Keith OHara <[email protected]> writes:

> David Kastrup <dak <at> gnu.org> writes:
>
>> Keith OHara <k-ohara5a5a <at> oco.net> writes:
>> >
>> > One can reverse the change, but possibly re-introduce issue 732, with :
>> >
>> > scm/markup.scm
>> > @@ -72,3 +72,3 @@ Example:
>> >                   (head (car stencils))
>> > -                 (xoff (+ space (interval-length (ly:stencil-extent head 
> X)))))
>> > + (xoff (+ space (interval-end (ly:stencil-extent head X)))))
>> >              (ly:stencil-add head
>> 
>> Can you?  I fail to get something like
>> 
>> \markup { x x x \hspace #-15 x x x }
>> 
>> to work, 
>
> Yes, I can.  If I make the change I indicated, then your example
> backspaces, and part (not all) of the complaint that started issue 732
> returns.

Ah, my problem was that I actually calculated the interval length
(bypassing negative distances I expected to see here), and \hspace
special-cases negative distances by placing _both_ stencil bounds to the
left.  That makes it impervious to the "empty interval" problem, but
results in data that is marvelously inconsistent, with a total length
of 0.

At any rate, I think I know slightly more about stencils since my "LGTM"
and venture a guess that the original "fix" was a bad idea.  It does not
seem to move to a useful place when the original stencil has a negative
left border.

-- 
David Kastrup


_______________________________________________
bug-lilypond mailing list
[email protected]
https://lists.gnu.org/mailman/listinfo/bug-lilypond

Reply via email to