day 25 C adjust for more inputs

day 25 C adjust for more inputs

I'm failing some of the alternative inputs here[1]; it looks it
happens only sometimes when count[s] is 1 rather than 0 (as in, my
code does not know whether the cut found includes or excludes s).  It
does not happen when count[s] is 0, but that does not happen on all
inputs.  I'd need more time to investigate why the C code is not
getting it right.  Meanwhile, my m4 code solved days 1-3 of that page
(but slower than my official input - the alternatives are indeed
bigger), but seemed to go into an inf-loop on other days.  I'm not
sure if my code is reliable to all official inputs or if I just got
lucky; nor what it is about the alternative inputs that sends my code
down the wrong path.

[1] https://github.com/mattcl/unofficial-aoc2023-inputs/blob/master/day_025/solutions.md