Hi, I'm doing a little housekeeping on my Rebol bundle. I've one command that takes a selection, evaluates the content and replaces the selection with the product. I'm using TM_SELECTION to determine the various selection modes:

• No selection: use TM_CURRENT_LINE as input
• Regular selection: use TM_SELECTED_TEXT as input
• Columnar selection: use TM_SELECTED_TEXT splitting on LF

I don't currently handle multiple caret selection, though I see that I can read from STDIN to obtain that input. The tricky part seems to be discerning how to break that input to match the respective selections. I could split by LF, though that doesn't work if any of the selections themselves span multiple lines. Otherwise it seems it's a walk through each line to match up with the TM_SELECTION entries (which I've not attempted thus far, in fairness).

Am I missing anything obvious here? Has there been any consideration of using some of the ASCII field separators (e.g. 0x1D-1F) to delimit these and permitting use of said separators as a way of directing the response to fill each of these selections respectively?

– Chris