Re: Add tab-tie-follow-engraver (issue2723043)

[n . puttock]

On 2010/11/12 04:12:48, Carl wrote:
> Here's a new version of the patch that tries to eliminate any scheme
calls, and
> demonstrates my problem.

> I need to do a check on the left bound to see if it's a grob, because
if I take
> the check out I get a Bus error.

> I don't know how to do such a check from C++ -- I think it's done by
casting,
> but I'm not sure.

Of course you know.  It's a pointer, so you check whether it's NULL.

if (left_bound)

> If I self-scm() the C++ Item * variable, then use the scheme routine
ly:grob? to
> check if it's a grob, if I get a Bus error.

The slur appears to be acknowledged twice, and the first time it has no
left bound (which implies it would be more efficient and safer to do
this checking in stop_translation_timestep ())

> As it stands, the code works on my machine.

It's back to square one though, isn't it?

The triple nested loop is horrible, and should be avoided at all costs;
this is partly why I suggested only acknowledging the ties and heads.

Cheers,
Neil

http://codereview.appspot.com/2723043/