Le 06/03/2013 16:32, Daniel Wagner a écrit :
Hi Tomasz,
On 03/06/2013 02:11 PM, Tomasz Bursztyka wrote:
The orignal code assumes that the builtin chain is empty which is
obviously always the right assumption :).
That is actually the bug in the original code: it might not be empty. If
it is, it does the right change, if not it modifies the 1 rule in the
chain.
I am looking at the current behaviour of __connman_iptables_append()
and I think it is broken. No patches applied nor have I have touched
this code in previous patches.
No shit sherlock, that's what I have told in response to your patch 7/9 ;-P
Look at iptables.c:
__connman_iptables_append() and __connman_iptables_insert()
Now go back to 1.11 before YOUR previous patchset:
look at iptables_append_rule() and iptables_insert_rule()
Tomasz
_______________________________________________
connman mailing list
connman@connman.net
http://lists.connman.net/listinfo/connman