On Monday, 21 February 2022 10:14:18 CET Sven Eckelmann wrote:
> On Sunday, 20 February 2022 19:36:14 CET Marek Lindner wrote:
> > diff --git a/alfred.h b/alfred.h
> > index 9ab92a2..598d38b 100644
> > --- a/alfred.h
> > +++ b/alfred.h
> > @@ -33,6 +33,9 @@
> > #define FIXED_TLV_LEN(__tlv_type) \
> > htons(sizeof(__tlv_type) - sizeof(__tlv_type.header))
> > 
> > +#define FIXED_TLV_LEN_PTR(__tlv_type) \
> > +       htons(sizeof(*__tlv_type) - sizeof(__tlv_type->header))
> > +
> 
> This doesn't apply on master. And also doesn't make a lot of sense, because
> you can now just use FIXED_TLV_LEN(*whatever). So I have dropped this part
> and applied the rest.

Your proposal did no compile and the master branch now also does not compile
on ipq40xx:

unix_sock.c: In function 'unix_sock_server_status':
alfred.h:34:46: error: 'status_op_mode' is a pointer; did you mean to use '->'?
  htons(sizeof(__tlv_type) - sizeof(__tlv_type.header))


> Btw. good that you've spotted some prints to stderr which should actually be
> to stdout. You only missed it for "mode: unknown". 

The alfred server running not in primary and not in secondary mode was stderr
worthy for me. But I also don't feel strongly about this.

Cheers,
Marek