Uwe Brauer writes:
In any case what is the reason that our version is so complicated?
It's not complicated. The computation is quite straightforward. The
reason for the additional computation is that our version accepts an
additional argument that does something useful, and GNU's doesn't. I
use the feature occasionally in TTYs.
Maybe it is from a time, when Xemacs had difficulties with large
No. The comment about overflow is simply saying that it's worth being
careful to do the computation (N * (BUFFERSIZE / 10)) rather than
((N * BUFFERSIZE) / 10), because the multiplication in the latter can
overflow and the format can't.
XEmacs-Beta mailing list