[9fans] byte order problem in aan?

[9fans] byte order problem in aan?

[Axel Belinfante]

I'm currently playing with uaan(*), a unix port of aan(8)
and I think we have a byte-order problem when we read/write
the aan Hdr over the network. Is that right? What to do?

Axel. (who doesn't mind discussing this off-list if that's prefered)

(*) for those understanding dutch: should a windows port be waan? :-)

Re: [9fans] byte order problem in aan?

[Charles Forsyth]

[-- Attachment #1: Type: text/plain, Size: 88 bytes --]

do what aan ought to have done ...
pack and unpack the values using a fixed byte order.

[-- Attachment #2: Type: message/rfc822, Size: 3189 bytes --]

From: Axel Belinfante <Axel.Belinfante@cs.utwente.nl>
To: 9fans@cse.psu.edu
Subject: [9fans] byte order problem in aan?
Date: Wed, 11 Feb 2004 16:49:37 +0100
Message-ID: <200402111549.i1BFnbq16810@zamenhof.cs.utwente.nl>

I'm currently playing with uaan(*), a unix port of aan(8)
and I think we have a byte-order problem when we read/write
the aan Hdr over the network. Is that right? What to do?

Axel. (who doesn't mind discussing this off-list if that's prefered)

(*) for those understanding dutch: should a windows port be waan? :-)

Re: [9fans] byte order problem in aan?

[Peter Bosch]

> do what aan ought to have done ...
> pack and unpack the values using a fixed byte order.

yup, this need to be fixed.  i'll do it tonight...

peter.

Re: [9fans] byte order problem in aan?

[Axel Belinfante]

> > do what aan ought to have done ...
> > pack and unpack the values using a fixed byte order.
>
> yup, this need to be fixed.  i'll do it tonight...

thanks. main reason I was asking, was to have a coordinated fix
(to keep my experimental uaan in sync with sources)

Axel.

Re: [9fans] byte order problem in aan?

[Peter Bosch]

>> > do what aan ought to have done ...
>> > pack and unpack the values using a fixed byte order.
>>
>> yup, this need to be fixed.  i'll do it tonight...
>
> thanks. main reason I was asking, was to have a coordinated fix
> (to keep my experimental uaan in sync with sources)

i'll use network order inside the header data structure.  the reality
is that the current code is also compiler dependent.  here's my proposed
data message:

typedef struct {
	uchar		nb[4];	// Number of data bytes in this message
	uchar		msg[4];	// Message number
	uchar		acked[4];	// Number of messages acked
} Hdr;

Re: [9fans] byte order problem in aan?

[boyd, rounin]

> do what aan ought to have done ...
> pack and unpack the values using a fixed byte order.

yup

Re: [9fans] byte order problem in aan?

[Axel Belinfante]

> > do what aan ought to have done ...
> > pack and unpack the values using a fixed byte order.
>
> yup

probably I should not react to this (but write more clear).
anyway, as I tried to explain in a later message,
I thought it would make sense to coordinate the
'fixed byte order' with the one(s) fixing aan for sources.