Re: Issue 3935: Use (pretty-print) for some IR props. (issue 95710044)

[markpolesky]

On 2014/05/30 10:45:36, dak wrote:
> >    ; (ly-type? vector) => #t

> That's rubbish.  None of the given types in ly-type? should trigger
for a
> vector.  And indeed it would appear that the definition of
ly:music-list? is
> broken and returns #t for anything that is not a list.

Your new commit still leaves this situation:
  (ly:music-list? '()) => #t

Is that intended?

> Instead of trying to work around obvious bugs when one finds them,
they should
> be reported and fixed.

Well, in my defense, I didn't realize it was a bug, but thanks for the
speedy fix.

> > That's correct, no quotes, hence the added quotes a few lines below:
> >    (string-append "\"" str "\"")

> But you'd get ("x" y") to print as (x y) then, wouldn't you?

Yes, I hadn't noticed that.  I had added `#:display #t' because without
it, even slightly long lines would get wrapped, even when (pretty-print)
was given a large value of #:width:

'(...
  (-1 . "accidentals.flatflat")
  (3/4
   .
   "accidentals.sharp.slashslash.stemstemstem")
  (1/4 . "accidentals.sharp.slashslash.stem")
  ...)

Is that a bug with (pretty-print #:width ...)?  Or am I misunderstanding
something?  Anyway, you are right; as it stands, my patch drops the
double-quotes with these strings:

'(...
  (-1 . accidentals.flatflat)
  (3/4 . accidentals.sharp.slashslash.stemstemstem)
  (1/4 . accidentals.sharp.slashslash.stem)
  ...)

What would you suggest?  Should I recurse into the prop-values to put
quotes around all those inner strings?  Or just accept the needless
line-wrapping (which is ugly to me)?  Do I need to file a guile bug?
For now, I'll remove the `#:display' since it's not right.

I've uploaded a new patch that incorporates your new commit and other
things, including a more maintainable definition of `ly-type?'.  Please
check it out.

Thanks!
- Mark

https://codereview.appspot.com/95710044/