From 069cb5c3a525ebcc19cc2927964258acaca87b13 Mon Sep 17 00:00:00 2001 From: Michael Biebl Date: Sun, 17 Jun 2018 14:54:10 +0200 Subject: New upstream version 1.11.90 --- shared/nm-utils/nm-enum-utils.c | 9 +++++++++ 1 file changed, 9 insertions(+) (limited to 'shared/nm-utils/nm-enum-utils.c') diff --git a/shared/nm-utils/nm-enum-utils.c b/shared/nm-utils/nm-enum-utils.c index 023f8385..b83c4207 100644 --- a/shared/nm-utils/nm-enum-utils.c +++ b/shared/nm-utils/nm-enum-utils.c @@ -64,9 +64,18 @@ _ASSERT_enum_values_info (GType type, * And then, when actually running against a newer library version where * @type knows the nick, we have this situation. * + * Another reason for specifying a nick both in @value_infos and @type, + * is to specify an alias which is not used with highest preference. For + * example, if you add an alias "disabled" for "none" (both numerically + * equal), then the first alias in @value_infos will be preferred over + * the name from @type. So, to still use "none" as preferred name, you may + * explicitly specify the "none" alias in @value_infos before "disabled". + * * However, what never is allowed, is to use a name (nick) to re-number * the value. That is, if both @value_infos and @type contain a particular * nick, their numeric values must agree as well. + * Allowing this, would be very confusing, because the name would have a different + * value from the regular GLib GEnum API. */ g_assert (enum_value->value == value_infos->value); } -- cgit 1.3.0-6-gf8a5