Re: [frogs] 836: output-suffix as a (Score) context property

[Ian Hulin <ian@xxxxxxxxxxxx>]

On 14/03/10 17:19, Graham Percival wrote:
> http://code.google.com/p/lilypond/issues/detail?id=836
>
> What's the latest on this issue?  The last news I see is from 16 Nov
> 2009, saying """I wasn't able to implement output-prefix and filename
> as "contexts" to \book easily, so we now have
> \bookOutputSuffix "newsuffix" and \bookOutputName "filename" which are
> scoped within each \book block.
>    
I went the music-function route because there was a feeling that 
contexts were properties which were reserved for the phase of music 
translation and producing grobs.  When I started working on this I 
reckoned contexts were Lilypond-speak for sets of attributes within the 
compiler, so it seemed reasonable to attach some new attributes to the 
the compilation unit (i.e an implicit or explicit \book block) to 
achieve this.  However, although the \book block is treated as OO object 
in the C++ code, I got very confused trying to work out what was at 
top-level, what was the object hierarchy and what was the context 
hierarchy.

So I worked out a way of doing it with music functions and parser 
variables instead, which met the demand from users for begin able to 
rename to backend output files simply from within code.

If people really want this functionality via a context-like interface 
I'll have another crack at it, but I don't really feel I understand the 
way Lily fits together well enough yet.

> There is also a regression test to cover output files having extended
> Latin accented
> characters (like č,š,ů,ř,ő,ž,ij,ě)."""
>
> There's also a "patch" label.
>
> - is the issue finished now?
> - is there any patch pending?
>    
I got lost in how the regression test files get generated. I thought it 
would be enough to have a file called something like
"output-filename-Zčeskýchluhůahájů.ly" - here's the file contents
\version "2.13.5"

\header {
  texidoc = "Using extended Latin characters like š and č
  in output file-names should not cause any problems with
  backend processors."

  dedication = "Lilypond Regression Test"
  title = "Má Vlast"
  subtitle = "Symphonic Tone Poem Cycle"
  instrument = "Orchestra"
  composer = "Bedřich Šmetana"
  opus = "Opus 64"
}
% Set up variables
globalIV = {\key f \major \time 4/4}
violinmIV = \relative c'' { \globalIV  d1 | }

%Now produce back-end files with names using extended accents

\book {
  \header {
   subsubtitle = "IV. Z českých luhů a hájů"
  }
  \score {
    \new Staff  \violinmIV
    \layout { }
    \midi { }
  }
}

But I couldn't work out how to stitch this in to the reg test suite 
using this name, since the make check seemed to insist on generating its 
own names on the fly, which made it dfficult writing a test testing to 
see whether files with a specific name were generated.

The CG Chap 8 has no content on regression tests so I'm stuck.

Cheers,

Ian

---
----
Join the Frogs!