[Coco] OS-9 windowing: What is GRP 206?

[Allen Huffman]

I am debugging a weird issue where a font will merge and display under NitrOS-9/6309, but not under the 6809 version. (The same code used to work in 1995.)

Two questions:

1. What is GRP 206 (not in docs)
2. Does the redundant data in stdfonts serve a purpose?

Long notes:

I am parsing the /dd/sys/stdfonts file, and I see it contains GPLoad sequences for the following group/buffer/style:

200/1/8 - 1K - Fnt_S8x8(1) = standard 8x8 font
200/2/6 - 1K - Fnt_S6x8(2) = standard 6x8 font
200/3/8 - 104 - Fnt_G8x8(3) = standard graphics font

After this are a bunch of tiny loads in to GRP 206, including some which are redundant.

206/4/5 - 8 bytes
206/3/5 - 8 bytes
206/2/5 - 8 bytes
206/1/5 - 8 bytes
206/5/5 - 7 bytes
205/6/5 - 6 bytes
206/4/5 - 8 bytes (repeated???)
206/3/5 - 8 bytes (repeated???)
206/2/5 - 8 bytes (repeated???)
206/1/6 - 8 bytes (repeated???)
...and so on...

Anyone know what GRP 206 is? And why are some repeated? Wouldn't that just erase the previous load? i.e.:

offset $889:
1b 2b ce 04 05 00 08 00 08 00 08 00 20 30 38 38 30 20 ff

ofset $8f8:
1b 2b ce 04 05 00 08 00 08 00 08 00 20 30 38 38 30 20 ff

offset $967:
1b 2b ce 04 05 00 08 00 08 00 08 00 20 30 38 38 30 20 ff

Why does it load them three times with the same data? There are multiples like this.

Are the stdfont files corrupted somehow with redundant data we could trim out?

		-- A