On 13 Mar 2000, at 11:54, Eli Zaretskii wrote:
> I'm not arguing that the current code in makeinfo is 100% fool-proof.
I'm sorry. I missed the shift. I assumed we were still talking about
makeinfo.
> I am arguing that if the read call reads only 100 characters, and if, by
> some magic, the programmer *knows* that will happen, it should be okay to
> ask for 200 bytes, and the read call should not bomb on us.
Well, you're getting into programming philosophy here. Personally, I'd
argue that my "magic" is more likely to be wrong than right, so I'd say
that it's good to check that the buffer I claim to be supplying is
accessible. Perhaps Cygnus could have been a bit more friendly by checking
only that part of the buffer needed by actual file data. But my viewpoint
is irrelevant here. ;-)
Thanks again for the explanation of your comment.
-- Dave Bryan
