it's hard to explain since i haven't analysed all the data yet and i still haven't got stable, repeatable results, but i'll try.
and yes, naturally i use c1 & c2 (one is set on AF, the other on manual, all other settings are identical)
as for the first of my 2 posts:
when in MF mode and using chdk it's fairly easy to change the endless range in MF, getting it back is just putting it in AF, focus far end and use half press + left to activate MF.
i thought that might have advantages, so i exploited that a bit until i lost half my focus range to endless and then it turned out not to be handy anymore

any value of the focus at any focal range can become -1, either because CF is not ready or misinterprets data or because i've cheated it in thinking that's where the range ends...however that will not stop it from focussing until it physically stops at the front lens assembly.
- i am trying to use the g1x for landscape time lapse (far end focus).
- because of the large sensor endless is just a bit too much, so i want to lower it a bit.
- since the cam will reboot and/or be turned off and on again i need to have settings saved in C1 or C2 for best results
there is not much to focus on in the beginning.
what i've experienced is that when i start in AF and move to AFl in a script i get widely varying focus values when it's set. it can also float away during a run, but also after a (re)boot.
Why you might ask, well the cam sometimes goes to 10 meters (there is nothing to see in that range) and might go up to 70 meters on yet another moment before it switches to AFL and sets zoom. from there zooming to my favorite point can be close and precise or far and it just end up somewhere in the neighbourhood.
I dont mind focus being off by 5 meters, i do mind if it is off by 30.
Although safety MF is off it still does the calculations on half press before shoot i think and that's where the focus(mm) values skew off.
when setting in MF i'm experiencing something different but also the same.
it appears that when my start is further than my actual setting point in MF chances are bigger that it will get closer to the set value.
because i'm already focussing close to the far end it is hard to get a higher, but still not -1 value set. that will remain not -1 between (re)boots
it also appears that the focus is not a stepper, so the CFP values are rounded.
can almost start writing a paper on this

but without solution for the change in focus i clearly don't know enough yet.