Re: Upgrading to 1.5.52

[Juergen Reuter]

On 27 May 2002, Laura Conrad wrote:
> ...
>     Juergen> Unfortunately, this currently does not work for
>     Juergen> TimeSignature for some internal reasons of lily (to be
>     Juergen> more precise, alist font selection in scm/font.scm);
>     Juergen> i.e. mensural time signatures currently do not work at
>     Juergen> all.  This should be fixed until 1.6.
>
> It looks to me (from running 1.5.58 on time.ly) that this
> still doesn't work.  This means that 1.5 is not feature complete from
> my point of view.
>

Correctly.  The problem is that, for fixing this error in a clean way, the
font selection mechanism needs to be revised.  Personally, I am voting for
replacing the user-visible font-family property by a global 16-bit font
character map that maps character names (or codes?) to pairs of
(font-family, character-code) which are not visible to the user, so that
the user only sees a unique global space of characters, but no font
families.  For easily switching between different font-series (e.g.
switching from modern font to mensural font by just changing a single
property in the font lookup mechanism), we still can introduce an
SCM functions that map characters *within* the 16-bit space, without
worrying about font-families.  I.e. we do some kind of preprocessing of
character codes within the 16-bit space, and the user simply selects the
preprocessing function(s).

However, this approach requires major changes, and I guess that Han-Wen
and others will vote for defering this to 1.7.  So, I fear the only way to
fix this is to add some preliminary hack, that switches between feta
and parmesan font in a hard-coded way.  Currently, scm/font.scm says:

        (timesig-symbol . ((font-family . music) ))

I.e. there is no access to the ancient font.

> I also ran 1.5.58 on ancient-font.ly, and there are some things that
> look pretty wierd to me.  Specifically, I would guess that the
> squiggles in the second staff of the third system should line up
> differently with the stems.
>

I will have a look for this during the debugging phase for 1.6 (i.e. after
the feature freeze).

As for features still missing for 1.6, I currently see:

* ambitus engraver (I am currently working on it; should be finished in a
  few days);

* mensural rests vertical alignment: I already submitted a patch (with
  some serious limitations), which was rejected because of its complexity;
  I want to give it a second try soon.

* time-signatures, this is broken in two ways: (a) regarding the font
  selection as described above; (b) the handling of mensural meter
  as discussed during the last two weeks.

Originally, I also planned to implement some support for mensural incipits
(starting the ambitus engraver); but I think this will be deferred to 1.7.

And, of course, there are lots of small bugs to be fixed.

Greetings,
Juergen