ColourTrans in 256 colour modes

I am trying to figure out what ColourTrans is doing in a 256 colour mode when it’s being asked to set a GCOL from a palette entry.

REM >coltest

MODE 28

FOR x% = 0 TO 31
  SYS "ColourTrans_SetGCOL", ((x% * 8) << 16), 0, 0, 0, 0 TO gcol%
  REM GCOL gcol%

  GFCOL% = VDU(153)
  GFTint% = VDU(157)

  PRINT x%, gcol%, GFCOL%, GFTint%

  RECTANGLE (x% * 32), 0, 32, 32
  VDU 5
  MOVE (x% * 32) - 128, 16
  PRINT gcol%
  VDU 4
NEXT

END

As can be seen, the boxes at the bottom are the expected green gradient.
Oddly, the VDU variables for current GCOL and TINT are always the same – how is ColourTrans not touching these?
Additionally, the returned GCOL number makes no sense. If I were to comment out the GCOL line following the SWI, it would go quite wrong. Plus… GCOL? Not GCOL&TINT?

What I’m looking for is: given an amount of red, green, or blue…return some value that can be used in the future to write that colour (or as close as it gets within the limitations of the mode) in the future. It seems like ColourTrans ought to be able to do this, but I’m just not making any sense of what I’m getting back from it. Am I missing something?

The colour and tint returned are packed into a single byte. See PRM 3-337.

colour% = gcol% >> 2
tint% = (gcol% AND 3) * 64

GCOL colour% TINT tint%

Does SetGCOL use an ECF even though bit 8 of R3 is clear?
I looked at PRINT VDU 151 … it’s 96. I can’t make heads or tails of that number.

The colour and tint returned are packed into a single byte.

Oh, right, it’s the one with the tint bits at the bottom, not the top. I was obviously “holding it wrong”.

Brilliant, thanks!