailutils/tests/mimehdr.at b/libmailutils/tests/mimehdr.at
new file mode 100644
index 0000000..dbe8f9d
--- /dev/null
+++ b/libmailutils/tests/mimehdr.at
@@ -0,0 +1,131 @@
+# This file is part of GNU Mailutils. -*- Autotest -*-
+# Copyright (C) 2011 Free Software Foundation, Inc.
+#
+# GNU Mailutils is free software; you can redistribute it and/or
+# modify it under the terms of the GNU General Public License as
+# published by the Free Software Foundation; either version 3, or (at
+# your option) any later version.
+#
+# GNU Mailutils is distributed in the hope that it will be useful, but
+# WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+# General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with GNU Mailutils. If not, see <http://www.gnu.org/licenses/>.
+
+# Warning: This text contains 8-bit UTF-8
+
+AT_BANNER(RFC 2231 header fields)
+
+dnl ---------------------------------------------------------------------
+dnl MIMEHDR([NAME], [KW], [OPT], [INPUT], [STDOUT = `'], [STDERR = `'])
+dnl
+m4_pushdef([MIMEHDR],[
+m4_pushdef([MU_TEST_GROUP],[mimehdr])
+m4_pushdef([MU_TEST_KEYWORDS],[mimehdr rfc2231])
+m4_pushdef([MU_TEST_COMMAND],[mimehdr $3])
+MU_GENERIC_TEST([$1],[$2],[$4],[],[$5],[$6])
+m4_popdef([MU_TEST_COMMAND])
+m4_popdef([MU_TEST_KEYWORDS])
+m4_popdef([MU_TEST_GROUP])
+])
+dnl ---------------------------------------------------------------------
+
+MIMEHDR([simple],[mimehdr00 mimehdr-simple],
+[],
+[message/external-body; access-type=URL;
+ URL="ftp://cs.utk.edu/pub/moore/bulk-mailer/bulk-mailer.tar";
+],
+[message/external-body
+access-type=URL
+URL=ftp://cs.utk.edu/pub/moore/bulk-mailer/bulk-mailer.tar
+])
+
+MIMEHDR([continuation],[mimehdr01 mimehdr-cont mimehdr-cont-00],
+[],
+[message/external-body; access-type=URL;
+ URL*0="ftp://";;
+ URL*1="cs.utk.edu/pub/moore/bulk-mailer/bulk-mailer.tar"
+],
+[message/external-body
+access-type=URL
+URL=ftp://cs.utk.edu/pub/moore/bulk-mailer/bulk-mailer.tar
+])
+
+MIMEHDR([charset (2047)],[mimehdr02 mimehdr-charset-rfc2047
mimehdr-charset-00],
+[],
+[attachment; charset=utf-8;
+
filename==?UTF-8?B?zrHPgc+HzrXOr86/IM6zzrnOsSDPhM63zr0gzrTOv866zrnOvM6xz4POr86x==?=
+],
+[attachment
+charset=utf-8
+filename=αÏÏείο για Ïην δοκιμαÏία
+])
+
+MIMEHDR([charset with language (2047)],[mimehdr03 mimehdr-charset-rfc2047
mimehdr-charset-01],
+[],
+[attachment; charset=utf-8;
+
filename==?UTF-8*el?B?zrHPgc+HzrXOr86/IM6zzrnOsSDPhM63zr0gzrTOv866zrnOvM6xz4POr86x==?=
+],
+[attachment
+charset=utf-8
+filename(lang:el/UTF-8)=αÏÏείο για Ïην δοκιμαÏία
+])
+
+MIMEHDR([no charset (2231)],[mimehdr04 mimehdr-no-charset-rfc2231
mimehdr-nocharset-00],
+[],
+[attachment; charset=utf-8;
+
filename*=%CE%B1%CF%81%CF%87%CE%B5%CE%AF%CE%BF%20%CE%B3%CE%B9%CE%B1%20%CF%84%CE%B7%CE%BD%20%CE%B4%CE%BF%CE%BA%CE%B9%CE%BC%CE%B1%CF%83%CE%AF%CE%B1
+],
+[attachment
+charset=utf-8
+filename=αÏÏείο για Ïην δοκιμαÏία
+])
+
+MIMEHDR([charset (2231)],[mimehdr05 mimehdr-charset-rfc2231
mimehdr-charset-rfc2231-00 mimehdr-charset-03],
+[],
+[attachment; charset=utf-8;
+
filename*=UTF-8''%CE%B1%CF%81%CF%87%CE%B5%CE%AF%CE%BF%20%CE%B3%CE%B9%CE%B1%20%CF%84%CE%B7%CE%BD%20%CE%B4%CE%BF%CE%BA%CE%B9%CE%BC%CE%B1%CF%83%CE%AF%CE%B1
+],
+[attachment
+charset=utf-8
+filename(lang:/UTF-8)=αÏÏείο για Ïην δοκιμαÏία
+])
+
+MIMEHDR([charset with language (2231)],[mimehdr06 mimehdr-charset-rfc2231
mimehdr-charset-rfc2231-01 mimehdr-charset-04],
+[],
+[attachment; charset=utf-8;
+
filename*=UTF-8'el_GR'%CE%B1%CF%81%CF%87%CE%B5%CE%AF%CE%BF%20%CE%B3%CE%B9%CE%B1%20%CF%84%CE%B7%CE%BD%20%CE%B4%CE%BF%CE%BA%CE%B9%CE%BC%CE%B1%CF%83%CE%AF%CE%B1
+],
+[attachment
+charset=utf-8
+filename(lang:el_GR/UTF-8)=αÏÏείο για Ïην δοκιμαÏία
+])
+
+MIMEHDR([charset with language and continuation (2231)],[mimehdr07
mimehdr-charset-rfc2231 mimehdr-charset-rfc2231-02 mimehdr-charset-05],
+[],
+[attachment; charset=utf-8;
+ filename*00*=UTF-8'el_GR'%CE%B1%CF%81%CF%87%CE%B5;
+ filename*01*=%CE%AF%CE%BF%20%CE%B3%CE;
+ filename*02*=%B9%CE%B1%20%CF%84%CE%B7;
+ filename*03*=%CE%BD%20%CE%B4%CE%BF%CE;
+ filename*04*=%BA%CE%B9%CE%BC%CE%B1%CF%83%CE%AF%CE%B1
+],
+[attachment
+charset=utf-8
+filename(lang:el_GR/UTF-8)=αÏÏείο για Ïην δοκιμαÏία
+])
+
+MIMEHDR([combined charset, lang and cset],[mimehdr08 mimehdr-comb
mimehdr-charset-rfc2231],
+[],
+[application/x-stuff
+ title*0*=us-ascii'en'This%20is%20even%20more%20
+ title*1*=%2A%2A%2Afun%2A%2A%2A%20
+ title*2="isn't it!"
+],
+[application/x-stuff
+title(lang:en/us-ascii)=This is even more ***fun*** isn't it!
+])
+
+m4_popdef([MIMEHDR])
diff --git a/libmailutils/tests/mimehdr.c b/libmailutils/tests/mimehdr.c
new file mode 100644
index 0000000..f02a252
--- /dev/null
+++ b/libmailutils/tests/mimehdr.c
@@ -0,0 +1,135 @@
+/* GNU Mailutils -- a suite of utilities for electronic mail
+ Copyright (C) 2005, 2007, 2009, 2010, 2011 Free Software Foundation,
+ Inc.
+
+ GNU Mailutils is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 3, or (at your option)
+ any later version.
+
+ GNU Mailutils is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with GNU Mailutils. If not, see <http://www.gnu.org/licenses/>. */
+
+#ifdef HAVE_CONFIG_H
+# include <config.h>
+#endif
+#include <stdlib.h>
+#include <string.h>
+#include <mailutils/assoc.h>
+#include <mailutils/header.h>
+#include <mailutils/message.h>
+#include <mailutils/mime.h>
+#include <mailutils/iterator.h>
+#include <mailutils/stream.h>
+#include <mailutils/stdstream.h>
+#include <mailutils/util.h>
+#include <mailutils/cstr.h>
+#include <mailutils/cctype.h>
+#include <mailutils/error.h>
+#include <mailutils/errno.h>
+
+struct named_param
+{
+ const char *name;
+ struct mu_mime_param const *param;
+};
+
+static int
+sort_names (void const *a, void const *b)
+{
+ struct named_param const *pa = a;
+ struct named_param const *pb = b;
+ return mu_c_strcasecmp (pa->name, pb->name);
+}
+
+static int
+print_named_param (void *item, void *data)
+{
+ struct named_param const *p = item;
+ struct mu_mime_param const *param = p->param;
+
+ mu_printf ("%s", p->name);
+ if (param->lang)
+ mu_printf ("(lang:%s/%s)", param->lang, param->cset);
+ mu_printf ("=%s\n", param->value);
+ return 0;
+}
+
+int
+main (int argc, char **argv)
+{
+ int i;
+ mu_stream_t tmp;
+ mu_transport_t trans[2];
+ char *value;
+ mu_assoc_t assoc;
+ mu_iterator_t itr;
+ mu_list_t list;
+ char *charset = NULL;
+
+ mu_set_program_name (argv[0]);
+ for (i = 1; i < argc; i++)
+ {
+ char *opt = argv[i];
+
+ if (strncmp (opt, "-debug=", 7) == 0)
+ mu_debug_parse_spec (opt + 7);
+ else if (strncmp (opt, "-charset=", 9) == 0)
+ charset = opt + 9;
+ else if (strcmp (opt, "-h") == 0 || strcmp (opt, "-help") == 0)
+ {
+ mu_printf ("usage: %s [-charset=cs] [-debug=SPEC]", mu_program_name);
+ return 0;
+ }
+ else
+ {
+ mu_error ("unknown option %s", opt);
+ return 1;
+ }
+ }
+
+ if (i != argc)
+ {
+ mu_error ("too many arguments");
+ return 1;
+ }
+
+ MU_ASSERT (mu_memory_stream_create (&tmp, MU_STREAM_RDWR));
+ MU_ASSERT (mu_stream_copy (tmp, mu_strin, 0, NULL));
+ MU_ASSERT (mu_stream_write (tmp, "", 1, NULL));
+ MU_ASSERT (mu_stream_ioctl (tmp, MU_IOCTL_TRANSPORT, MU_IOCTL_OP_GET,
+ trans));
+
+ MU_ASSERT (mu_mime_header_parse ((char*)trans[0], charset, &value, &assoc));
+
+ mu_printf ("%s\n", value);
+ MU_ASSERT (mu_list_create (&list));
+ MU_ASSERT (mu_assoc_get_iterator (assoc, &itr));
+ for (mu_iterator_first (itr); !mu_iterator_is_done (itr);
+ mu_iterator_next (itr))
+ {
+ const char *name;
+ struct mu_mime_param *p;
+ struct named_param *np;
+
+ mu_iterator_current_kv (itr, (const void **)&name, (void**)&p);
+ np = malloc (sizeof (*np));
+ if (!np)
+ abort ();
+ np->name = name;
+ np->param = p;
+ MU_ASSERT (mu_list_append (list, np));
+ }
+ mu_iterator_destroy (&itr);
+
+ mu_list_sort (list, sort_names);
+ mu_list_foreach (list, print_named_param, NULL);
+
+ return 0;
+}
+
diff --git a/libmailutils/tests/testsuite.at b/libmailutils/tests/testsuite.at
index 4c1b427..b9b889c 100644
--- a/libmailutils/tests/testsuite.at
+++ b/libmailutils/tests/testsuite.at
@@ -96,4 +96,6 @@ m4_include([imapio.at])
m4_include([scantime.at])
m4_include([strftime.at])
-m4_include([fsaf.at])
\ No newline at end of file
+m4_include([fsaf.at])
+
+m4_include([mimehdr.at])
hooks/post-receive
--
GNU Mailutils
| [Prev in Thread] |
Current Thread |
[Next in Thread] |
- [SCM] GNU Mailutils branch, master, updated. release-2.2-540-g3c5aed3,
Sergey Poznyakoff <=