>>>>> On Tue, 03 Jun 2003 18:59:54 +0900, itojun@iijlab.net said:

> 	which standard defines the behavior?  

There is no standard. Actually programs which use gethostent_r()
should die. 

> if there's no standard the programmmer who assumed solaris/glibc
> behavior is the one to be punished.

I don't think so. See below.

> you can't blame ISC (nor solaris/glibc).

Note that gethostent_r() has been already existed before the IRS
library.
If the IRS people need different semantics from Solaris/glibc one,
fine, then they should choose different name.
Intentionally defining different semantics for those functions
is definitely NIH. That's what we should avoid.
--
soda