Re: Issue 4212: fix out-of-bounds index in division_maior() (issue 189420043 by address@hidden)

[benko . pal]

On 2015/01/02 08:56:03, dak wrote:
> On 2015/01/02 07:21:19, benko.pal wrote:
> > On 2015/01/01 23:08:56, Dan Eble wrote:
> > > On 2015/01/01 22:57:31, benko.pal wrote:
> > > >
https://codereview.appspot.com/189420043/diff/1/lily/breathing-sign.cc
> > > > File lily/breathing-sign.cc (right):
> > > >
> > > >
> > >
> >

https://codereview.appspot.com/189420043/diff/1/lily/breathing-sign.cc#newcode122
> > lily/breathing-sign.cc:122: if (ydim[DOWN] < val && line_pos.begin
() < it - 1)
> > I'd rather write
> > line_pos.begin () + 1 < it
> > but good catch anyway, thanks!

> Actually, all of those checks (existing and proposed) look like
undefined
> behavior since they calculate a possibly non-existing iterator and
compare with
> it.  With most compilers and implementations things will probably
work, but one
> can't really rely on it.  In particular, the compiler is allowed to
make
> line_pos.begin () < whatever the same as line_pos.begin () !=
whatever.

it - 1 was used above as it[-1]; it exists, because ydim is a positive
range
(last if) strictly smaller than the whole staff (widening by negative
amount),
so it must be larger than begin, as asserted.

https://codereview.appspot.com/189420043/