Re: [Lilypond-auto] Issue 2962 in lilypond: account for markup-system-spacing when filling pages

[lilypond]

Comment #2 on issue 2962 by address@hidden: account for markup-system-spacing  
when filling pages
http://code.google.com/p/lilypond/issues/detail?id=2962

Personally, I think the page breaker needs to get rewritten.  The problem  
with the current code is that it does the page breaking  
calculations "abstractly", then creates the page after that.  So we have  
two page builders working independently that need to come to similar  
results.

One consequence is that we have problems in page breaking that are almost  
impossible to track down.  Another consequence is that, for example, the  
page break decisions don't take skylines into account which is clearly  
deficient.  Also it is impossible to, say, format a footnote block  
differently than purely vertically.

Just right now I have a hard time getting the page breaker to _not_  
consider footnote-footer-padding in case there is no footer.  At the point  
the footnote spacing is calculated "abstractly", the information about a  
footer being present or not is not reasonably available.

The primitives for stacking stencils either are O(1) in performance/memory  
usage or can be made to do so, or appropriate equivalents in the page  
breaker can be implemented.  So we could actually remove the redundancy  
with only a constant factor performance hit.  Having only a single instance  
responsible for page building and page breaking would get rid of a lot of  
ways to get almost-but-not-quite working page break decisions.

--
You received this message because this project is configured to send all  
issue notifications to this address.
You may adjust your notification preferences at:
https://code.google.com/hosting/settings