Re: Issue 787 in lilypond: script stack order fails with 2 more notes

[Carl Sorensen]

On 11/30/09 2:09 PM, "address@hidden" <address@hidden>
wrote:

> Updates:
>         Status: Started
>         Labels: -fixed_2_13_9
> 
> Comment #6 on issue 787 by n.puttock: script stack order fails with 2 more
> notes
> http://code.google.com/p/lilypond/issues/detail?id=787
> 
> This is still not quite right, as Reinhold has found out with the regtest
> accidental-suggestion.ly.
> 
> If there's only one stacking direction, the new code causes a segfault,
> since it
> calls scm_car (null) in the other direction.
> 
> This fails,
> 
> \relative c' {
>    c4-"1"-"2"
> }
> 
> whereas this works:
> 
> \relative c' {
> c4-"1"^"2"
> }
> 
> Then there's another issue with scm->double robustness; if you apply
> Reinhold's patch
> for the above bug, the following still fails due to some grob properties
> returning
> '() if unset (I haven't checked exactly but it's probably
> default_outside_staff):
> 
> \relative c' {
>    c4\turn ^"1"^"2"
> }

Ok,

I'm pretty sure it's fixed now.

I've modified Reinhold's patch slightly and applied it.

I've also replaced scm_to_double calls with robust_scm2double, so we
shouldn't get any failures there.

I have successfully completed make test.

There is a question I have though,

In your opinion, should

c4^"1"^"2"\turn 

put the \turn at the top or the bottom of the stack?

Right now, the \turn shows up at the bottom of the stack, because it doesn't
have an outside_staff_priority so it defaults to 0.

If you think it should go to the top, I could add some more code to give its
outside_staff_priority a default of 0 if it were first in the list, and the
same as the previous element if it weren't first in the list.  Is this
behavior desirable?

A  patch has been posted at

http://codereview.appspot.com/163048


Thanks,

Carl