On 16/09/2010 13:42, Heiko Oberdiek wrote:
> \outer\def \reserveinserts#1{%
>    \begingroup
>      \@tempcnta=\numexpr#1\relax
>      \ifnum\@tempcnta<\z@
>        \PackageWarning{etex}{%
>          Negative number (\the\@tempcnta) for %
>          \@backslashchar reserveinserts,\MessageBreak
>          assuming 0%
>        }%
>        \@tempcnta=\z@
>      \fi
>      \count@\count10 % count
>      \ifnum\count11>\count@ \count@\count11 \fi % dimen
>      \ifnum\count12>\count@ \count@\count12 \fi % skip
>      \ifnum\count14>\count@ \count@\count14 \fi % box
>      \count@=\numexpr\et@xins-\count@\relax
>      \PackageInfo{etex}{%
>        Maximum for \@backslashchar reserveinserts is \the\count@
>      }%
>      \ifnum\@tempcnta>\count@
>        \PackageError{etex}{%
>          \the\@tempcnta\space insertion register%
>          \ifnum\@tempcnta=\@ne \space is \else s are \fi
>          requested,\MessageBreak
>          \ifnum\count@<\@ne
>            but no more registers can be reserved%
>          \else
>            but only \the\count@\space register%
>            \ifnum\count@=\@ne\else s\fi\space can be reserved%
>          \fi
>        }\@ehc
>        \@tempcnta=\count@
>      \fi
>      \global\insc@unt\numexpr\et@xins-\@tempcnta\relax
>    \endgroup
> }

As Chris says, this seems more sensible than the current definition as 
it gives meaningful messages when it's not possible to reserve the space 
requested. Now, who's actually in charge of etex.sty?
-- 
Joseph Wright