X-Git-Url: https://diplodocus.org/git/nmh/blobdiff_plain/e2eebc77cc570f77d906bc0cc6890092068468cb..f1920d78123667716f2321d37ce37628603b2700:/uip/mhfixmsg.c diff --git a/uip/mhfixmsg.c b/uip/mhfixmsg.c index af6cd889..65abd513 100644 --- a/uip/mhfixmsg.c +++ b/uip/mhfixmsg.c @@ -16,6 +16,9 @@ #define MHFIXMSG_SWITCHES \ X("decodetext 8bit|7bit", 0, DECODETEXTSW) \ X("nodecodetext", 0, NDECODETEXTSW) \ + X("decodetypes", 0, DECODETYPESW) \ + X("crlflinebreaks", 0, CRLFLINEBREAKSSW) \ + X("nocrlflinebreaks", 0, NCRLFLINEBREAKSSW) \ X("textcharset", 0, TEXTCHARSETSW) \ X("notextcharset", 0, NTEXTCHARSETSW) \ X("reformat", 0, REFORMATSW) \ @@ -80,6 +83,9 @@ typedef struct fix_transformations { int reformat; int replacetextplain; int decodetext; + char *decodetypes; + /* Whether to use CRLF linebreaks, per RFC 2046 Sec. 4.1.1, par.1. */ + int lf_line_endings; char *textcharset; } fix_transformations; @@ -89,12 +95,14 @@ static int get_multipart_boundary (CT, char **); static int replace_boundary (CT, char *, char *); static int fix_types (CT, svector_t, int *); static char *replace_substring (char **, const char *, const char *); +static char *remove_parameter (char *, const char *); static int fix_multipart_cte (CT, int *); static int set_ce (CT, int); static int ensure_text_plain (CT *, CT, int *, int); static int find_textplain_sibling (CT, int, int *); static int insert_new_text_plain_part (CT, int, CT); static CT build_text_plain_part (CT); +static int insert_into_new_mp_alt (CT *, int *); static CT divide_part (CT); static void copy_ctinfo (CI, CI); static int decode_part (CT); @@ -104,12 +112,14 @@ static CT build_multipart_alt (CT, CT, int, int); static int boundary_in_content (FILE **, char *, const char *); static void transfer_noncontent_headers (CT, CT); static int set_ct_type (CT, int type, int subtype, int encoding); -static int decode_text_parts (CT, int, int *); +static int decode_text_parts (CT, int, const char *, int *); +static int should_decode(const char *, const char *, const char *); static int content_encoding (CT, const char **); static int strip_crs (CT, int *); static int convert_charsets (CT, char *, int *); static int fix_always (CT, int *); static int write_content (CT, char *, char *, int, int); +static void set_text_ctparams(CT, char *, int); static int remove_file (char *); static void report (char *, char *, char *, char *, ...); static void pipeser (int); @@ -133,6 +143,8 @@ main (int argc, char **argv) { fx.fixtypes = NULL; fx.replacetextplain = 0; fx.decodetext = CE_8BIT; + fx.decodetypes = "text,application/ics"; /* Default, per man page. */ + fx.lf_line_endings = 0; fx.textcharset = NULL; if (nmh_init(argv[0], 1)) { return 1; } @@ -177,6 +189,17 @@ main (int argc, char **argv) { case NDECODETEXTSW: fx.decodetext = 0; continue; + case DECODETYPESW: + if (! (cp = *argp++) || *cp == '-') + adios (NULL, "missing argument to %s", argp[-2]); + fx.decodetypes = cp; + continue; + case CRLFLINEBREAKSSW: + fx.lf_line_endings = 0; + continue; + case NCRLFLINEBREAKSSW: + fx.lf_line_endings = 1; + continue; case TEXTCHARSETSW: if (! (cp = *argp++) || (*cp == '-' && cp[1])) adios (NULL, "missing argument to %s", argp[-2]); @@ -323,7 +346,10 @@ main (int argc, char **argv) { } ctp = cts; - if ((ct = parse_mime (file))) { *ctp++ = ct; } + if ((ct = parse_mime (file))) { + set_text_ctparams(ct, fx.decodetypes, fx.lf_line_endings); + *ctp++ = ct; + } } else { /* * message(s) are coming from a folder @@ -364,7 +390,10 @@ main (int argc, char **argv) { char *msgnam; msgnam = m_name (msgnum); - if ((ct = parse_mime (msgnam))) { *ctp++ = ct; } + if ((ct = parse_mime (msgnam))) { + set_text_ctparams(ct, fx.decodetypes, fx.lf_line_endings); + *ctp++ = ct; + } } } @@ -443,7 +472,7 @@ mhfixmsgsbr (CT *ctp, const fix_transformations *fx, char *outfile) { ensure_text_plain (ctp, NULL, &message_mods, fx->replacetextplain); } if (status == OK && fx->decodetext) { - status = decode_text_parts (*ctp, fx->decodetext, &message_mods); + status = decode_text_parts (*ctp, fx->decodetext, fx->decodetypes, &message_mods); } if (status == OK && fx->textcharset != NULL) { status = convert_charsets (*ctp, fx->textcharset, &message_mods); @@ -554,7 +583,16 @@ fix_boundary (CT *ct, int *message_mods) { } free (part_boundary); + } else { + /* Couldn't fix the boundary. Report failure so that mhfixmsg + doesn't modify the message. */ + status = NOTOK; } + } else { + /* No multipart struct, even though the content type is + CT_MULTIPART. Report failure so that mhfixmsg doesn't modify + the message. */ + status = NOTOK; } } @@ -885,6 +923,51 @@ replace_substring (char **str, const char *old, const char *new) { } } +/* + * Remove a name=value parameter, given just its name, from a header value. + */ +char * +remove_parameter (char *str, const char *name) { + /* It looks to me, based on the BNF in RFC 2045, than there can't + be whitespace betwwen the parameter name and the "=", or + between the "=" and the parameter value. */ + char *param_name = concat (name, "=", NULL); + char *cp; + + if ((cp = strstr (str, param_name))) { + char *start, *end; + size_t count = 1; + + /* Remove any leading spaces, before the parameter name. */ + for (start = cp; + start > str && isspace ((unsigned char) *(start-1)); + --start) { + continue; + } + /* Remove a leading semicolon. */ + if (start > str && *(start-1) == ';') { --start; } + + end = cp + strlen (name) + 1; + if (*end == '"') { + /* Skip past the quoted value, and then the final quote. */ + for (++end ; *end && *end != '"'; ++end) { continue; } + ++end; + } else { + /* Skip past the value. */ + for (++end ; *end && ! isspace ((unsigned char) *end); ++end) {} + } + + /* Count how many characters need to be moved. Include + trailing null, which is accounted for by the + initialization of count to 1. */ + for (cp = end; *cp; ++cp) { ++count; } + (void) memmove (start, end, count); + } + + free (param_name); + + return str; +} static int fix_multipart_cte (CT ct, int *message_mods) { @@ -1063,51 +1146,72 @@ ensure_text_plain (CT *ct, CT parent, int *message_mods, int replacetextplain) { free (type_subtype); if (! has_text_plain) { - /* Parent is a multipart/alternative. Insert a new - text/plain subpart. */ - const int inserted = - insert_new_text_plain_part (*ct, new_subpart_number, - parent); - if (inserted) { - ++*message_mods; - if (verbosw) { - report (NULL, parent->c_partno, parent->c_file, - "insert text/plain part"); + struct multipart *mp = (struct multipart *) parent->c_ctparams; + struct part *part; + int siblings = 0; + + for (part = mp->mp_parts; part; part = part->mp_next) { + if (*ct != part->mp_part) { + ++siblings; } - } else { - status = NOTOK; } - } - } else { - /* Slip new text/plain part into a new multipart/alternative. */ - CT tp_part = build_text_plain_part (*ct); - if (tp_part) { - CT mp_alt = build_multipart_alt (*ct, tp_part, CT_MULTIPART, - MULTI_ALTERNATE); - if (mp_alt) { - struct multipart *mp = - (struct multipart *) mp_alt->c_ctparams; + if (siblings) { + /* Parent is a multipart/related. Insert a new + text/plain subpart in a new multipart/alternative. */ + if (insert_into_new_mp_alt (ct, message_mods)) { + /* Not an error if text/plain couldn't be added. */ + } + } else { + /* There are no siblings, so insert a new text/plain + subpart, and change the parent type from + multipart/related to multipart/alternative. */ + const int inserted = + insert_new_text_plain_part (*ct, new_subpart_number, + parent); + + if (inserted) { + HF hf; - if (mp && mp->mp_parts) { - mp->mp_parts->mp_part = tp_part; - /* Make the new multipart/alternative the parent. */ - *ct = mp_alt; + parent->c_subtype = MULTI_ALTERNATE; + parent->c_ctinfo.ci_subtype = getcpy ("alternative"); + if (! replace_substring (&parent->c_ctline, "/related", + "/alternative")) { + advise (NULL, + "did not find multipart/related in %s", + parent->c_ctline); + } - ++*message_mods; - if (verbosw) { - report (NULL, (*ct)->c_partno, (*ct)->c_file, - "insert text/plain part"); + /* Update Content-Type header field. */ + for (hf = parent->c_first_hf; hf; hf = hf->next) { + if (! strcasecmp (TYPE_FIELD, hf->name)) { + if (replace_substring (&hf->value, "/related", + "/alternative")) { + ++*message_mods; + if (verbosw) { + report (NULL, parent->c_partno, + parent->c_file, + "insert text/plain part"); + } + + /* Remove, e.g., type="text/html" from + multipart/alternative. */ + remove_parameter (hf->value, "type"); + break; + } else { + advise (NULL, "did not find multipart/" + "related in header %s", + hf->value); + } + } } } else { - free_content (tp_part); - free_content (mp_alt); - status = NOTOK; + /* Not an error if text/plain couldn't be inserted. */ } - } else { - status = NOTOK; } - } else { + } + } else { + if (insert_into_new_mp_alt (ct, message_mods)) { status = NOTOK; } } @@ -1240,6 +1344,43 @@ build_text_plain_part (CT encoded_part) { } +/* Slip new text/plain part into a new multipart/alternative. */ +static int +insert_into_new_mp_alt (CT *ct, int *message_mods) { + CT tp_part = build_text_plain_part (*ct); + int status = OK; + + if (tp_part) { + CT mp_alt = build_multipart_alt (*ct, tp_part, CT_MULTIPART, + MULTI_ALTERNATE); + if (mp_alt) { + struct multipart *mp = (struct multipart *) mp_alt->c_ctparams; + + if (mp && mp->mp_parts) { + mp->mp_parts->mp_part = tp_part; + /* Make the new multipart/alternative the parent. */ + *ct = mp_alt; + + ++*message_mods; + if (verbosw) { + report (NULL, (*ct)->c_partno, (*ct)->c_file, + "insert text/plain part"); + } + } else { + free_content (tp_part); + free_content (mp_alt); + status = NOTOK; + } + } else { + status = NOTOK; + } + } else { + /* Not an error if text/plain couldn't be built. */ + } + + return status; +} + static CT divide_part (CT ct) { CT new_part; @@ -1642,11 +1783,39 @@ set_ct_type (CT ct, int type, int subtype, int encoding) { static int -decode_text_parts (CT ct, int encoding, int *message_mods) { +decode_text_parts (CT ct, int encoding, const char *decodetypes, int *message_mods) { int status = OK; + int lf_line_endings = 0; switch (ct->c_type) { - case CT_TEXT: + case CT_MULTIPART: { + struct multipart *m = (struct multipart *) ct->c_ctparams; + struct part *part; + + /* Should check to see if the body for this part is encoded? + For now, it gets passed along as-is by InitMultiPart(). */ + for (part = m->mp_parts; status == OK && part; part = part->mp_next) { + status = decode_text_parts (part->mp_part, encoding, decodetypes, message_mods); + } + break; + } + + case CT_MESSAGE: + if (ct->c_subtype == MESSAGE_EXTERNAL) { + struct exbody *e = (struct exbody *) ct->c_ctparams; + + status = decode_text_parts (e->eb_content, encoding, decodetypes, message_mods); + } + break; + + default: + if (! should_decode(decodetypes, ct->c_ctinfo.ci_type, ct->c_ctinfo.ci_subtype)) { + break; + } + + lf_line_endings = + ct->c_ctparams && ((struct text *) ct->c_ctparams)->lf_line_endings; + switch (ct->c_encoding) { case CE_BASE64: case CE_QUOTED: { @@ -1698,7 +1867,9 @@ decode_text_parts (CT ct, int encoding, int *message_mods) { report (NULL, ct->c_partno, ct->c_file, "decode%s", ct->c_ctline ? ct->c_ctline : ""); } - strip_crs (ct, message_mods); + if (lf_line_endings) { + strip_crs (ct, message_mods); + } } else { status = NOTOK; } @@ -1710,39 +1881,49 @@ decode_text_parts (CT ct, int encoding, int *message_mods) { } case CE_8BIT: case CE_7BIT: - strip_crs (ct, message_mods); + if (lf_line_endings) { + strip_crs (ct, message_mods); + } break; default: break; } break; - - case CT_MULTIPART: { - struct multipart *m = (struct multipart *) ct->c_ctparams; - struct part *part; - - /* Should check to see if the body for this part is encoded? - For now, it gets passed along as-is by InitMultiPart(). */ - for (part = m->mp_parts; status == OK && part; part = part->mp_next) { - status = decode_text_parts (part->mp_part, encoding, message_mods); - } - break; } - case CT_MESSAGE: - if (ct->c_subtype == MESSAGE_EXTERNAL) { - struct exbody *e = (struct exbody *) ct->c_ctparams; + return status; +} - status = decode_text_parts (e->eb_content, encoding, message_mods); - } - break; - default: - break; +/* Determine if the part with type[/subtype] should be decoded, according to + decodetypes (which came from the -decodetypes switch). */ +static int +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. */ + + int found_match = 0; + char *delimited_decodetypes = concat(",", decodetypes, ",", NULL); + char *delimited_type = concat(",", type, ",", NULL); + + if (nmh_strcasestr(delimited_decodetypes, delimited_type)) { + found_match = 1; + } else if (subtype != NULL) { + char *delimited_type_subtype = + concat(",", type, "/", subtype, ",", NULL); + + if (nmh_strcasestr(delimited_decodetypes, delimited_type_subtype)) { + found_match = 1; + } + free(delimited_type_subtype); } - return status; + free(delimited_type); + free(delimited_decodetypes); + + return found_match; } @@ -2167,6 +2348,44 @@ write_content (CT ct, char *input_filename, char *outfile, int modify_inplace, } +/* + * parse_mime() does not set lf_line_endings in struct text, so use this function to do it. + * It touches the parts the decodetypes identifies. + */ +static void +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; + struct part *part; + + for (part = m->mp_parts; part; part = part->mp_next) { + set_text_ctparams(part->mp_part, decodetypes, lf_line_endings); + } + break; + } + + case CT_MESSAGE: + if (ct->c_subtype == MESSAGE_EXTERNAL) { + struct exbody *e = (struct exbody *) ct->c_ctparams; + + set_text_ctparams(e->eb_content, decodetypes, lf_line_endings); + } + break; + + default: + if (should_decode(decodetypes, ct->c_ctinfo.ci_type, ct->c_ctinfo.ci_subtype)) { + if (ct->c_ctparams == NULL) { + if ((ct->c_ctparams = (struct text *) mh_xcalloc (1, sizeof (struct text))) == NULL) { + adios (NULL, "out of memory"); + } + } + ((struct text *) ct->c_ctparams)->lf_line_endings = lf_line_endings; + } + } +} + + /* * If "rmmproc" is defined, call that to remove the file. Otherwise, * use the standard MH backup file.