More lilypond testing

I decided it was taking too much mental energy thinking about whether
to convert to 2.6 from 2.0, and I should just decide to do it or not
do it.

So I picked two pieces, one vocal, and the other a dance, both
unbarred parts.

The dance piece converts without problems, but looks close to
identical in both versions. You can see if you agree, but I don’t see
any compelling reason to do a lot of work to go from this
to
this.

For the vocal piece, using abc2ly directly on the ABC works pretty
well, except that some of the lilypond snippets I have in %%ly
directives in the ABC need to be changed. However, I immediately
found a major problem with printing a second verse if you use
convert-ly on the lily 2.0 abc2ly output. So it isn’t really an
option to just switch to 2.6 and convert the previous pieces in a
large book. When I reported this as a bug, I was told something close
to that convert-ly doesn’t convert lyrics between 2.0 and 2.6.
(Actual email from Eric Sandberg: addlyrics (2.0) is converted to oldaddlyrics, which is strongly deprecated,
and pretty unsupported. It is known that it sometimes doesn’t work.

There is no good automated way to convert from oldaddlyrics to lyricsto, so
I’m afraid you’ll have to fix it manually.)

The thread has continued; see the
mailing list archives for further discussion.

This suggests that the Dowlands and Morleys are probably never
going to make it to 2.6. Maybe the next big project. But it does
mean that I have to keep a 2.0 environment working.

So I would say that the only obvious advantage to me of 2.6 over
2.0 is the unicode support, which is only important if I do more
Polish music, which I don’t really have any plans for. The default
font for printing lyrics has changed, and I suppose I could get used
to it, but I can’t say it bowls me over with its elegance. You can
see the difference if you look at
the
2.0 output
and
the
2.6 output.

There have been major changes to the underlying technology between 2.0
and 2.6. TeX is no longer the underlying layout engine, and lily is
producing its own postscript. So given this, it’s actually pretty
amazing that more things haven’t broken. But since I have TeX
installed and working well on my system, that isn’t really a reason
for me to change. Previously I’ve been motivated to do all the work
that converting both the lilypond source and the scripts that produce
the lilypond source by the promise of better looking music at the end
of the tunnel, but I don’t see that I have that in this case.

So until someone shows me a compelling reason to do otherwise, the
Serpent
Publications production environment is going to remain on 2.0.

I do have a script that allows me to test the current CVS version
of lilypond. When I’m feeling virtuous about contributing to open
source software, I will continue to smoketest the abc2ly and
convert-ly scripts to make sure they aren’t obviously broken. And
maybe this will show me that there are benefits to some future version
of lilypond that justify the conversion pain.