On Wed, Oct 23, 2024 at 01:03:10PM +0200, Phil Sutter wrote:
> On Tue, Oct 22, 2024 at 05:07:25PM +0200, Pablo Neira Ayuso wrote:
> > On Tue, Oct 22, 2024 at 04:55:33PM +0200, Phil Sutter wrote:
> > > On Tue, Oct 22, 2024 at 03:48:12PM +0200, Pablo Neira Ayuso wrote:
> > > > On Tue, Oct 22, 2024 at 03:08:01PM +0200, Phil Sutter wrote:
> > > > > On Tue, Oct 22, 2024 at 02:30:58PM +0200, Phil Sutter wrote:
> > > > > [...]
> > > > > > - With your patch applied, 20 rules fail (in both variants). Is this
> > > > > > expected or a bug on my side?
> > > > >
> > > > > OK, so most failures are caused by my test kernel not having
> > > > > CONFIG_IP_VS_IPV6 enabled.
> > > > >
> > > > > Apart from that, there is a minor bug in introduced libip6t_recent.t in
> > > > > that it undoes commit d859b91e6f3ed ("extensions: recent: New kernels
> > > > > support 999 hits") by accident. More interesting though, it's reported
> > > > > twice, once for fast mode and once for normal mode. I'll see how I can
> > > > > turn off error reporting in fast mode, failing tests are repeated
> > > > > anyway.
> > > >
> > > > Would you point me to the relevant line in the libip6t_recent.t?
> > >
> > > It is in line 7, I had changed the supposed-to-fail --hitcount value of
> > > 999 to 65536.
> >
> > This was already fixed in v2, correct?
>
> Ah, you're right. I didn't notice your v2.
>
> If you're OK with it, I'll apply your v3 with the following changes:
> - Describe 'iptables' param in _run_test_file()
> - Drop duplicate 'endswith' test from _run_test_file()
> - Print results with command name suffixed for libxt tests (it is more
> consistent wrt. tests count)
Patch applied with mentioned changes. Thanks!
