fix parse_float bug

Author: jcupitt

ChangeLog

@@ -2,6 +2,7 @@ started 8.5 24/1/17
 - add max_slope to lhist
 - gaussnoise goes via vips8 now
 - add snake option to array join [Joe Padfield]
+- parse_float was broken for numbers starting "0."
 
 started 8.4.1 25/9/16
 - simplify nip2-icon.rc build, bgilbert

TODO

@@ -1,3 +1,9 @@
+- should parse_float / _int etc. allow leading and trailing spaces?
+
+    (split is_space x)?0
+
+- add entropy menu item
+
 - sharpen should use new interface?
 
 - can we call affine from nip2 vips_call? do we need a double array?

share/nip2/start/_convert.def

@@ -189,7 +189,7 @@ parse_float l
 
 	ipart = parse_int parts?0;
 	sign 
-		= 1, ipart > 0
+		= 1, ipart >= 0
 		= -1;
 	fpart
 		= 0, parts?1 == [];