Re: reduce unbound within \context lambda

[Aaron Hill]

On 2019-03-02 3:11 am, Thomas Morley wrote:
> fold, reduce any many other procedures are not part of core-guile, but
> of the SRFI-1 module.
> Nowadays SRFI-1 is present in every ly-file. Not in \paper and \layout, 
> though.

That's what I get for not reading the Guile docs closely enough.  I knew 
the section was within SRFI-1, but it was titled "Fold, Unfold & Map".  
Since map appeared to work, I assumed that SRFI-1 was included and 
something was amiss with fold/reduce.  But I now see that map "is 
extended with respect to R5RS, because the argument lists may have 
different lengths."

That got me thinking.  Since SRFI-1 defines its own map, the one in 
\layout or \paper would be the core R5RS version and not take kindly to 
lists of differing lengths:

%%%%
\version "2.19.82"
#(display (map + '(1 2 3) '(1 2 3 4)))
\paper { #(display (map + '(1 2 3) '(1 2 3 4))) }
%%%%

Sure enough:

====
GNU LilyPond 2.19.82
Processing `map-srfi-1.ly'
Parsing...(2 4 6)
map-srfi-1.ly:3:11: error: GUILE signaled an error for the expression 
beginning
here
\paper { #
          (display (map + '(1 2 3) '(1 2 3 4))) }
Argument 3 out of range: (1 2 3 4)
fatal error: failed files: "map-srfi-1.ly"
The terminal process terminated with exit code: 1
====

> I don't know how one could change it.
> Thus I usually define things outside from layout/paper:

It seems \header and \midi also fail like \layout and \paper, so that is 
a perfectly reasonable approach to completely avoid these issues.

For fun, I tried adding an explicit use-modules:

%%%%
\version "2.19.82"
\paper {
  #(use-modules (srfi srfi-1))
  #(display (map + '(1 2 3) '(1 2 3 4)))
}
%%%%

====
GNU LilyPond 2.19.82
Processing `map-srfi-1.ly'
Parsing...WARNING: #f: `break' imported from both #f and (srfi srfi-1)
(2 4 6)
Success: compilation successfully completed
====

Nope.  I don't like the look of that warning.  :-/


-- Aaron Hill