[dvipdfmx] improving color stack in xdvipdfmx

[Shunsaku Hirata]

Dear Philip-san,

Maybe you are caring about dvipdfmx vs. xdvipdfmx differences?
These days both programs are the same (single source code) but
having different program names.

All changes in dvipdfmx will also be available in xdvipdfmx and no
additional works for XeTeX are required.

Thanks,
Shunsaku Hirata

2020年6月24日(水) 19:59 Philip Taylor <P.Taylor at hellenic-institute.uk>:
>
> Shunsaku-san —
>
> I'm currently working on extending dvipdfmx to support various color
> models. (mostly finished but testing remains)
>
> This is implemented just as an extension to the existing color special
> of dvipdfmx, pdf:bcolor and pdf:ecolor, of the form:
>
>   \special{pdf:bcolor @colorspace [1.0 0.4 0]}
>   ...
>   \special{pdf:ecolor}
>
> where @colorspace is an object representing color space defined
> somewhere before. (/DeviceCMYK and other device color model
> names can also be used in place of object reference)
>
> For example, to use ICCBased color space,
>
>   \special{pdf:fstream @iccp (sRGB.icc) << /N 3 >>}
>   \special{pdf:obj @colorspace_srgb [ /ICCBased @iccp ]}
>   \special{pdf:bcolor @colorspace_srgb [0.0 1.0 0.0]}
>   Hello World!
>   \special{pdf:ecolor}
>
> Here, ICC profile "sRGB.icc" is loaded. Other color models are also
> supported. First define color space by pdf:obj and just refer to that
> object in pdf:bolor.
>
> Does this look OK?
>
>
> Thanks,
> Shunsaku Hirata
>
>
> I am not sufficiently experienced in this field to say whether or not it looks "OK", but it certainly looks extremely useful.  I hope that Ulrike can offer a more expert oninion.  If you were to make this change, would it automatically carry through to xdvipdfmx (as slaved by XeTeX), or would that require additional work on your part ?
>
> Philip Taylor
>
>