.. im going to have to ponder this for a bit an make up my mind what direction to go ...
I think you can save it. Please bear in mind that I'm thinking 'Atlas' here, and the ease of setting up change gears, and especially the ease of engaging/disengaging the spindle gear.
I'm thinking that: If you haven't disturbed the threading/change gears, and the toolpost positioning, that all the possible combinations of thread re-engagement matching is the same as the tooth count of the spindle gear, no matter how the leadscrew is engaged.
If you were to:
Back-out the cross-slide a little, enuff to just clear the threads.
Move the carriage to the threaded area.
Re-engage the leadscrew.
Disengage the spindle gear from the changegear.
Slowly turn the chuck 'till the workpiece thread aligns with the cutting tool.
Re-engage the spindle gear.
Rotate the chuck by hand to remove all slack in the system.
Advance the crossfeed to check tool fit in the old threads.
Closely examine the tool/thread alignment.
Disengage the spindle gear, rotate chuck one tooth at-a-time, re-engage spindle gear.
Re-examine, re-index by one tooth until aligned.
Since the tool must move by one thread spacing for each rotation of the chuck, the total combination of possible tool positions is limited by the spindle gear tooth count.
Of course, this assumes that the only change to the setup was inadvertently disengaging the leadscrew.
Please tell me this is wrong...