Re: Barcheck warnings

[Erik Sandberg]

On Friday 09 January 2004 01.47, Han-Wen Nienhuys wrote:
> address@hidden writes:
> > One rule could be to not print the "barcheck failure at a/b" warning in a
> > bar, if a "barcheck failure at a/b" with the same a and b occured on the
> > last | of the same staff. But probably there also exist even more clever
> > mechanisms.
>
> Your proposal seems clever enough. I've added this, but you have to
> set barCheckSynchronize to #f for it to work.

My own objection against it is that it gives two warnings for each mistake, 
which should be reducable to one: You can always assume that the first 
warning that occurs is the mistake. Say it's a barcheck failure at a/b. Now 
we can expect that if we made no error in the next staff, there should be a 
barcheck failure at (b-a)/b. So the wisest thing would actually be to issue a 
warning only if the next staff has a barcheck failure different from (b-a)/b 
(if there is no barcheck failure in that staff, then the probable error is 
that the time sig changed and we forgot to say this, so no more warning would 
be needed).

But this is getting complicated. IMHO the only important difference is between 
O(1) and O(# of bars), the rest is just cosmetics.

(and I hope the barCheckSynchronize can be defaulted to #t sometime in the 
future if it works fine, barcheck failures is not exactly something you plan 
to do :) )

> > Also, I have noticed that midi and paper seem to have separate barcheck
> > mechanisms, the paper's being a bit more intelligent. Included in my wish
> > is that both barcheck mechanisms should follow the same rules.
>
> Can you demonstrate this? This hasn't been the case for some time now.

\version "2.1.5"
\score {
 \notes { a2 | a a | a a | a a | a a | a a | a a }
 \paper {}
 \midi {\tempo 4=100}
}

gives output

-----------------------------------------------------------
lilypond (GNU LilyPond) 2.1.5
Running usr...
Now processing: `test.ly'
Parsing...Interpreting music...
/tmp/test.ly:4:3: warning: barcheck failed at: 1/2:
a2
   | a a | a a | a a | a a | a a | a a
[7]
Preprocessing graphical objects...
Calculating line breaks... [3][6][7]
paper output to `test.tex'...

Interpreting music...
/tmp/test.ly:4:3: warning: barcheck failed at: 1/2:
a2
   | a a | a a | a a | a a | a a | a a

/tmp/test.ly:4:9: warning: barcheck failed at: 1/2:
a2 | a a
         | a a | a a | a a | a a | a a

/tmp/test.ly:4:15: warning: barcheck failed at: 1/2:
a2 | a a | a a
               | a a | a a | a a | a a

/tmp/test.ly:4:21: warning: barcheck failed at: 1/2:
a2 | a a | a a | a a
                     | a a | a a | a a

/tmp/test.ly:4:27: warning: barcheck failed at: 1/2:
a2 | a a | a a | a a | a a
                           | a a | a a

/tmp/test.ly:4:33: warning: barcheck failed at: 1/2:
a2 | a a | a a | a a | a a | a a
                                 | a a

MIDI output to `test.midi'...
Track ...


Analyzing test.tex...
Running latex...
Running dvips...
Running ps2pdf...
DVI output to `test.dvi'...
MIDI output to `test.midi'...
PDF output to `test.pdf'...
PS output to `test.ps'...
-------------------------------------------------------
i.e., midi gives O(n) warnings and paper gives 1 warning.

Erik