replace 10 by 10.0, so that the conversion is not done at runtime

diff --git a/lang/cem/libcc/gen/ecvt.c b/lang/cem/libcc/gen/ecvt.c
index 84ca9f3..43a0bf9 100644 (file)
--- a/lang/cem/libcc/gen/ecvt.c
+++ b/lang/cem/libcc/gen/ecvt.c
@@ -26,8 +26,8 @@ cvt(value, ndigit, decpt, sign, ecvtflag)
        int ndigit, *decpt, *sign;
 {
        double intpart, fractpart;
-       static char buf[NDIGITS];
-       char buf1[NDIGITS];
+       static char buf[NDIGITS+1];
+       char buf1[NDIGITS+1];
        register char *pe = buf1;
        register  char *pb;
        int pointpos = 0;
@@ -47,7 +47,7 @@ cvt(value, ndigit, decpt, sign, ecvtflag)
                do {    /* get digits of integer part, low order digit
                           first
                        */
-                       value = modf(intpart/10, &intpart);
+                       value = modf(intpart/10.0, &intpart);
                        /* compensate for rounding errors, because
                           the conversion to "int" truncates
                        */
@@ -56,7 +56,7 @@ cvt(value, ndigit, decpt, sign, ecvtflag)
                                pe = &buf1[0];
                                while (pb < &buf1[NDIGITS]) *pe++ = *pb++;
                        }
-                       *pe++ = (int)((value+.05) * 10) + '0';
+                       *pe++ = (int)((value+.05) * 10.0) + '0';
                        pointpos++;
                } while (intpart != 0);
                pb = buf;
@@ -66,7 +66,7 @@ cvt(value, ndigit, decpt, sign, ecvtflag)
                pb = &buf[0];
                if (value > 0) {
                        fractpart = value;
-                       while ((value = value*10) < 1) {
+                       while ((value = value*10.0) < 1) {
                                fractpart = value;
                                pointpos--;
                        }
@@ -85,7 +85,7 @@ cvt(value, ndigit, decpt, sign, ecvtflag)
                if (pe > &buf[NDIGITS]) pe = &buf[NDIGITS];
        }
        while (pb <= pe) {
-               fractpart = modf(fractpart * 10, &value);
+               fractpart = modf(fractpart * 10.0, &value);
                *pb++ = (int)value + '0';
        }
        pb = pe;