On Mar 04, 2005, at 14:52, S�ren Hansen wrote:
fre, 04 03 2005 kl. 13:23 +0300, skrev Paul Bagyenda:- the move you suggested, of list_destroy(lto,NULL) in mmsfromemail.c is potentially dodgy since that variable is reset in the meantime. So in some circumstances, a leak could have resulted. I have made some changes around that.
I just took at second look... If you believe that lto can be reset in the meantime then the call to mms_log should definitely be moved, too.
No, you had put it in the right place (save for the small problem). Before it was being destroyed before use. Never wise.
- In mmsproxy.c, the change you had suggested on when to send DLR to
the sender pointed out a bug, but potentially introduced another:
According to spec, we must send DLR if the recipient inserts the
X-Mms-Report-Allowed header with a value of "Yes" *or* if they do not
insert the header (i.e. default is "Yes"). (Of course we must also only
send DLR if the sender asked for it.) Introduced what I hope is the
right logic.
Ok, I missed that one in the spec. So now it's something like ( ( (allow_report == NULL) || (octstr_case_compare(allow_report, octstr_imm("Yes"))) ) && e->dlr )
Right?
Yes.
-- S�ren Hansen <[EMAIL PROTECTED]> _______________________________________________ Devel mailing list [email protected] http://mbuni.org/mailman/listinfo/devel_mbuni.org
_______________________________________________ Devel mailing list [email protected] http://mbuni.org/mailman/listinfo/devel_mbuni.org
