X-Git-Url: https://diplodocus.org/git/nmh/blobdiff_plain/4a78cbcd4fa986d9c1e7bd0a5a4bdb619faeb7cb..5c34a38ad47b5bec1535a4c6c3272f8b4b2364bd:/uip/mhparse.c diff --git a/uip/mhparse.c b/uip/mhparse.c index 9489a8e4..4509e764 100644 --- a/uip/mhparse.c +++ b/uip/mhparse.c @@ -80,6 +80,19 @@ struct k2v SubApplication[] = { { NULL, APPLICATION_UNKNOWN } /* this one must be last! */ }; +/* + * Mapping of names of CTE types in mhbuild directives + */ +static struct k2v EncodingType[] = { + { "8bit", CE_8BIT }, + { "qp", CE_QUOTED }, + { "q-p", CE_QUOTED }, + { "quoted-printable", CE_QUOTED }, + { "b64", CE_BASE64 }, + { "base64", CE_BASE64 }, + { NULL, 0 }, +}; + /* mhcachesbr.c */ int find_cache (CT, int, int *, char *, char *, int); @@ -122,6 +135,7 @@ static int readDigest (CT, char *); static int get_leftover_mp_content (CT, int); static int InitURL (CT); static int openURL (CT, char **); +static size_t param_len(PM, int, size_t, int *); struct str2init str2cts[] = { { "application", CT_APPLICATION, InitApplication }, @@ -205,12 +219,12 @@ parse_mime (char *file) fflush (fp); if (ferror (stdin)) { - unlink (file); + (void) m_unlink (file); advise ("stdin", "error reading"); return NULL; } if (ferror (fp)) { - unlink (file); + (void) m_unlink (file); advise (file, "error writing"); return NULL; } @@ -222,7 +236,7 @@ parse_mime (char *file) if (!(ct = get_content (fp, file, 1))) { if (is_stdin) - unlink (file); + (void) m_unlink (file); advise (NULL, "unable to decode %s", file); return NULL; } @@ -838,6 +852,47 @@ magic_skip: cp++; } + /* + * Get any extension directives (right now just the content transfer + * encoding, but maybe others) that we care about. + */ + + if (magic && *cp == '*') { + /* + * See if it's a CTE we match on + */ + struct k2v *kv; + + dp = ++cp; + while (*cp != '\0' && ! isspace((unsigned char) *cp)) + cp++; + + if (dp == cp) { + advise (NULL, "invalid null transfer encoding specification"); + return NOTOK; + } + + if (*cp != '\0') + *cp++ = '\0'; + + ct->c_reqencoding = CE_UNKNOWN; + + for (kv = EncodingType; kv->kv_key; kv++) { + if (strcasecmp(kv->kv_key, dp) == 0) { + ct->c_reqencoding = kv->kv_value; + break; + } + } + + if (ct->c_reqencoding == CE_UNKNOWN) { + advise (NULL, "invalid CTE specification: \"%s\"", dp); + return NOTOK; + } + + while (isspace ((unsigned char) *cp)) + cp++; + } + /* * Check if anything is left over */ @@ -953,7 +1008,8 @@ InitText (CT ct) { char buffer[BUFSIZ]; char *chset = NULL; - char **ap, **ep, *cp; + char *cp; + PM pm; struct k2v *kv; struct text *t; CI ci = &ct->c_ctinfo; @@ -974,13 +1030,13 @@ InitText (CT ct) ct->c_ctparams = (void *) t; /* scan for charset parameter */ - for (ap = ci->ci_attrs, ep = ci->ci_values; *ap; ap++, ep++) - if (!strcasecmp (*ap, "charset")) + for (pm = ci->ci_first_pm; pm; pm = pm->pm_next) + if (!strcasecmp (pm->pm_name, "charset")) break; /* check if content specified a character set */ - if (*ap) { - chset = *ep; + if (pm) { + chset = pm->pm_value; t->tx_charset = CHARSET_SPECIFIED; } else { t->tx_charset = CHARSET_UNSPECIFIED; @@ -1012,7 +1068,8 @@ InitMultiPart (CT ct) { int inout; long last, pos; - char *cp, *dp, **ap, **ep; + char *cp, *dp; + PM pm; char *bp, buffer[BUFSIZ]; struct multipart *m; struct k2v *kv; @@ -1057,15 +1114,15 @@ InitMultiPart (CT ct) * required for multipart messages. */ bp = 0; - for (ap = ci->ci_attrs, ep = ci->ci_values; *ap; ap++, ep++) { - if (!strcasecmp (*ap, "boundary")) { - bp = *ep; + for (pm = ci->ci_first_pm; pm; pm = pm->pm_next) { + if (!strcasecmp (pm->pm_name, "boundary")) { + bp = pm->pm_value; break; } } /* complain if boundary parameter is missing */ - if (!*ap) { + if (!pm) { advise (NULL, "a \"boundary\" parameter is mandatory for \"%s/%s\" type in message %s's %s: field", ci->ci_type, ci->ci_subtype, ct->c_file, TYPE_FIELD); @@ -1270,7 +1327,7 @@ InitMessage (CT ct) case MESSAGE_PARTIAL: { - char **ap, **ep; + PM pm; struct partial *p; if ((p = (struct partial *) calloc (1, sizeof(*p))) == NULL) @@ -1278,25 +1335,25 @@ InitMessage (CT ct) ct->c_ctparams = (void *) p; /* scan for parameters "id", "number", and "total" */ - for (ap = ci->ci_attrs, ep = ci->ci_values; *ap; ap++, ep++) { - if (!strcasecmp (*ap, "id")) { - p->pm_partid = add (*ep, NULL); + for (pm = ci->ci_first_pm; pm; pm = pm->pm_next) { + if (!strcasecmp (pm->pm_name, "id")) { + p->pm_partid = add (pm->pm_value, NULL); continue; } - if (!strcasecmp (*ap, "number")) { - if (sscanf (*ep, "%d", &p->pm_partno) != 1 + if (!strcasecmp (pm->pm_name, "number")) { + if (sscanf (pm->pm_value, "%d", &p->pm_partno) != 1 || p->pm_partno < 1) { invalid_param: advise (NULL, "invalid %s parameter for \"%s/%s\" type in message %s's %s field", - *ap, ci->ci_type, ci->ci_subtype, + pm->pm_name, ci->ci_type, ci->ci_subtype, ct->c_file, TYPE_FIELD); return NOTOK; } continue; } - if (!strcasecmp (*ap, "total")) { - if (sscanf (*ep, "%d", &p->pm_maxno) != 1 + if (!strcasecmp (pm->pm_name, "total")) { + if (sscanf (pm->pm_value, "%d", &p->pm_maxno) != 1 || p->pm_maxno < 1) goto invalid_param; continue; @@ -1411,21 +1468,21 @@ no_body: int params_external (CT ct, int composing) { - char **ap, **ep; + PM pm; struct exbody *e = (struct exbody *) ct->c_ctparams; CI ci = &ct->c_ctinfo; ct->c_ceopenfnx = NULL; - for (ap = ci->ci_attrs, ep = ci->ci_values; *ap; ap++, ep++) { - if (!strcasecmp (*ap, "access-type")) { + for (pm = ci->ci_first_pm; pm; pm = pm->pm_next) { + if (!strcasecmp (pm->pm_name, "access-type")) { struct str2init *s2i; CT p = e->eb_content; for (s2i = str2methods; s2i->si_key; s2i++) - if (!strcasecmp (*ep, s2i->si_key)) + if (!strcasecmp (pm->pm_value, s2i->si_key)) break; if (!s2i->si_key) { - e->eb_access = *ep; + e->eb_access = pm->pm_value; e->eb_flags = NOTOK; p->c_encoding = CE_EXTERNAL; continue; @@ -1439,46 +1496,46 @@ params_external (CT ct, int composing) return NOTOK; continue; } - if (!strcasecmp (*ap, "name")) { - e->eb_name = *ep; + if (!strcasecmp (pm->pm_name, "name")) { + e->eb_name = pm->pm_value; continue; } - if (!strcasecmp (*ap, "permission")) { - e->eb_permission = *ep; + if (!strcasecmp (pm->pm_name, "permission")) { + e->eb_permission = pm->pm_value; continue; } - if (!strcasecmp (*ap, "site")) { - e->eb_site = *ep; + if (!strcasecmp (pm->pm_name, "site")) { + e->eb_site = pm->pm_value; continue; } - if (!strcasecmp (*ap, "directory")) { - e->eb_dir = *ep; + if (!strcasecmp (pm->pm_name, "directory")) { + e->eb_dir = pm->pm_value; continue; } - if (!strcasecmp (*ap, "mode")) { - e->eb_mode = *ep; + if (!strcasecmp (pm->pm_name, "mode")) { + e->eb_mode = pm->pm_value; continue; } - if (!strcasecmp (*ap, "size")) { - sscanf (*ep, "%lu", &e->eb_size); + if (!strcasecmp (pm->pm_name, "size")) { + sscanf (pm->pm_value, "%lu", &e->eb_size); continue; } - if (!strcasecmp (*ap, "server")) { - e->eb_server = *ep; + if (!strcasecmp (pm->pm_name, "server")) { + e->eb_server = pm->pm_value; continue; } - if (!strcasecmp (*ap, "subject")) { - e->eb_subject = *ep; + if (!strcasecmp (pm->pm_name, "subject")) { + e->eb_subject = pm->pm_value; continue; } - if (!strcasecmp (*ap, "url")) { + if (!strcasecmp (pm->pm_name, "url")) { /* * According to RFC 2017, we have to remove all whitespace from * the URL */ - char *u, *p = *ep; - e->eb_url = u = mh_xmalloc(strlen(*ep) + 1); + char *u, *p = pm->pm_value; + e->eb_url = u = mh_xmalloc(strlen(pm->pm_value) + 1); for (; *p != '\0'; p++) { if (! isspace((unsigned char) *p)) @@ -1488,8 +1545,8 @@ params_external (CT ct, int composing) *u = '\0'; continue; } - if (composing && !strcasecmp (*ap, "body")) { - e->eb_body = getcpy (*ep); + if (composing && !strcasecmp (pm->pm_name, "body")) { + e->eb_body = getcpy (pm->pm_value); continue; } } @@ -1623,7 +1680,7 @@ static int openBase64 (CT ct, char **file) { int bitno, cc, digested; - int fd, len, skip, own_ct_fp = 0; + int fd, len, skip, own_ct_fp = 0, text = ct->c_type == CT_TEXT; uint32_t bits; unsigned char value, b; char *cp, *ep, buffer[BUFSIZ]; @@ -1646,12 +1703,6 @@ openBase64 (CT ct, char **file) } if (*file == NULL) { - char *tempfile; - if ((tempfile = m_mktemp2(NULL, invo_name, NULL, NULL)) == NULL) { - adios(NULL, "unable to create temporary file in %s", - get_temp_dir()); - } - ce->ce_file = add (tempfile, NULL); ce->ce_unlink = 1; } else { ce->ce_file = add (*file, NULL); @@ -1669,19 +1720,22 @@ openBase64 (CT ct, char **file) cp = context_find (buffer); } if (cp != NULL && *cp != '\0') { - if (ce->ce_unlink) { - /* Temporary file already exists, so we rename to - version with extension. */ - char *file_org = strdup(ce->ce_file); - ce->ce_file = add (cp, ce->ce_file); - if (rename(file_org, ce->ce_file)) { - adios (ce->ce_file, "unable to rename %s to ", file_org); - } - free(file_org); - - } else { - ce->ce_file = add (cp, ce->ce_file); - } + if (ce->ce_unlink) { + /* Create temporary file with filename extension. */ + if ((ce->ce_file = m_mktemps(invo_name, cp, NULL, NULL)) == NULL) { + adios(NULL, "unable to create temporary file in %s", + get_temp_dir()); + } + } else { + ce->ce_file = add (cp, ce->ce_file); + } + } else if (*file == NULL) { + char *tempfile; + if ((tempfile = m_mktemp2(NULL, invo_name, NULL, NULL)) == NULL) { + adios(NULL, "unable to create temporary file in %s", + get_temp_dir()); + } + ce->ce_file = add (tempfile, NULL); } if ((ce->ce_fp = fopen (ce->ce_file, "w+")) == NULL) { @@ -1745,17 +1799,20 @@ openBase64 (CT ct, char **file) test_end: if ((bitno -= 6) < 0) { b = (bits >> 16) & 0xff; - putc ((char) b, ce->ce_fp); + if (!text || b != '\r') + putc ((char) b, ce->ce_fp); if (digested) MD5Update (&mdContext, &b, 1); if (skip < 2) { b = (bits >> 8) & 0xff; - putc ((char) b, ce->ce_fp); + if (! text || b != '\r') + putc ((char) b, ce->ce_fp); if (digested) MD5Update (&mdContext, &b, 1); if (skip < 1) { b = bits & 0xff; - putc ((char) b, ce->ce_fp); + if (! text || b != '\r') + putc ((char) b, ce->ce_fp); if (digested) MD5Update (&mdContext, &b, 1); } @@ -1885,12 +1942,6 @@ openQuoted (CT ct, char **file) } if (*file == NULL) { - char *tempfile; - if ((tempfile = m_mktemp2(NULL, invo_name, NULL, NULL)) == NULL) { - adios(NULL, "unable to create temporary file in %s", - get_temp_dir()); - } - ce->ce_file = add (tempfile, NULL); ce->ce_unlink = 1; } else { ce->ce_file = add (*file, NULL); @@ -1908,19 +1959,22 @@ openQuoted (CT ct, char **file) cp = context_find (buffer); } if (cp != NULL && *cp != '\0') { - if (ce->ce_unlink) { - /* Temporary file already exists, so we rename to - version with extension. */ - char *file_org = strdup(ce->ce_file); - ce->ce_file = add (cp, ce->ce_file); - if (rename(file_org, ce->ce_file)) { - adios (ce->ce_file, "unable to rename %s to ", file_org); - } - free(file_org); - - } else { - ce->ce_file = add (cp, ce->ce_file); - } + if (ce->ce_unlink) { + /* Create temporary file with filename extension. */ + if ((ce->ce_file = m_mktemps(invo_name, cp, NULL, NULL)) == NULL) { + adios(NULL, "unable to create temporary file in %s", + get_temp_dir()); + } + } else { + ce->ce_file = add (cp, ce->ce_file); + } + } else if (*file == NULL) { + char *tempfile; + if ((tempfile = m_mktemp2(NULL, invo_name, NULL, NULL)) == NULL) { + adios(NULL, "unable to create temporary file in %s", + get_temp_dir()); + } + ce->ce_file = add (tempfile, NULL); } if ((ce->ce_fp = fopen (ce->ce_file, "w+")) == NULL) { @@ -2114,12 +2168,6 @@ open7Bit (CT ct, char **file) } if (*file == NULL) { - char *tempfile; - if ((tempfile = m_mktemp2(NULL, invo_name, NULL, NULL)) == NULL) { - adios(NULL, "unable to create temporary file in %s", - get_temp_dir()); - } - ce->ce_file = add (tempfile, NULL); ce->ce_unlink = 1; } else { ce->ce_file = add (*file, NULL); @@ -2137,19 +2185,22 @@ open7Bit (CT ct, char **file) cp = context_find (buffer); } if (cp != NULL && *cp != '\0') { - if (ce->ce_unlink) { - /* Temporary file already exists, so we rename to - version with extension. */ - char *file_org = strdup(ce->ce_file); - ce->ce_file = add (cp, ce->ce_file); - if (rename(file_org, ce->ce_file)) { - adios (ce->ce_file, "unable to rename %s to ", file_org); - } - free(file_org); - - } else { - ce->ce_file = add (cp, ce->ce_file); - } + if (ce->ce_unlink) { + /* Create temporary file with filename extension. */ + if ((ce->ce_file = m_mktemps(invo_name, cp, NULL, NULL)) == NULL) { + adios(NULL, "unable to create temporary file in %s", + get_temp_dir()); + } + } else { + ce->ce_file = add (cp, ce->ce_file); + } + } else if (*file == NULL) { + char *tempfile; + if ((tempfile = m_mktemp2(NULL, invo_name, NULL, NULL)) == NULL) { + adios(NULL, "unable to create temporary file in %s", + get_temp_dir()); + } + ce->ce_file = add (tempfile, NULL); } if ((ce->ce_fp = fopen (ce->ce_file, "w+")) == NULL) { @@ -2158,18 +2209,18 @@ open7Bit (CT ct, char **file) } if (ct->c_type == CT_MULTIPART) { - char **ap, **ep; + PM pm; CI ci = &ct->c_ctinfo; len = 0; fprintf (ce->ce_fp, "%s: %s/%s", TYPE_FIELD, ci->ci_type, ci->ci_subtype); len += strlen (TYPE_FIELD) + 2 + strlen (ci->ci_type) + 1 + strlen (ci->ci_subtype); - for (ap = ci->ci_attrs, ep = ci->ci_values; *ap; ap++, ep++) { + for (pm = ci->ci_first_pm; pm; pm = pm->pm_next) { putc (';', ce->ce_fp); len++; - snprintf (buffer, sizeof(buffer), "%s=\"%s\"", *ap, *ep); + snprintf (buffer, sizeof(buffer), "%s=\"%s\"", , *ep); if (len + 1 + (cc = strlen (buffer)) >= CPERLIN) { fputs ("\n\t", ce->ce_fp); @@ -2370,12 +2421,12 @@ openFile (CT ct, char **file) if (ferror (gp)) { admonish (ce->ce_file, "error reading"); - unlink (cachefile); + (void) m_unlink (cachefile); } else if (ferror (fp)) { admonish (cachefile, "error writing"); - unlink (cachefile); + (void) m_unlink (cachefile); } fclose (fp); } @@ -2586,12 +2637,12 @@ openFTP (CT ct, char **file) if (ferror (gp)) { admonish (ce->ce_file, "error reading"); - unlink (cachefile); + (void) m_unlink (cachefile); } else if (ferror (fp)) { admonish (cachefile, "error writing"); - unlink (cachefile); + (void) m_unlink (cachefile); } fclose (fp); } @@ -2861,7 +2912,7 @@ openURL (CT ct, char **file) if (ferror(gp)) { admonish(ce->ce_file, "error reading"); - unlink(cachefile); + (void) m_unlink (cachefile); } } umask(mask); @@ -3303,3 +3354,124 @@ bad_quote: *header_attrp = cp; return OK; } + +/* + * Create a string based on a list of output parameters. Assume that this + * parameter string will be appended to an existing header, so start out + * with the separator (;). Perform RFC 2231 encoding when necessary. + */ + +char * +output_params(size_t initialwidth, PM params) +{ + char *paramout = NULL; + char line[CPERLIN + 1], *q; + int curlen, index, eightbit; + size_t valoff; + + while (params != NULL) { + index = 0; + valoff = 0; + q = line; + + curlen = param_len(params, index, valoff, &eightbit); + + /* + * If this won't fit on the line, start a new one. Save room in + * case we need a semicolon on the end + */ + + if (initialwidth + curlen > CPERLIN - 1) { + paramout = add(";\n\t", paramout); + initialwidth = 8; + } + + /* + * Loop until we get a parameter that fits within a line. + */ + + while (initialwidth + curlen > CPERLIN - 1) { + int curvallen = strlen(params->pm_value + valoff) - + (initialwidth + curlen - (CPERLIN - 1)); + + /* + * curvallen holds how many characters we take from this + * current value. Make sure it's at least 1. + */ + + if (curvallen < 1) + curvallen = 1; + } + params = params->pm_next; + } + + return paramout; +} + +/* + * Calculate the size of a parameter. Include any necessary encoding. + * Start the length computation from where "offset" is marked. + */ + +static size_t +param_len(PM pm, int index, size_t valueoff, int *eightbit) +{ + char *start = pm->pm_value + valueoff, *p; + size_t len = 0; + + *eightbit = 0; + + /* + * Add up the length. First, start with the parameter name, and include + * the equal sign. + */ + + len += strlen(pm->pm_name) + 1; + + /* + * Scan the parameter value. If we find an 8-bit character, then + * we need to compute the locale name for the length. + */ + + for (p = start; *p != '\0'; p++) { + if (! isascii((unsigned char) *p)) { + *eightbit = 1; + break; + } + } + + /* + * If we've got 8-bit character, put the locale on the front (if we're + * doing part 0. Also compute the length of the string based on the + * encoding we need to do. + */ + + if (*eightbit) { + len++; /* For the encoding we need to do */ + if (index == 0) { + len += strlen(write_charset_8bit()) + 2; /* Plus extra '' */ + } else { + /* + * We know we definitely need to include an index. + * This will get the length wrong if we have more than 99 + * sections. I can live with that. + */ + len += 2; /* * */ + if (index > 9) + len++; + } + for (p = start; *p != '\0'; p++) { + if (isparamencode(*p)) + len += 3; + else + len++; + } + } else { + /* + * Much simpler! Don't forget opening and closing quotes. + */ + len += strlen(start) + 2; + } + + return len; +}