Re: [Lilypond-auto] Issue 2444 in lilypond: special characters in output-suffix should not be translated

[lilypond]

Comment #1 on issue 2444 by address@hidden: special characters in  
output-suffix should not be translated
http://code.google.com/p/lilypond/issues/detail?id=2444

This is due to code in scm/lily-library.scm in get-outfile-name:
    ;; Allow all ASCII alphanumerics, including accents
    (if (string? output-suffix)
        (set! result
              (format #f "~a-~a"
                      result
                      (string-regexp-substitute
                       "[^-[:alnum:]]"
                       "_"
                       output-suffix))))

This behavior has been there from the start:
commit cacc27ec40e1d28f6667bc0f29789637a3ed09ad
Author: Ian Hulin <address@hidden>
Date:   Thu Nov 12 22:20:36 2009 +0000

    Tracker 836: Add facility to change output file-name for a \book block

    ...or to set a suffix to prevent multiple files over-writing each other  
during a
    compilation. This change allows user to do this via functions rather  
than having
    to do so by using set! and define on parser variables in Scheme.

You'll see that this was present in the first iteration of  
<URL:http://codereview.appspot.com/150044> already.  So the change was  
intentional.  The surrounding material suggests that somewhere measures  
have been taken to let the output-filename admit accented characters (some  
file name code translation?).  It is my guess that the suffix replacement  
was done to make sure that no such characters creep into the suffix, to  
avoid having to recode them there as well.

Somebody (TM) should check where the output filename translation happens,  
and how, and make sure that the suffix does not get treated any worse.