Re: [Lilypond-auto] Issue 2783 in lilypond: wrong placement of timesignature

[lilypond]

Comment #32 on issue 2783 by address@hidden: wrong placement of  
timesignature
http://code.google.com/p/lilypond/issues/detail?id=2783

Confirming that commit d10ec4f5a95d5 before 2.15.13 caused Lilypond to  
count 'line-positions and use that count instead of 'line-count  
if 'line-positions is defined.  This moved the time signature in ČÁRY  
off-center.

So if we revert 0b5d0acd, the example at the start of this issue has its  
time-signature raised to position-one, so the 3 in 3/4 pokes over the top  
line.  (Please post an image if you can confirm.) I think everyone would  
consider this even worse.

It would not be too hard to reverse the re-definition of  
Staff_symbol::line_count() from d10ec4f5a95d5 and continue to have the  
patch work the same, internal ledger-lines for issue 1193, but I am not  
recommending we do so.

LilyPond does print the time signature, when using \tabFullNotation,  
off-center on guitar tab so that it is centered on a line.  All the other  
sources of tablature I have or can find (not much) put time signatures in  
the center regardless of lines.

So putting time signature at position zero, or center, regardless of  
line-count might be the most helpful thing to do.

Pál, your examples all seem to work more simply with time-signature at  
zero, at least as a starting point.  I see in your patch how centering  
solved inconsistencies introduced by d10ec4f5a95d5, while being consistent  
LilyPond's unusual placement for even line-count.  Centering on zero would  
be less confusing for anybody else who has been using 'line-positions #(-4  
-2 0 2)