Re: recent change to __go_draw_axes__

[Ben Abbott]

On Feb 9, 2013, at 5:27 PM, John W. Eaton wrote:

> After this change:
> 
>  # HG changeset patch
>  # User Ben Abbott  <address@hidden>
>  # Date 1360366055 18000
>  # Node ID 1e35b64ac31ade05453c03bacb9df73ab05b387e
>  # Parent  5b2126a8c84fa4624c737e5921362b79a0e879d4
>  Pad binary data to gnuplot with a space
> 
>  * scripts/plot/private/__go_draw_axes__.m: Add a space to the end of each
>  binary record.  Fixes the "nset obj 1;\n <- line 0: invalid command".
> 
>  diff --git a/scripts/plot/private/__go_draw_axes__.m 
> b/scripts/plot/private/__go_draw_axes__.m
>  --- a/scripts/plot/private/__go_draw_axes__.m
>  +++ b/scripts/plot/private/__go_draw_axes__.m
>  @@ -2027,6 +2027,7 @@
>         endif
>       endif
>     endif
>  +  fprintf (plot_stream, "\n # pad binary data with a space\n")
> 
>   endfunction
> 
> I'm seeing output like this when making some of the figures for the
> manual:
> 
>  $ ../../run-octave -f -q -H -p . --eval "geometryimages ('voronoi', 'eps');"
> 
>  multiplot> 
> �-���?�;[G�&�?Ȟ�ف�?;�����?0a$"�?J���?X�x����?�F��?�T3t��?�+ܣ��?�Q���w�?���:��?j=���6�?�
>             ^
>             line 0: 
> 
> 
>  multiplot> 
> k���?y�,Hy��?�&���?���&���?Lî\��?�D��Zx�?/-Y(!��?1`�+�:�?�5�6�e�?�)ӻ���?����C�?�{i�7��?��,
>                            
> �?j�򫟇�?gD�}J��?���/N�?�������?R5b���?�����?_�W���?�7N��~�?�(a��?g�g�Q�?YwW�u�?)�—�#�?"��KB�?�������?�j���?�ker��?z��?��?�����?��
>        �e�?:I�fj[�?3t��}
>  ����?�Z                �?�����%�?=��?�<�����?N/*  d�?Z�
>              ^
>             line 0: 
> 
> 
> What problem was the extra blank line attempting to solve?  Maybe it
> works with some versions of gnuplot and not others?  I'm using 4.6.0.
> 
> jwe


John,

I'm using ...

        $ gnuplot --version
        gnuplot 4.6 patchlevel 1

The 4.6-1 I never encountered the "nset obj 1" problem, and the change had no 
effect.  I recall that the gnuplot list had a thread complaining that the 
binary segments were difficult to maintain and that it might be a good idea to 
deprecate them.  As far as I know, that never happened, but it does indicate 
the bug may be there.  

I'll try Ubuntu via VBox and see if I can duplication your result.

Ben