Re: [Lilypond-auto] Issue 3777 in lilypond: Code improvements to satb.ly built-in template

[lilypond]

Comment #18 on issue 3777 by address@hidden: Code improvements to satb.ly  
built-in template
http://code.google.com/p/lilypond/issues/detail?id=3777

The difference between $x and #x is twofold: $x is analyzed in the lexer,  
given a particular token category based on its value, and then further  
interpreted.  That means that things like
x = { c d e f }
$x
don't work since the parser requires lookahead of one token to decide  
whether this is going to turn into
x = { c d e f }
\addlyrics { Hi! }
and when $x is evaluated by the lexer, things go awry.  #x just _reads_ x  
without looking at its value and tells the parser "this is going to end up  
as a Scheme expression".  Then it does the assignment to x, and only  
afterwards gets to evaluate the value of the already read expression x.   
This works nicer, but it requires the _parser_ to interpret the value as  
desired.

Now for \layout, you almost always want a _copy_ of the layout rather than  
the original to play with.  Since $x also copies while #x only interprets,  
supporting #x for layout may lead to bad usage easily.  Not that this was  
an intentional decision.

The point is that the meaning of any #xxx expression is not wired into the  
Parser grammar: the grammar cannot distinguish any #x from any other.   
Instead, the code executed  by #xxx has to go and analyze the set of values  
and take appropriate action that usually cannot (or cannot easily) result  
in different interpretations depending on the _type_ of expression as  
looking at the value is a _semantic_ action, and the ways of letting those  
feedback into the _syntactic_ analysis are limited and cumbersome.

--
You received this message because this project is configured to send all  
issue notifications to this address.
You may adjust your notification preferences at:
https://code.google.com/hosting/settings