Subject: Re: Hesiod uid maps
To: Greg Hudson <ghudson@MIT.EDU>
From: Michael Graff <>
List: tech-userlevel
Date: 03/07/1999 12:38:41
This quesiton is unrelated to your uid/passwd change.

Do you really want a function that is marked as "internal to..." to
have a weak alias?  I may not fully understand the purpose of this,
but I thoght it was to allow us to provide a function but allow it to
be overridden at link time.

If it is a purely internal function (not part of whatever API the
hesiod resolving code uses) should it be something other than

I guess the main question is when should there be a weak alias, and
when should there not be one?

Greg Hudson <ghudson@MIT.EDU> writes:

> *** /mit/netbsd/src/lib/libc/net/hesiod.c	Tue Mar  2 17:48:41 1999
> --- hesiod.c	Sat Mar  6 17:40:14 1999
> ***************
> *** 75,80 ****
> --- 75,81 ----
>   __weak_alias(hesiod_end,_hesiod_end);
>   __weak_alias(hesiod_to_bind,_hesiod_to_bind);
>   __weak_alias(hesiod_resolve,_hesiod_resolve);
> + __weak_alias(hesiod__uidresolve,_hesiod__uidresolve);
>   __weak_alias(hesiod_free_list,_hesiod_free_list);
>   __weak_alias(hes_init,_hes_init);
>   __weak_alias(hes_to_bind,_hes_to_bind);