error info handling in MIT krb5 code
Nicolas.Williams at sun.com
Wed Mar 29 18:46:01 EST 2006
On Wed, Mar 29, 2006 at 06:37:44PM -0500, Ken Raeburn wrote:
> On Mar 29, 2006, at 18:23, Nicolas Williams wrote:
> >Perhaps this is an argument for redesigning log files, facilities and
> >protocols, but that's way beyond what MIT can do in the scope of
> >krb5 :/
> >I'm sure :)
> Certainly for this next release! :-)
On the double!
> >The way we handle exceptional cases is this: the localization team
> >file bugs when they run into strings that are difficult to
> >localize, and
> >engineering will fix them.
> Nice... you have a localization team... :-)
We have, otherwise we couldn't have the business model we have.
> >Minimizing changes for vendors who've integrated MIT krb5 source and
> >done some localization work already would be nice though... :^)
> Well, I think we'd be happy to discuss how to get their changes back
> into our source tree...
What I meant is that finding a way to avoid peppering code with
gettext() calls would be good; see my comments about ssh in OpenSolaris.
More information about the krbdev