Mercurial > pidgin.yaz
changeset 12434:1811afd29b57
[gaim-migrate @ 14741]
OSCAR, at least, is guilty of setting status messages on statuses
which do not accept a message with some regularity. I'm betting other
protocols have similar errors. This causes such errors to be logged
as an error to the debug window, so that they can be tracked down and
fixed.
In general, we seem to be abusing g_return_if_fail -- it's intended
to be an *assertion*, not a shortcut for normal return conditions.
I think it complains (or can complain) to the debug window itself ...
maybe we need to make it louder.
committer: Tailor Script <tailor@pidgin.im>
| author | Ethan Blanton <elb@pidgin.im> |
|---|---|
| date | Fri, 09 Dec 2005 05:07:10 +0000 |
| parents | 74d5a423433c |
| children | 8f98014a4e7d |
| files | src/status.c |
| diffstat | 1 files changed, 12 insertions(+), 1 deletions(-) [+] |
line wrap: on
line diff
--- a/src/status.c Fri Dec 09 05:05:06 2005 +0000 +++ b/src/status.c Fri Dec 09 05:07:10 2005 +0000 @@ -865,7 +865,18 @@ /* Make sure this attribute exists and is the correct type. */ attr_value = gaim_status_get_attr_value(status, id); - g_return_if_fail(attr_value != NULL); + /* This used to be g_return_if_fail, but it's failing a LOT, so + * let's generate a log error for now. */ + /* g_return_if_fail(attr_value != NULL); */ + if (attr_value == NULL) { + gaim_debug_error("status", + "Attempted to set status attribute '%s' for " + "status '%s', which is not legal. Fix " + "this!\n", id, + gaim_status_type_get_name(gaim_status_get_type(status))); + return; + + } g_return_if_fail(gaim_value_get_type(attr_value) == GAIM_TYPE_STRING); gaim_value_set_string(attr_value, value);