The tool is supposed to work line-based, so a) the newline character is
implicit and b) imho, does not belong to the string which is then further
processed.
Also, this allows '$' to match the end of a line when applying a regex to the
line, no matter if the REG_NEWLINE flag has been used when compiling the regex
or not.
while (42)
{
+ size_t len;
+
status = cu_tail_readline (obj, buf, buflen);
if (status != 0)
{
if (buf[0] == 0)
break;
+ len = strlen (buf);
+ while (len > 0) {
+ if (buf[len - 1] != '\n')
+ break;
+ buf[len - 1] = '\0';
+ }
+
status = callback (data, buf, buflen);
if (status != 0)
{