Re: make pretty-print available in ly files (issue 222810043 by address@hidden)

[david . nalesnik]

On 2015/04/08 11:14:54, dak wrote:
> On 2015/04/08 10:43:09, thomasmorley651 wrote:
> > On 2015/04/08 10:21:55, dak wrote:
> >
> > > value->lily-string is probably the most useful reasonably generic
printing
> > > facility right now.  But nobody really knows it.
> >
> > At least, I wasn't aware of it, because of (in a ly-file)
> >
> > #(display value->lily-string)
> > -> Unbound variable: value->lily-string

> Like with much of the Scheme-level printing you'll need
> #(use-modules (scm display-lily))

> It's not been around for long in its current form (and name) anyway.
It became
> an exported function under that name with
> commit 8d8e8aec6388fbb08ed2219884b82ecf53a9dbcd
> Author: David Kastrup <mailto:address@hidden>
> Date:   Wed Jan 1 20:54:15 2014 +0100

>      Provide value->lily-string function

> because I really needed that facility in order to make nice error
messages.
> It's in 2.19.1 and 2.18.0 I think, as part of issue 3770.

But doesn't this just return a string?  Also, it requires a parser
argument.


> At some point of time, somebody needs to make some sort of plan along
the lines
> of "what kind of functions and programming facilities are needed for
working
> with LilyPond in a coherent and predictable and comfortable manner"
and put it
> into code.  It is quite too often that one runs into the "I cannot
believe
> LilyPond does not have something obvious for such an obvious task like
$x"
> situation.

And of course documenting it in a convenient way, like "Resources for
..." Can't think of a nice summary word and "easier tampering" or
"hacking" wouldn't send the right message.

It's worth noting that the ~y option IS available within ly files,
though not in scm files, where format has been redefined as
simple-format.  So something like this produces nice results:

#(format #t "~y" (procedure-source grob::rhythmic-location))

==>

(lambda (grob)
  "Return a pair consisting of the measure number and moment within
   the measure of grob @var{grob}."
  (let* ((item (if (grob::has-interface
                     grob
                     'spanner-interface)
                 (ly:spanner-bound grob LEFT)
                 grob))
         (col (ly:item-get-column item)))
    (if (ly:grob? col)
      (ly:grob-property col (quote rhythmic-location))
      '())))

______________

So I guess the way to proceed now is to make a patch reverting this
commit, and then just push it to staging as soon as it passes tests,
rather than going through a review cycle?


https://codereview.appspot.com/222810043/