On Tue, Feb 03, 2026 at 12:45:27PM +0100, Theo Buehler wrote:
> On Tue, Feb 03, 2026 at 11:29:06AM +0100, Claudio Jeker wrote:
> > This is another diff that just moves bits around to simplify the big
> > filterset diff I have next.
> > 
> > Introduce rde_filter_copy() that takes care of copying a filter rule with
> 
> The signature feels more like rde_filter_dup(), although that would
> suggest you need to error check... Shrug.

You're right dup is probably the better term. The error checking is
handled similar to e.g. xstrdup in other programs. I don't think this
is a big issue here. Especially since the caller would probably fatal
anyway.

I adjusted the diff and used rde_filter_dup().
 
> > all depenencies to a new object. Also check that peer_apply_out_filter()
> > does not return an old list for new peers.
> 
> Fine with me. Unless I'm missing something that can't actually happen
> right now, but I assume this is bound to change.

Not really, I just dislike that the return value of
peer_apply_out_filter() is unused in that one call.
 
> > 
> > Move rde_l3vpn_import() to rde_filter.c since it works on a
> > struct filter_set to match against communities.
> > This is needed since the new code will use a different thing for sets.
> 
> ok tb
> 

-- 
:wq Claudio