[Bug] r.univar: Wrong cell number for other zones than the first

[ninsbl]

Describe the bug
When r.univar is called with zones and the g-flag for shell script output, the "cells" attribute gets always the number of cells in the first zone encountered, which does not correspond to the actual number of cells in the zone (or in the region in total).

The issue is likely this line:

grass/raster/r.univar/stats.c

Line 157 in 2edf389

fprintf(stdout, "cells=%lu\n", stats->size);

which should read:

fprintf(stdout, "cells=%lu\n", stats[z].size);

To Reproduce
Steps to reproduce the behavior:

g.region -p raster=landclass96
r.univar map=elevation zone=landclass96 -g

and compare the values in "cells"

Expected behavior
The value of cells should represent the number of zones also in shell script style.

System description (please complete the following information):

• Operating System: Fedora Linux 37
• GRASS GIS version 8.3.dev