darktable: Watermarking the photo caption - how?

Hi,
I use the latest windows build of darktable.
What do I want to achieve?
I want to imprint the image caption onto the lower part of the exported image.
So I found the watermark function.
I now created a new file “Caption-Text.svg” as copy of simple-text.svg which contains:

<?xml version="1.0" encoding="UTF-8" standalone="no"?>
<svg xmlns="http://www.w3.org/2000/svg"
     x="0"
     y="0">

  <text x="0" y="5px" text-anchor="start"
        fill="$(WATERMARK_COLOR)"
        font-family="$(WATERMARK_FONT_FAMILY)"
        font-style="$(WATERMARK_FONT_STYLE)"
        font-weight="$(WATERMARK_FONT_WEIGHT)"
        font-size="10px">$(Xmp.dc.description)</text>
</svg>

I can now choose this new svg in the GUI, but sadly it does not work. It does display “$(WATERMARK_TEXT)” as if I was still using the simple-text.svg.

Any hints?

How can $(WATERMARK_TEXT) be displayed since you do not have this in your SVG?

Just tested on Linux, using your caption-text.svg I have no output if my description is empty and otherwise I do have the description appearing as watermark. So looks iike something wrong on your side. We need more information to help.

I was wondering the exact same. I now deleted the svg file and recreated it without capital letters as “caption-text.svg”. Now it works. Even if I rename it to capitals again. So I can not recreate the issue now. Lets hope it doesn’t come back.
Anyhow what still doesnt work now is the font choosing. Using the above caption-text.svg it does display the description always in a sort of default font, regardless what I choose in the GUI.
Only once I type at least one letter into the “Text” field in the GUI, it immediately changes to the chosen font. Is that how it works on Linux as well?

Yes changing fonts works for me on GNU/Linux.

Thanks for the feedback.