On Fri, 19 Sep 2003 itojun@iijlab.net wrote:

> >It seems to me that we've introduced extra complexity for absolutely no
> >reason at all.  Because the program exits immediately in the failure
> >case (without jumping through a pile of clean-up functions), there's
> >no chance that an invalid buffer or buffersize will be used in further
> >code.
>
> 	if we leave bad practice in code people would cut-and-paste it and use
> 	it, causing bad mistakes.  so i am being pedantic.

Wouldn't it be sufficient to just add a comment saying that this behavior
is ok because we exit if there's a failure?

Take care,

Bill