In article <22348.932402582@coconut.itojun.org>,
	itojun@iijlab.net writes:
> Why don't we just allocate h_name separately?

Why should I? The library function is fine.

> Is there good reason for single malloc()?

Yes, it is faster.

It would be no big problem to modify the library function to align the
"h_addr_list" on a long word boundary, so we don't have to discuss how
do that here.

It is about how compatible our "libc" has to be with broken applications.

	Regards

-- 
Matthias Scheler                                http://home.owl.de/~tron/