Re: Fixes slope errors from incorrect X extents in Beam::print. (issue 5293060)

[Keith OHara]

On Mon, 24 Oct 2011 22:19:41 -0700, address@hidden <address@hidden> wrote:

> Before this patch, the x_span of beams was only ever calculated between the 
> first normal stem and last normal stem of a beam (omitting any trailing beamage 
> on the left or right coming from breaks and/or stemlets).  If it has a 
> consistent slope, however, the x_span of a broken part of a beam should be the 
> whole length, as the trailing beamage on the right and/or left are part of the 
> length between two stems.  This is where the difference comes from.

I see.
When a beam is broken, and we ask for consistent-broken-slope, you are 
imagining an extra stem at the breakpoint.
Then Beam::set_stem_lengths() returns the end-point of that imaginary stem, instead of 
the last real stem, for use in "quantized-positions", because that made it 
easier to match the heights of the broken beam.

> When you say "what use-case will break when we choose consistent-broken-slope after 
> this patch ?", I'm not sure what you mean.

I saw that 'consistent-broken-slope changed the return value of x_span() for broken 
beams, which was used in conjunction with "quantized-positions" for 
Beam::print().  From that I concluded that either :
(a) the old return value was wrong, in which case I wondered why you corrected 
it only sometimes, or
(b) the change would break something that used to work, and we would only 
notice when 'consistent-broken-slope=#t.
I did not consider the possibility that
(c) you made a compensating change to "quantized-positions"