Compiler complaints about char vs unsigned char

[jonsmirl]

I had to change this, because isdigit() is internally implemented with an array on my system and the compiler just refused to allow a possible negative index into the array. gcc 12.2.0

jonsmirl@ares:~/lowpan/components/forth$ git diff floatops.c
diff --git a/components/forth/floatops.c b/components/forth/floatops.c
index 3276f69..d3db981 100755
--- a/components/forth/floatops.c
+++ b/components/forth/floatops.c
@@ -11,7 +11,7 @@
 #include <stdlib.h>
 #include <ctype.h>
 
-char *ecvt(double n, int ndigits, int *decpt, int *sign);
+unsigned char *ecvt(double n, int ndigits, int *decpt, int *sign);
 char *fcvt(double n, int ndigits, int *decpt, int *sign);
 
 #ifdef NeXT
@@ -233,7 +233,7 @@ fintop(int op, cell *sp, cell *up)
                } else {
                        int decpt, sign;
                        int places = (int)V(FNUMPLACES);
-                       char *digits = ecvt(ftos, places+1, &decpt, &sign);
+                       unsigned char *digits = ecvt(ftos, places+1, &decpt, &sign);
                        char *s = floatstr;
                        if (sign)
                                *s++ = '-';
@@ -310,7 +310,7 @@ fintop(int op, cell *sp, cell *up)
                {
                    int len = *sp++;
                    char *adr = (char *)(*sp++);
-                   char *radr;
+                   unsigned char *radr;
                    int decpt, sign;
 
                    radr = ecvt(ftos, len, &decpt, &sign);