From: Baptiste Daroussin <bapt@FreeBSD.org>
Date: Tue, 27 Aug 2024 09:18:48 +0000 (+0200)
Subject: mlmmj: add a parser to extract content-type's mime and boundary
X-Git-Tag: RELEASE_1_5_0~24
X-Git-Url: http://git.ipfire.org/gitweb.cgi?a=commitdiff_plain;h=f3f22ca867a1495a437736257ec07c8c565b1047;p=thirdparty%2Fmlmmj.git

mlmmj: add a parser to extract content-type's mime and boundary
---

diff --git a/include/mlmmj.h b/include/mlmmj.h
index a45b76ae..ca352c92 100644
--- a/include/mlmmj.h
+++ b/include/mlmmj.h
@@ -135,6 +135,8 @@ void ml_init(struct ml *ml);
 bool ml_open(struct ml *ml, bool checkuser);
 void ml_close(struct ml *ml);
 bool send_probe(struct ml *ml, const char *addr);
+void parse_content_type(strlist *, char **mime_type, char **boundary);
+char *find_in_list(strlist *, const char *pattern);
 
 #define MY_ASSERT(expression) if (!(expression)) { \
 			errno = 0; \
diff --git a/src/mlmmj.c b/src/mlmmj.c
index 6c2ce866..aa2cc275 100644
--- a/src/mlmmj.c
+++ b/src/mlmmj.c
@@ -30,6 +30,7 @@
 #include <fcntl.h>
 #include <stdio.h>
 #include <string.h>
+#include <strings.h>
 #include <time.h>
 #include <unistd.h>
 #include <err.h>
@@ -113,6 +114,68 @@ send_probe(struct ml *ml, const char *addr)
 	return (true);
 }
 
+char *
+find_in_list(strlist *l, const char *pattern)
+{
+	size_t len;
+
+	if (l == NULL)
+		return (NULL);
+	if (pattern == NULL)
+		return (NULL);
+	len = strlen(pattern);
+
+	tll_foreach(*l, el) {
+		char *str = el->item;
+		if (str == NULL)
+			continue;
+		if (strncasecmp(str, pattern, len) == 0)
+			return (str);
+	}
+	return (NULL);
+}
+
+void
+parse_content_type(strlist *allhdrs, char **mime_type, char **boundary)
+{
+	*boundary = NULL;
+	*mime_type = NULL;
+	char *pos;
+	size_t len = 0;
+
+	char *hdr = find_in_list(allhdrs, "Content-Type:");
+	if (hdr == NULL)
+		return; /* No Content-Type */
+	pos = hdr + 13;
+	while (*pos && (*pos == ' ' || *pos == '\t'))
+		++pos;
+	if (*pos == '"') {
+		++pos;
+		while (pos[len] && pos[len] != '"')
+			++len;
+	} else {
+		while (pos[len] && pos[len] != ' ' && pos[len] != '\t' && pos[len] != ';')
+			++len;
+	}
+	if (len > 0)
+		*mime_type = xstrndup(pos, len);
+	pos += len;
+	len = 0;
+	while (*pos && strncasecmp(pos, "boundary=", 9))
+		++pos;
+	if (*pos == '\0')
+		return; /* No boundaries */
+	pos += 9;
+	if (*pos == '"') {
+		++pos;
+		while (pos[len] && pos[len] != '"')
+			++len;
+	} else {
+		while(pos[len] && pos[len] != ' ' && pos[len] != '\t' && pos[len] != ';')
+			++len;
+	}
+	*boundary = xstrndup(pos, len);
+}
 
 bool
 parse_lastdigest(char *line, long *lastindex, time_t *lasttime,
diff --git a/tests/mlmmj.c b/tests/mlmmj.c
index 655b76e9..c7c43889 100644
--- a/tests/mlmmj.c
+++ b/tests/mlmmj.c
@@ -158,6 +158,8 @@ ATF_TC_WITHOUT_HEAD(gethdrline);
 ATF_TC_WITHOUT_HEAD(readlf);
 ATF_TC_WITHOUT_HEAD(mod_get_addr_type);
 ATF_TC_WITHOUT_HEAD(send_probe);
+ATF_TC_WITHOUT_HEAD(parse_content_type);
+ATF_TC_WITHOUT_HEAD(find_in_list);
 
 ATF_TC_BODY(random_int, tc)
 {
@@ -2861,6 +2863,66 @@ ATF_TC_BODY(send_probe, tc)
 	}
 }
 
+ATF_TC_BODY(find_in_list, tc)
+{
+	strlist bla = tll_init();
+	ATF_REQUIRE(find_in_list(NULL, NULL) == NULL);
+	ATF_REQUIRE(find_in_list(&bla, NULL) == NULL);
+	ATF_REQUIRE(find_in_list(&bla, "test") == NULL);
+	tll_push_back(bla, NULL);
+	ATF_REQUIRE(find_in_list(&bla, "test") == NULL);
+	tll_push_back(bla, "plop");
+	ATF_REQUIRE(find_in_list(&bla, "test") == NULL);
+	tll_push_back(bla, "test");
+	ATF_REQUIRE_STREQ(find_in_list(&bla, "test"), "test");
+	tll_pop_back(bla);
+	tll_push_back(bla, "test longer");
+	ATF_REQUIRE_STREQ(find_in_list(&bla, "test"), "test longer");
+}
+
+ATF_TC_BODY(parse_content_type, tc)
+{
+	char *m, *b;
+	strlist hdrs = tll_init();
+	parse_content_type(NULL, &m, &b);
+	ATF_REQUIRE(m == NULL);
+	ATF_REQUIRE(b == NULL);
+	tll_push_back(hdrs, "From: plop");
+	parse_content_type(&hdrs, &m, &b);
+	ATF_REQUIRE(m == NULL);
+	ATF_REQUIRE(b == NULL);
+	tll_push_back(hdrs, "Content-Type:");
+	parse_content_type(&hdrs, &m, &b);
+	ATF_REQUIRE(m == NULL);
+	ATF_REQUIRE(b == NULL);
+	tll_pop_back(hdrs);
+	tll_push_back(hdrs, "Content-Type: multipart/mixed");
+	parse_content_type(&hdrs, &m, &b);
+	ATF_REQUIRE_STREQ(m, "multipart/mixed");
+	free(m);
+	ATF_REQUIRE(b == NULL);
+	tll_pop_back(hdrs);
+	tll_push_back(hdrs, "Content-Type: multipart/mixed;");
+	parse_content_type(&hdrs, &m, &b);
+	ATF_REQUIRE_STREQ(m, "multipart/mixed");
+	free(m);
+	ATF_REQUIRE(b == NULL);
+	tll_pop_back(hdrs);
+	tll_push_back(hdrs, "Content-Type: multipart/mixed;\tboundary=bla");
+	parse_content_type(&hdrs, &m, &b);
+	ATF_REQUIRE_STREQ(m, "multipart/mixed");
+	free(m);
+	ATF_REQUIRE_STREQ(b,"bla");
+	free(b);
+	tll_pop_back(hdrs);
+	tll_push_back(hdrs, "Content-Type: multipart/mixed;\tboundary=bla;");
+	parse_content_type(&hdrs, &m, &b);
+	ATF_REQUIRE_STREQ(m, "multipart/mixed");
+	free(m);
+	ATF_REQUIRE_STREQ(b,"bla");
+	free(b);
+}
+
 ATF_TP_ADD_TCS(tp)
 {
 	ATF_TP_ADD_TC(tp, random_int);
@@ -2948,6 +3010,8 @@ ATF_TP_ADD_TCS(tp)
 	ATF_TP_ADD_TC(tp, readlf);
 	ATF_TP_ADD_TC(tp, mod_get_addr_type);
 	ATF_TP_ADD_TC(tp, send_probe);
+	ATF_TP_ADD_TC(tp, parse_content_type);
+	ATF_TP_ADD_TC(tp, find_in_list);
 
 	return (atf_no_error());
 }