17:13:19Inlinethe correction of (count (- width (mod column tab-width)))) to (if (= tab-width 0) 1 (- tab-width (mod column tab-width))))) and then putting immediately a line (decf offset count) solves the problem
17:13:54Inlinei.e. before all the rest of (when (looking-at-spaces) ... stuff begins
17:14:16Inlinethen we prevent the division by zero error and don't change count at all
17:33:40Inlineinitially the loop does not stop at 0 when you observe the >= stuff, it has one more cycle left so to say, but with 0 it bails out and no chance of preventing that, but when you give 1 as a result of the mod operation and decrement just afterwards then one didn't change the cycle count and the relative offset differences (offset offset2) remain constant, otherwise we introduce a off-by-one error and you get something
17:35:15Inlinemy previous <= was a total noob fail, since it corresponds to a no-op, and i was wondering why it was so fast (cause it didn't change the tabification of the opened file at all)