Re: [GLISS] differentiating pre/post/neutral commands

[David Kastrup]

Thomas Morley <address@hidden> writes:

> That said, now some thoughts ontopic:
> In most cases I'm quite fine with the current lily-syntax concerning
> the pre/post-commands.
> But there are cases I'd wish to be more consequent.
>
> Example:
> <c-1\2\rightHandFinger #2  g>
> Sometimes backslash, sometimes dash!

Dash is required for single-character shorthands like -. -- -^ -_ -3 and
similar.  It is also required when a _music_ function returning an
articulation follows (like \tweak ... -.).  It is not required when an
event function or a variable containing an articulation follows.  You
_can_ always add it.

> Well,
> <c-1\2-\rightHandFinger #2  g'>1
> works, too.
>
> But
> <c-1-\2-\rightHandFinger #2  g'>1
> not.

Why wouldn't it? [trying it out] Pfffffft.  You are right.  I consider
this a bug worth fixing.

> BTW, writing this post and looking for more examples I was surprised,
> that sometimes adding a dash to a post-command works already:
>
> c1-\startTextSpan f-\stopTextSpan

It should always work, without exception.  It is just that sometimes it
is optional.

> Summarize:
> I'd vote for no extrem changes to the syntax, but to improve it
> slightly in a more consequent manner.
> Maybe it should be _possible_ to prepend a dash to every post-command.

That's definitely the intent.

> Concerning music-functions, I think we should keep the current
> behaviour. Music-functions mostly affect a larger section of  music,
> appending them to any single event would be curious, except this  _is_
> the wanted behaviour:
> c-\parenthesize-|

I would like to have this become equivalent to c\parenthesize-| but it
still requires some work to allow LilyPond to _first_ call a music
function, _then_ decide whether the result is a post-event to be
attached to the previous note.

-- 
David Kastrup