Re: [Patch 1/1] CLD: Introduce the "New CLD" API
Pete Zaitcev <[email protected]>
| Newsgroups | org.kernel.vger.hail-devel |
|---|---|
| Message-ID | <[email protected]> |
On Mon, 08 Feb 2010 23:55:03 -0500 Jeff Garzik <[email protected]> wrote: > >>> - if (cldc_getaddr(&host_list, hostb, NULL)) > >>> + if (cldc_getaddr(&host_list, hostb,&ncld_log)) > >>> return 1001; > > > >> Logging pointer should be supplied by the caller... that is the larger > >> bug. Needing a no-op log function is just an indication of that. > > > > I did it before, but then every client, without exception, needed > > to include<stdarg.h> and provide the log. And what for? This is > > STUPID and proof is in the pudding: your getsrvinfo does not do > > anything of the sort. > > Possibly true -- but it is worse to introduce internal inconsistencies. > It is far better to remove debugging statements completely than to > create inconsistent usage. Hmm, you may be right about that. > > Might as well replace cldc_getadd with getsrvinfo > > and then everyone becomes happy. > > Seems a good fit :) How far along were you with getsrvinfo? I only have the header that you sent to the list just before Christmas. -- Pete