* distribution for complete copyright information.
*/
-#include <h/mh.h>
-#include <h/fmt_scan.h>
-#include <h/mime.h>
-#include <h/mhparse.h>
+#include "h/mh.h"
+#include "sbr/m_name.h"
+#include "sbr/m_gmprot.h"
+#include "sbr/m_getfld.h"
+#include "sbr/getarguments.h"
+#include "sbr/concat.h"
+#include "sbr/seq_setprev.h"
+#include "sbr/seq_setcur.h"
+#include "sbr/seq_save.h"
+#include "sbr/smatch.h"
+#include "sbr/fmt_rfc2047.h"
+#include "sbr/cpydata.h"
+#include "sbr/trimcpy.h"
+#include "sbr/m_convert.h"
+#include "sbr/m_backup.h"
+#include "sbr/getfolder.h"
+#include "sbr/folder_read.h"
+#include "sbr/context_save.h"
+#include "sbr/context_replace.h"
+#include "sbr/context_find.h"
+#include "sbr/readconfig.h"
+#include "sbr/ambigsw.h"
+#include "sbr/path.h"
+#include "sbr/print_version.h"
+#include "sbr/print_help.h"
+#include "sbr/error.h"
+#include "h/fmt_scan.h"
+#include "h/mime.h"
+#include "h/mhparse.h"
#include "h/done.h"
-#include <h/utils.h>
-#include <h/signals.h>
+#include "h/utils.h"
+#include "h/signals.h"
#include "sbr/m_maildir.h"
#include "sbr/m_mktemp.h"
#include "sbr/mime_type.h"
X("decodetext 8bit|7bit|binary", 0, DECODETEXTSW) \
X("nodecodetext", 0, NDECODETEXTSW) \
X("decodetypes", 0, DECODETYPESW) \
+ X("decodeheaderfieldbodies utf-8", 0, DECODEHEADERFIELDBODIESSW) \
+ X("nodecodeheaderfieldbodies", 0, NDECODEHEADERFIELDBODIESSW) \
X("crlflinebreaks", 0, CRLFLINEBREAKSSW) \
X("nocrlflinebreaks", 0, NCRLFLINEBREAKSSW) \
X("textcharset", 0, TEXTCHARSETSW) \
int replacetextplain;
int decodetext;
char *decodetypes;
+ char *decodeheaderfieldbodies; /* Either NULL or "utf-8". */
/* Whether to use CRLF linebreaks, per RFC 2046 Sec. 4.1.1, par.1. */
int lf_line_endings;
char *textcharset;
static int less_restrictive (int, int);
static int convert_charsets (CT, char *, int *);
static int fix_always (CT, int *);
+static int decode_header_field_bodies (CT, int *);
static int fix_filename_param (char *, char *, PM *, PM *);
static int fix_filename_encoding (CT);
static int write_content (CT, const char *, char *, FILE *, int, int);
int
-main (int argc, char **argv) {
+main (int argc, char **argv)
+{
int msgnum;
char *cp, *file = NULL, *folder = NULL;
char *maildir = NULL, buf[100], *outfile = NULL;
fx.replacetextplain = 0;
fx.decodetext = CE_8BIT;
fx.decodetypes = "text,application/ics"; /* Default, per man page. */
+ fx.decodeheaderfieldbodies = NULL;
fx.lf_line_endings = 0;
fx.textcharset = NULL;
}
fx.decodetypes = cp;
continue;
+ case DECODEHEADERFIELDBODIESSW:
+ if (! (cp = *argp++) || *cp == '-') {
+ die("missing argument to %s", argp[-2]);
+ }
+ fx.decodeheaderfieldbodies = cp;
+ if (strcasecmp (cp, "utf-8") && strcasecmp (cp, "utf8")) {
+ /* Because UTF-8 strings can't have embedded nulls. Other
+ encodings support that, too, but we won't bother to
+ enumerate them. */
+ die("-decodeheaderfieldbodies only supports utf-8");
+ }
+ continue;
+ case NDECODEHEADERFIELDBODIESSW:
+ fx.decodeheaderfieldbodies = NULL;
+ continue;
case CRLFLINEBREAKSSW:
fx.lf_line_endings = 0;
continue;
*/
static int
mhfixmsgsbr (CT *ctp, char *maildir, const fix_transformations *fx,
- FILE **infp, char *outfile, FILE **outfp) {
+ FILE **infp, char *outfile, FILE **outfp)
+{
/* Store input filename in case one of the transformations, i.e.,
fix_boundary(), rewrites to a tmp file. */
char *input_filename = maildir
&message_mods);
update_cte (*ctp);
}
+ if (status == OK && fx->decodeheaderfieldbodies) {
+ status = decode_header_field_bodies(*ctp, &message_mods);
+ }
if (status == OK && fx->textcharset != NULL) {
status = convert_charsets (*ctp, fx->textcharset, &message_mods);
}
*/
static int
copy_input_to_output (const char *input_filename, FILE *infp,
- const char *output_filename, FILE *outfp) {
+ const char *output_filename, FILE *outfp)
+{
int in = fileno (infp);
int out = fileno (outfp);
int status = OK;
* Fix mismatched outer level boundary.
*/
static int
-fix_boundary (CT *ct, int *message_mods) {
+fix_boundary (CT *ct, int *message_mods)
+{
struct multipart *mp;
int status = OK;
* Find boundary at end of multipart.
*/
static int
-get_multipart_boundary (CT ct, char **part_boundary) {
+get_multipart_boundary (CT ct, char **part_boundary)
+{
char buffer[NMH_BUFSIZ];
char *end_boundary = NULL;
off_t begin = (off_t) ct->c_end > (off_t) (ct->c_begin + sizeof buffer)
* Open and copy ct->c_file to file, replacing the multipart boundary.
*/
static int
-replace_boundary (CT ct, char *file, char *boundary) {
+replace_boundary (CT ct, char *file, char *boundary)
+{
FILE *fpin, *fpout;
int compnum, state;
char buf[NMH_BUFSIZ], name[NAMESZ];
* Fix Content-Type header to reflect the content of its part.
*/
static int
-fix_types (CT ct, svector_t fixtypes, int *message_mods) {
+fix_types (CT ct, svector_t fixtypes, int *message_mods)
+{
int status = OK;
switch (ct->c_type) {
* Replace a substring, allocating space to hold the new one.
*/
char *
-replace_substring (char **str, const char *old, const char *new) {
+replace_substring (char **str, const char *old, const char *new)
+{
char *cp;
if ((cp = strstr (*str, old))) {
* Remove a name=value parameter, given just its name, from a header value.
*/
char *
-remove_parameter (char *str, const char *name) {
+remove_parameter (char *str, const char *name)
+{
/* It looks to me, based on the BNF in RFC 2045, than there can't
be whitespace between the parameter name and the "=", or
between the "=" and the parameter value. */
* 8 bit.
*/
static int
-fix_composite_cte (CT ct, int *message_mods) {
+fix_composite_cte (CT ct, int *message_mods)
+{
int status = OK;
if (ct->c_type == CT_MESSAGE || ct->c_type == CT_MULTIPART) {
* Set content encoding.
*/
static int
-set_ce (CT ct, int encoding) {
+set_ce (CT ct, int encoding)
+{
const char *ce = ce_str (encoding);
const struct str2init *ctinit = get_ce_method (ce);
* Make sure each text part has a corresponding text/plain part.
*/
static int
-ensure_text_plain (CT *ct, CT parent, int *message_mods, int replacetextplain) {
+ensure_text_plain (CT *ct, CT parent, int *message_mods, int replacetextplain)
+{
int status = OK;
switch ((*ct)->c_type) {
*/
static int
find_textplain_sibling (CT parent, int replacetextplain,
- int *new_subpart_number) {
+ int *new_subpart_number)
+{
struct multipart *mp = (struct multipart *) parent->c_ctparams;
struct part *part, *prev;
bool has_text_plain = false;
* Insert a new text/plain part.
*/
static int
-insert_new_text_plain_part (CT ct, int new_subpart_number, CT parent) {
+insert_new_text_plain_part (CT ct, int new_subpart_number, CT parent)
+{
struct multipart *mp = (struct multipart *) parent->c_ctparams;
struct part *new_part;
* Create a text/plain part to go along with non-plain sibling part.
*/
static CT
-build_text_plain_part (CT encoded_part) {
+build_text_plain_part (CT encoded_part)
+{
CT tp_part = divide_part (encoded_part);
char *tmp_plain_file = NULL;
* Slip new text/plain part into a new multipart/alternative.
*/
static int
-insert_into_new_mp_alt (CT *ct, int *message_mods) {
+insert_into_new_mp_alt (CT *ct, int *message_mods)
+{
CT tp_part = build_text_plain_part (*ct);
int status = OK;
* Clone a MIME part.
*/
static CT
-divide_part (CT ct) {
+divide_part (CT ct)
+{
CT new_part;
NEW0(new_part);
* Copy the content info from one part to another.
*/
static void
-copy_ctinfo (CI dest, CI src) {
+copy_ctinfo (CI dest, CI src)
+{
PM s_pm, d_pm;
dest->ci_type = src->ci_type ? mh_xstrdup (src->ci_type) : NULL;
* Decode content.
*/
static int
-decode_part (CT ct) {
+decode_part (CT ct)
+{
char *tmp_decoded;
int status;
FILE *file;
* be in the future for other than text types.
*/
static int
-reformat_part (CT ct, char *file, char *type, char *subtype, int c_type) {
+reformat_part (CT ct, char *file, char *type, char *subtype, int c_type)
+{
int output_subtype, output_encoding;
const char *reason = NULL;
char *cp, *cf;
* Fill in a multipart/alternative part.
*/
static CT
-build_multipart_alt (CT first_alt, CT new_part, int type, int subtype) {
+build_multipart_alt (CT first_alt, CT new_part, int type, int subtype)
+{
char *boundary_prefix = "----=_nmh-multipart";
char *boundary = concat (boundary_prefix, first_alt->c_partno, NULL);
char *boundary_indicator = "; boundary=";
* Check that the boundary does not appear in the content.
*/
static int
-boundary_in_content (FILE **fp, char *file, const char *boundary) {
+boundary_in_content (FILE **fp, char *file, const char *boundary)
+{
char buffer[NMH_BUFSIZ];
size_t bytes_read;
bool found_boundary = false;
* Remove all non-Content headers.
*/
static void
-transfer_noncontent_headers (CT old, CT new) {
+transfer_noncontent_headers (CT old, CT new)
+{
HF hp, hp_prev;
hp_prev = hp = old->c_first_hf;
* Set content type.
*/
static int
-set_ct_type (CT ct, int type, int subtype, int encoding) {
+set_ct_type (CT ct, int type, int subtype, int encoding)
+{
char *typename = ct_type_str (type);
char *subtypename = ct_subtype_str (type, subtype);
/* E.g, " text/plain" */
*/
static int
decode_text_parts (CT ct, int encoding, const char *decodetypes,
- int *message_mods) {
+ int *message_mods)
+{
int status = OK;
int lf_line_endings = 0;
* decodetypes (which came from the -decodetypes switch).
*/
static int
-should_decode(const char *decodetypes, const char *type, const char *subtype) {
+should_decode(const char *decodetypes, const char *type, const char *subtype)
+{
/* Quick search for matching type[/subtype] in decodetypes: bracket
decodetypes with commas, then search for ,type, and ,type/subtype, in
it. */
* to a string explaining why.
*/
static int
-content_encoding (CT ct, const char **reason) {
+content_encoding (CT ct, const char **reason)
+{
CE ce = &ct->c_cefile;
int encoding = CE_7BIT;
* Strip carriage returns from content.
*/
static int
-strip_crs (CT ct, int *message_mods) {
+strip_crs (CT ct, int *message_mods)
+{
char *charset = content_charset (ct);
int status = OK;
* of the part C-T-E's.
*/
static void
-update_cte (CT ct) {
+update_cte (CT ct)
+{
const int least_restrictive_enc = least_restrictive_encoding (ct);
if (least_restrictive_enc != CE_UNKNOWN &&
* within a message.
*/
static int
-least_restrictive_encoding (CT ct) {
+least_restrictive_encoding (CT ct)
+{
int encoding = CE_UNKNOWN;
switch (ct->c_type) {
* CE_8BIT is less restrictive than CE_7BIT.
*/
static int
-less_restrictive (int encoding, int second_encoding) {
+less_restrictive (int encoding, int second_encoding)
+{
switch (second_encoding) {
case CE_BINARY:
return encoding != CE_BINARY;
* Convert character set of each part.
*/
static int
-convert_charsets (CT ct, char *dest_charset, int *message_mods) {
+convert_charsets (CT ct, char *dest_charset, int *message_mods)
+{
int status = OK;
switch (ct->c_type) {
* headers, respectively.
*/
static int
-fix_always (CT ct, int *message_mods) {
+fix_always (CT ct, int *message_mods)
+{
int status = OK;
switch (ct->c_type) {
}
+/*
+ * Decodes UTF-8 encoded header values. Similar to fix_filename_param(), but
+ * does not modify any MIME parameter values.
+ */
+static int
+decode_header_field_bodies (CT ct, int *message_mods)
+{
+ int status = OK;
+
+ switch (ct->c_type) {
+ case CT_MULTIPART: {
+ struct multipart *m = (struct multipart *) ct->c_ctparams;
+ struct part *part;
+
+ for (part = m->mp_parts; status == OK && part; part = part->mp_next) {
+ status = decode_header_field_bodies (part->mp_part, message_mods);
+ }
+ break;
+ }
+
+ case CT_MESSAGE:
+ if (ct->c_subtype == MESSAGE_EXTERNAL) {
+ struct exbody *e = (struct exbody *) ct->c_ctparams;
+
+ status = decode_header_field_bodies (e->eb_content, message_mods);
+ }
+ break;
+ }
+
+ HF hf;
+
+ for (hf = ct->c_first_hf; hf; hf = hf->next) {
+ /* Only decode UTF-8 values. */
+ if (hf->value && has_suffix(hf->value, "?=\n") &&
+ (! strncasecmp (hf->value, " =?utf8?", 8) ||
+ ! strncasecmp (hf->value, " =?utf-8?", 9))) {
+ /* Looks like an RFC 2047 encoded parameter. */
+ char decoded[PATH_MAX + 1];
+
+ if (decode_rfc2047 (hf->value, decoded, sizeof decoded)) {
+ const size_t len = strlen(decoded);
+
+ /* decode_rfc2047() could truncate if the buffer fills up.
+ Detect and discard if that happened. */
+ if (len < sizeof(decoded) - 1 && strcmp(hf->value, decoded)) {
+ hf->value = mh_xrealloc (hf->value, len + 1);
+ strncpy (hf->value, decoded, len + 1);
+ ++*message_mods;
+ }
+ } else {
+ inform("failed to decode %s parameter %s", hf->name, hf->value);
+ status = NOTOK;
+ }
+ }
+ }
+
+ return status;
+}
+
+
/*
* Factor out common code for loops in fix_filename_encoding().
*/
static int
-fix_filename_param (char *name, char *value, PM *first_pm, PM *last_pm) {
+fix_filename_param (char *name, char *value, PM *first_pm, PM *last_pm)
+{
bool fixed = false;
if (has_prefix(value, "=?") && has_suffix(value, "?=")) {
* headers, respectively.
*/
static int
-fix_filename_encoding (CT ct) {
+fix_filename_encoding (CT ct)
+{
PM pm;
HF hf;
int fixed = 0;
*/
static int
write_content (CT ct, const char *input_filename, char *outfile, FILE *outfp,
- int modify_inplace, int message_mods) {
+ int modify_inplace, int message_mods)
+{
int status = OK;
if (modify_inplace) {
* function to do it. It touches the parts the decodetypes identifies.
*/
static void
-set_text_ctparams(CT ct, char *decodetypes, int lf_line_endings) {
+set_text_ctparams(CT ct, char *decodetypes, int lf_line_endings)
+{
switch (ct->c_type) {
case CT_MULTIPART: {
struct multipart *m = (struct multipart *) ct->c_ctparams;
* use the standard MH backup file.
*/
static int
-remove_file (const char *file) {
+remove_file (const char *file)
+{
if (rmmproc) {
char *rmm_command = concat (rmmproc, " ", file, NULL);
int status = system (rmm_command);
* Output formatted message to user.
*/
static void
-report (char *what, char *partno, char *filename, char *message, ...) {
+report (char *what, char *partno, char *filename, char *message, ...)
+{
va_list args;
char *fmt;
projects / nmh / blobdiff