Ensure fgets read at least one byte before modifying string

If a file has a \0 byte (binary file, strange encoding, corruption),
fgets() can return a string starting with a \0 byte - check for that
before checking to see if the byte before the \0 is a \n, so we don't
reach back before the start of our memory buffer.

Also check that the penultimate byte is a \n before we chop it off,
in case we're reading from a file missing a newline, or a line longer
than fit in the buffer provided to fgets().

Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>

1 files changed, 6 insertions, 1 deletions

diff --git a/man.c b/man.c
index e8432f0..b430060 100644
--- a/man.c
+++ b/man.c
@@ -290,7 +290,12 @@ ReadMandescFile(SectionList ** section_list, char *path)
     snprintf(mandesc_file, sizeof(mandesc_file), "%s/%s", path, MANDESC);
     if ((descfile = fopen(mandesc_file, "r")) != NULL) {
         while (fgets(string, BUFSIZ, descfile) != NULL) {
-            string[strlen(string) - 1] = '\0';  /* Strip off the CR. */
+            size_t len = strlen(string);
+
+            if (len == 0)
+                continue;
+            if (string[len - 1] == '\n')
+                string[len - 1] = '\0';  /* Strip off the CR. */
 
             if (streq(string, NO_SECTION_DEFAULTS)) {
                 use_defaults = FALSE;