}
if (opp == NULL) {
- /* Check for duplicated profile entries. But only on this
- very first call from context_read(), when opp is NULL. */
+ /* Check for duplicated non-null profile entries. Also ignore
+ profile entries that are just "#", because that's what the
+ mh-profile man page suggests using for comments. Only do
+ this check on the very first call from context_read(), when
+ opp is
+ NULL. */
for (np = m_defs; np; np = np->n_next) {
/* Yes, this is O(N^2). The profile should be small enough so
that's not a performance problem. */
- struct node *np2;
- for (np2 = np->n_next; np2; np2 = np2->n_next) {
- if (!mh_strcasecmp (np->n_name, np2->n_name)) {
- admonish (NULL, "multiple \"%s\" profile components "
- "in %s, ignoring \"%s\"",
- np->n_name, defpath, np2->n_field);
+ if (strlen (np->n_name) > 0 && strcmp ("#", np->n_name)) {
+ struct node *np2;
+ for (np2 = np->n_next; np2; np2 = np2->n_next) {
+ if (! mh_strcasecmp (np->n_name, np2->n_name)) {
+ printf ("%d\n", strlen (np->n_name));
+ admonish (NULL, "multiple \"%s\" profile components "
+ "in %s, ignoring \"%s\"",
+ np->n_name, defpath, np2->n_field);
+ }
}
}
}