Am Di., 6. Jan. 2026 um 10:45 Uhr schrieb Werner LEMBERG <[email protected]>:
>
>
> [See thread starting at
> https://lists.gnu.org/archive/html/lilypond-user/2026-01/msg00026.html]
>
> > 2.12.-epsfile is different from the 2.25. version.
> >
> > 2.12.-epsfile internally calls the old `eps-file->stencil` seems to
> > respect the bounding-box for the final stencil.
> > 2.25.-epsfile internally calls the pretty new image-markup. There you
> > can find the comment:
> > ;; Stencil starts at (0, 0), even for an EPS image whose
> > bounding
> > ;; box does not start at (0, 0).
> > I.e. whatever bounding-box is processed, it is translated to start a (0, 0)
> >
> > This feels like a design-decision. I've no clue about the reasoning
> > and didn't investigate further.
>
> Jean, can you please tell us your reasoning? I think we should
> document this implementation difference – even we missed it that an
> example in the documentation changed...
>
>
> Werner
Actually, the difference occurs between 2.12. and 2.14. already.
The definition of `eps-file->stencil` was changed, and there is the comment:
; We need to shift the whole eps to (0,0), otherwise it will appear
; displaced in lilypond (displacement will depend on the scaling!)
Thus not a new issue.
Cheers,
Harm
