Hi Daniel,
Hi Tomasz,
What do you think about this? Yes I agree the __connman_log_init()
is dirty but we can't really add '-d' because that belongs also
to iptables command set. And enabling __connman_log_init() always
is bit too much in my opinion.
Looks fine to me, iptables-test is a test tool anyway.
I just tried, sure the output is a bit cryptic and verbose compared to
real iptables but I think it's fine.
It should help actually to figure out some bugs, if any, in src/iptables.c
And anyway, when testing new rules addition/deletion, dev should always
check via the real iptables if everything is correct.
Just 2 comments below:
diff --git a/tools/iptables-test.c b/tools/iptables-test.c
index 95d0af4..bef9251 100644
--- a/tools/iptables-test.c
+++ b/tools/iptables-test.c
@@ -34,6 +34,7 @@ enum iptables_command {
IPTABLES_COMMAND_POLICY,
IPTABLES_COMMAND_CHAIN_INSERT,
IPTABLES_COMMAND_CHAIN_DELETE,
+ IPTABLES_COMMAND_DUMP,
IPTABLES_COMMAND_UNKNOWN,
};
@@ -42,11 +43,12 @@ int main(int argc, char *argv[])
enum iptables_command cmd = IPTABLES_COMMAND_UNKNOWN;
char *table = NULL, *chain = NULL, *rule = NULL, *tmp;
Fix table so it gets "filter" as a default value. If you run
./iptables-test without -t it should imply that "fitler" table is used.
(like with real iptables)
Currently, not using -t makes it segfault on commit.
int err, c, i;
+ char *debug;
Unused so you can nuke it.
Merge this patch + my comments to your patch 4 and you get my ACK.
Cheers,
Tomasz
_______________________________________________
connman mailing list
[email protected]
http://lists.connman.net/listinfo/connman
