Re: Problems making doc after commit by Mark P

[Carl Sorensen]

On 12/6/10 5:07 AM, "Neil Puttock" <address@hidden> wrote:

> On 6 December 2010 09:26, Mark Polesky <address@hidden> wrote:
> 
>> Some of the error messages:
>> 
>>  error: failed files: "d7/lily-75221729.ly"
>>  error: Children (4 3 2 1 0) exited with errors.
>>  [...]
>>  Child returned 1
>>  make[3]: *** [out-www/collated-files.texi] Error 1
>>  rm out-www/weblinks.itexi
>>  make[3]: Leaving directory `[...]/input/regression'
>>  make[2]: *** [WWW-1] Error 2
>>  make[2]: Leaving directory `[...]/input'
>>  make[1]: *** [WWW-1] Error 2
>>  make[1]: Leaving directory `[...]'
>>  make: *** [doc-stage-1] Error 2
>> 
>> I'll revert my commit #407df9e which introduced these
>> problems, but can anyone explain what's going on?
> 
> Have you looked further up the log file?  I noticed several of the
> following errors from ghostscript:
> 
> Error: /undefined in output-scale
> Operand stack:
>    output-scale   2.83465
> Execution stack:
>    %interp_exit   .runexec2   --nostringval--   --nostringval--
> --nostringval--   2   %stopped_push   --nostringval--
> --nostringval--   --nostringval--   false   1   %stopped_push   1878
> 1   3   %oparray_pop   1877   1   3   %oparray_pop   --nostringval--
> 1861   1   3   %oparray_pop   1755   1   3   %oparray_pop
> --nostringval--   %errorexec_pop   .runexec2   --nostringval--
> --nostringval--   --nostringval--   2   %stopped_push
> --nostringval--   --nostringval--
> Dictionary stack:
>    --dict:1160/1684(ro)(G)--   --dict:0/20(G)--   --dict:110/200(L)--
> Current allocation mode is local
> Last OS error: 2
> Current file position is 8548
> 
> All the failed files have one thing in common: a global staff-size
> setting, which causes output-scale to become a rational:
> 
> \paper {
>   #(display output-scale)
> }
> 
> -> 1.75729901757299
> 
> but
> 
> #(set-global-staff-size 16)
> \paper {
>   #(display output-scale)
> }
> 
> -> 10160/7227
> 
> I suspect it's the conversion of output-scale to a string in
> framework-ps.scm which causes the problem, since number->string
> doesn't change a rational to a floating point value (needs
> exact->inexact).

For clarification, the following patch may help:

diff --git a/scm/output-ps.scm b/scm/output-ps.scm
index ac7cc53..670f8ec 100644
--- a/scm/output-ps.scm
+++ b/scm/output-ps.scm
@@ -145,7 +145,7 @@
 ~a
 ~a print_glyphs")
          postscript-font-name
-         size
+         (exact->inexact size)
          (string-join (map (lambda (x) (apply glyph-spec x))
                            (reverse w-x-y-named-glyphs)) "\n")
          (length w-x-y-named-glyphs)))

HTH,

Carl