X-Git-Url: https://diplodocus.org/git/nmh/blobdiff_plain/c982212dc21459effbb206a7de2f6306f68bdefe..d7e260c9a302e465a2c40e75ff8bd506768a7dbf:/uip/mhparse.c

diff --git a/uip/mhparse.c b/uip/mhparse.c
index f2f3a246..444bf7d5 100644
--- a/uip/mhparse.c
+++ b/uip/mhparse.c
@@ -23,8 +23,6 @@
 
 extern int debugsw;
 
-extern pid_t xpid;	/* mhshowsbr.c  */
-
 /* cache policies */
 extern int rcachesw;	/* mhcachesbr.c */
 extern int wcachesw;	/* mhcachesbr.c */
@@ -183,19 +181,6 @@ struct str2init str2methods[] = {
 };
 
 
-int
-pidcheck (int status)
-{
-    if ((status & 0xff00) == 0xff00 || (status & 0x007f) != SIGQUIT)
-	return status;
-
-    fflush (stdout);
-    fflush (stderr);
-    done (1);
-    return 1;
-}
-
-
 /*
  * Main entry point for parsing a MIME message or file.
  * It returns the Content structure for the top level
@@ -838,7 +823,7 @@ magic_skip:
             if (ct->c_dispo_type &&
 		!get_param(ct->c_dispo_first, "filename", '_', 1)) {
 		add_param(&ct->c_dispo_first, &ct->c_dispo_last, "filename",
-			  r1bindex(ci->ci_magic, '/'));
+			  r1bindex(ci->ci_magic, '/'), 0);
 	    }
         }
 	else
@@ -2478,13 +2463,6 @@ openFTP (CT ct, char **file)
 	return NOTOK;
     }
 
-    if (xpid) {
-	if (xpid < 0)
-	    xpid = -xpid;
-	pidcheck (pidwait (xpid, NOTOK));
-	xpid = 0;
-    }
-
     /* Get the buffer ready to go */
     bp = buffer;
     buflen = sizeof(buffer);
@@ -2684,13 +2662,6 @@ openMail (CT ct, char **file)
 	return NOTOK;
     }
 
-    if (xpid) {
-	if (xpid < 0)
-	    xpid = -xpid;
-	pidcheck (pidwait (xpid, NOTOK));
-	xpid = 0;
-    }
-
     /* Get buffer ready to go */
     bp = buffer;
     buflen = sizeof(buffer);
@@ -2821,13 +2792,6 @@ openURL (CT ct, char **file)
 	return NOTOK;
     }
 
-    if (xpid) {
-	if (xpid < 0)
-	    xpid = -xpid;
-	pidcheck (pidwait (xpid, NOTOK));
-	xpid = 0;
-    }
-
     ce->ce_unlink = (*file == NULL);
     caching = 0;
     cachefile[0] = '\0';
@@ -3255,19 +3219,20 @@ parse_header_attrs (const char *filename, const char *fieldname,
     struct sectlist {
 	char *value;
 	int index;
+	int len;
 	struct sectlist *next;
-    } *sp;
+    } *sp, *sp2;
     struct parmlist {
 	char *name;
 	char *charset;
 	char *lang;
 	struct sectlist *sechead;
 	struct parmlist *next;
-    } *pp, *phead = NULL;
+    } *pp, *pp2, *phead = NULL;
 
     while (*cp == ';') {
-	char *dp, *vp, *up, c, *nameptr, *valptr, *charset, *lang;
-	int encoded = 0, partial = 0, index = 0, len = 0;
+	char *dp, *vp, *up, *nameptr, *valptr, *charset = NULL, *lang = NULL;
+	int encoded = 0, partial = 0, len = 0, index = 0;
 
 	cp++;
 	while (isspace ((unsigned char) *cp))
@@ -3313,15 +3278,15 @@ parse_header_attrs (const char *filename, const char *fieldname,
 	 * If there's a * and one or more digits, then it's section N.
 	 *
 	 * Remember we can have one or the other, or both.  cp points to
-	 * beginning of name, up points to the last character in the
+	 * beginning of name, up points past the last character in the
 	 * parameter name.
 	 */
 
-	for (vp = cp; vp <= up; vp++) {
-	    if (*vp == '*' && vp < up) {
+	for (vp = cp; vp < up; vp++) {
+	    if (*vp == '*' && vp < up - 1) {
 		partial = 1;
 		continue;
-	    } else if (*vp == '*' && vp == up) {
+	    } else if (*vp == '*' && vp == up - 1) {
 	    	encoded = 1;
 	    } else if (partial) {
 		if (isdigit((unsigned char) *vp))
@@ -3357,7 +3322,8 @@ parse_header_attrs (const char *filename, const char *fieldname,
 	     */
 	    if (index == 0) {
 	    	vp = dp;
-		while (*vp != '\'' && !isspace((unsigned char) *vp))
+		while (*vp != '\'' && !isspace((unsigned char) *vp) &&
+							*vp != '\0')
 		    vp++;
 		if (*vp == '\'') {
 		    if (vp != dp) {
@@ -3368,36 +3334,41 @@ parse_header_attrs (const char *filename, const char *fieldname,
 		    } else {
 			charset = NULL;
 		    }
+		    vp++;
 		} else {
-		    advise(NULL, "missing charset in message %s's %s:
+		    advise(NULL, "missing charset in message %s's %s: "
 			   "field\n%*s(parameter %s)", filename, fieldname,
 			   strlen(invo_name) + 2, "", nameptr);
 		    free(nameptr);
 		    return NOTOK;
 		}
-	    }
+		dp = vp;
 
-	    vp = (dp = vp);
-	    while (*vp != '\'' && !isspace((unsigned char) *vp))
-		vp++;
+		while (*vp != '\'' && !isspace((unsigned char) *vp) &&
+							*vp != '\0')
+		    vp++;
 
-	    if (*vp == '\'') {
-		if (vp != dp) {
-		    len = vp - dp;
-		    lang = mh_xmalloc(len + 1);
-		    strncpy(lang, dp, len);
-		    lang[len] = '\0';
+		if (*vp == '\'') {
+		    if (vp != dp) {
+			len = vp - dp;
+			lang = mh_xmalloc(len + 1);
+			strncpy(lang, dp, len);
+			lang[len] = '\0';
+		    } else {
+			lang = NULL;
+		    }
+		    vp++;
 		} else {
-		    charset = NULL;
+		    advise(NULL, "missing language tag in message %s's %s: "
+			   "field\n%*s(parameter %s)", filename, fieldname,
+			   strlen(invo_name) + 2, "", nameptr);
+		    free(nameptr);
+		    if (charset)
+			free(charset);
+		    return NOTOK;
 		}
-	    } else {
-		advise(NULL, "missing language tag in message %s's %s: "
-		       "field\n%*s(parameter %s)", filename, fieldname,
-		       strlen(invo_name) + 2, "", nameptr);
-		free(nameptr);
-		if (charset)
-		     free(charset);
-		return NOTOK;
+
+		dp = vp;
 	    }
 
 	    /*
@@ -3407,8 +3378,7 @@ parse_header_attrs (const char *filename, const char *fieldname,
 	     * length so we can allocate the correct buffer size.
 	     */
 
-	    for (dp = vp, len = 0; *vp != '\0' && !isspace((unsigned char *vp);
-	    							vp++) {
+	    for (vp = dp, len = 0; istoken(*vp); vp++) {
 		if (*vp == '%') {
 		     if (*(vp + 1) == '\0' ||
 				!isxdigit((unsigned char) *(vp + 1)) ||
@@ -3432,8 +3402,8 @@ parse_header_attrs (const char *filename, const char *fieldname,
 
 	    up = valptr = mh_xmalloc(len + 1);
 
-	    for (vp = dp; *vp != '\0' && !isspace((unsigned char *vp); vp++) {
-		if (*vp = '%') {
+	    for (vp = dp; istoken(*vp); vp++) {
+		if (*vp == '%') {
 		    *up++ = decode_qp(*(vp + 1), *(vp + 2));
 		    vp += 2;
 		} else {
@@ -3442,6 +3412,7 @@ parse_header_attrs (const char *filename, const char *fieldname,
 	    }
 
 	    *up = '\0';
+	    cp = vp;
 	} else {
 	    /*
 	     * A "normal" string.  If it's got a leading quote, then we
@@ -3451,13 +3422,12 @@ parse_header_attrs (const char *filename, const char *fieldname,
 	     * buffer.
 	     */
 
-	    dp = vp;
 	    len = 0;
 
 	    if (*dp == '"') {
 		for (cp = dp + 1;;) {
 		    switch (*cp++) {
-		    case '\0:
+		    case '\0':
 bad_quote:
 		        advise (NULL,
 				"invalid quoted-string in message %s's %s: "
@@ -3484,9 +3454,30 @@ bad_quote:
 		    break;
 		}
 
-		valptr = mh_xmalloc(len + 1);
-	    } else 
+	    } else {
+		for (cp = dp; istoken (*cp); cp++) {
+		    len++;
+		}
+	    }
 
+	    valptr = mh_xmalloc(len + 1);
+
+	    if (*dp == '"') {
+	    	int i;
+		for (cp = dp + 1, vp = valptr, i = 0; i < len; i++) {
+		    if (*cp == '\\') {
+			cp++;
+		    }
+		    *vp++ = *cp++;
+		}
+		cp++;
+	    } else {
+	    	strncpy(valptr, cp = dp, len);
+		cp += len;
+	    }
+
+	    valptr[len] = '\0';
+	}
 
 	/*
 	 * If 'partial' is set, we don't allocate a parameter now.  We
@@ -3506,64 +3497,66 @@ bad_quote:
 	    if (pp == NULL) {
 		pp = mh_xmalloc(sizeof(*pp));
 		memset(pp, 0, sizeof(*pp));
+		pp->name = nameptr;
 		pp->next = phead;
 		phead = pp;
 	    }
 
-	    if (index == 0 && encoded) {
-
-	pm = mh_xmalloc(sizeof(*pm));
-	memset(pm, 0, sizeof(*pm));
-
-	/* This is all mega-bozo and needs cleanup */
-	vp = (pm->pm_name = add (cp, NULL)) + (up - cp);
-	*vp = '\0';
-
-	/* Now store the attribute value. */
+	    /*
+	     * Insert this into the section linked list
+	     */
 
-	vp = pm->pm_name + (dp - cp);
+	    sp = mh_xmalloc(sizeof(*sp));
+	    memset(sp, 0, sizeof(*sp));
+	    sp->value = valptr;
+	    sp->index = index;
+	    sp->len = len;
 
-	if (*dp == '"') {
-	    for (cp = ++dp, dp = vp;;) {
-		switch (*cp++) {
-		    case '\0':
-bad_quote:
-		        advise (NULL,
-				"invalid quoted-string in message %s's %s: "
-                                "field\n%*s(parameter %s)",
+	    if (pp->sechead == NULL || pp->sechead->index > index) {
+		sp->next = pp->sechead;
+		pp->sechead = sp;
+	    } else {
+		for (sp2 = pp->sechead; sp2 != NULL; sp2 = sp2->next) {
+		    if (sp2->index == sp->index) {
+			advise (NULL, "duplicate index (%d) in message "
+				"%s's %s: field\n%*s(parameter %s)", sp->index,
 				filename, fieldname, strlen(invo_name) + 2, "",
-				pm->pm_name);
+				nameptr);
 			return NOTOK;
+		    }
+		    if (sp2->index < sp->index &&
+			(sp2->next == NULL || sp2->next->index > sp->index)) {
+			sp->next = sp2->next;
+			sp2->next = sp;
+			break;
+		    }
+		}
 
-		    case '\\':
-			*dp++ = c;
-			if ((c = *cp++) == '\0')
-			    goto bad_quote;
-			/* else fall... */
+		if (sp2 == NULL) {
+		    advise(NULL, "Internal error: cannot insert partial "
+		    	   "param in message %s's %s: field\n%*s(parameter %s)",
+			   filename, fieldname, strlen(invo_name) + 2, "",
+			   nameptr);
+		    return NOTOK;
+		}
+	    }
 
-		    default:
-			*dp++ = c;
-			continue;
+	    /*
+	     * Save our charset and lang tags.
+	     */
 
-		    case '"':
-			*dp = '\0';
-			break;
-		}
-		break;
+	    if (index == 0 && encoded) {
+		if (pp->charset)
+		    free(pp->charset);
+	    	pp->charset = charset;
+		if (pp->lang)
+		    free(pp->lang);
+		pp->lang = lang;
 	    }
 	} else {
-	    for (cp = dp, dp = vp; istoken (*cp); cp++, dp++)
-		continue;
-	    *dp = '\0';
-	}
-	pm->pm_value = getcpy(vp); 
-	if (!*vp) {
-	    advise (NULL,
-		    "invalid parameter in message %s's %s: "
-                    "field\n%*s(parameter %s)",
-		    filename, fieldname, strlen(invo_name) + 2, "",
-		    pm->pm_name);
-	    return NOTOK;
+	    pm = add_param(param_head, param_tail, nameptr, valptr, 1);
+	    pm->pm_charset = charset;
+	    pm->pm_lang = lang;
 	}
 
 	while (isspace ((unsigned char) *cp))
@@ -3573,20 +3566,70 @@ bad_quote:
             get_comment (filename, fieldname, &cp, commentp) == NOTOK) {
 	    return NOTOK;
         }
+    }
 
-	if (*param_head == NULL) {
-	    *param_head = pm;
-	    *param_tail = pm;
-	} else {
-	    (*param_tail)->pm_next = pm;
-	    *param_tail = pm;
+    /*
+     * Now that we're done, reassemble all of the partial parameters.
+     */
+
+    for (pp = phead; pp != NULL; ) {
+    	char *p, *q;
+	size_t tlen = 0;
+	int pindex = 0;
+	for (sp = pp->sechead; sp != NULL; sp = sp->next) {
+	    if (sp->index != pindex++) {
+		advise(NULL, "missing section %d for parameter in "
+		       "message %s's %s: field\n%*s(parameter %s)", pindex - 1,
+		       filename, fieldname, strlen(invo_name) + 2, "",
+		       pp->name);
+		return NOTOK;
+	    }
+	    tlen += sp->len;
 	}
+
+	p = q = mh_xmalloc(tlen + 1);
+	for (sp = pp->sechead; sp != NULL; ) {
+	    memcpy(q, sp->value, sp->len);
+	    q += sp->len;
+	    free(sp->value);
+	    sp2 = sp->next;
+	    free(sp);
+	    sp = sp2;
+	}
+
+	p[tlen] = '\0';
+
+	pm = add_param(param_head, param_tail, pp->name, p, 1);
+	pm->pm_charset = pp->charset;
+	pm->pm_lang = pp->lang;
+	pp2 = pp->next;
+	free(pp);
+	pp = pp2;
     }
 
     *header_attrp = cp;
     return OK;
 }
 
+/*
+ * Return the charset for a particular content type.  Return pointer is
+ * only valid until the next call to content_charset().
+ */
+
+char *
+content_charset (CT ct) {
+    static char *ret_charset = NULL;
+
+    if (ret_charset != NULL) {
+	free(ret_charset);
+    }
+
+    ret_charset = get_param(ct->c_ctinfo.ci_first_pm, "charset", '?', 0);
+
+    return ret_charset ? ret_charset : "US-ASCII";
+}
+
+
 /*
  * 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
@@ -3994,14 +4037,14 @@ normal_param(PM pm, char *output, size_t len, size_t valuelen,
  */
 
 PM
-add_param(PM *first, PM *last, const char *name, const char *value)
+add_param(PM *first, PM *last, char *name, char *value, int nocopy)
 {
     PM pm = mh_xmalloc(sizeof(*pm));
 
     memset(pm, 0, sizeof(*pm));
 
-    pm->pm_name = getcpy(name);
-    pm->pm_value = getcpy(value);
+    pm->pm_name = nocopy ? name : getcpy(name);
+    pm->pm_value = nocopy ? value : getcpy(value);
 
     if (*first) {
 	(*last)->pm_next = pm;
@@ -4014,6 +4057,33 @@ add_param(PM *first, PM *last, const char *name, const char *value)
     return pm;
 }
 
+/*
+ * Either replace a current parameter with a new value, or add the parameter
+ * to the parameter linked list.
+ */
+
+PM
+replace_param(PM *first, PM *last, char *name, char *value, int nocopy)
+{
+    PM pm;
+
+    for (pm = *first; pm != NULL; pm = pm->pm_next) {
+	if (strcasecmp(name, pm->pm_name) == 0) {
+	    /*
+	     * If nocopy is set, it's assumed that we own both name
+	     * and value.  We don't need name, so we discard it now.
+	     */
+	    if (nocopy)
+		free(name);
+	    free(pm->pm_value);
+	    pm->pm_value = nocopy ? value : getcpy(value);
+	    return pm;
+	}
+    }
+
+    return add_param(first, last, name, value, nocopy);
+}
+
 /*
  * Retrieve a parameter value from a parameter linked list.  If the parameter
  * value needs converted to the local character set, do that now.
@@ -4049,7 +4119,10 @@ char *get_param_value(PM pm, char replace)
     int utf8;
     iconv_t cd;
     ICONV_CONST char *p;
+#else /* HAVE_ICONV */
+    char *p;
 #endif /* HAVE_ICONV */
+
     char *q;
 
     /*
@@ -4124,9 +4197,10 @@ char *get_param_value(PM pm, char replace)
     *q = '\0';
 
     return buffer;
-#endif /* HAVE_ICONV */
 
 noiconv:
+#endif /* HAVE_ICONV */
+
     /*
      * Take everything non-ASCII and substituite the replacement character
      */