Re: Possible bug in EPS export for Network Miditower object? (fwd)

[Cyrille Chepelov <cyrille chepelov org>]

Le Thu, May 23, 2002, à 03:52:46PM +0100, Andrew Ferrier a écrit:

> I would say that this is increasingly looking like a bug in the
> EPS export routines. Unfortunately I don't have the time to go
> through them and find it but I hope this helps the developers
> track it down.

Do you mean,

static void
set_linewidth(RendererEPS *renderer, real linewidth)
{  /* 0 == hairline **/
  if (linewidth == 0.0) linewidth=.1; /* Adobe's advice */
  fprintf(renderer->file, "%f slw\n", (double) linewidth);
}


???

Line width 0 "as fine as the device allows". Adobe strongly warns against 
using device-dependent things in a Postscript file (think about what happens
on a 2400dpi output device).

You'd probably much better using a line width of 0.01 if this is really what
you want to see.

What *is* buggy, though, is that we get an inconsistent behaviour. However,
I'll call this inconsistency a feature of the device-dependentness of line
width 0 for the duration of the 0.90.RC# time frame.

(yes, I know. Many places in the code call set_linewidth(this,0). This is
bad).

	-- Cyrille

-- 
Grumpf.