Re: [PATCH Try#11 3/4] cfg80211: Radiotap parser

[Johannes Berg <johannes@sipsolutions.net>]

[-- Attachment #1: Type: text/plain, Size: 936 bytes --]

On Tue, 2007-06-12 at 22:42 -0700, Michael Wu wrote:
> On Tuesday 12 June 2007 06:04, andy@warmcat.com wrote:
> > +typedef enum  {
> > +       RADIOTAP_PARSER_OK = 0,
> > +       RADIOTAP_PARSER_DONE,
> > +       RADIOTAP_PARSER_INVALID
> > +} ieee80211_radiotap_parser_retcode_t;

> Yuck. I much prefer the standard error codes used in the previous version..

I would as well, but as I outlined before we basically have four things
that can happen
 * we found a next item we understood
 * we have a parser error
 * we reached the end
 * we found an item we can't support (and thus we don't support any
   further ones either assuming we add support sequentially)

I would like to be able to distinguish all these cases, the previous
version couldn't distinguish between the last two. However, I'm not
convinced that error codes are reasonable for something that isn't an
error (i.e. reaching the end)

johannes

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 190 bytes --]