small (1 and 2 word) floating point literals not parsed correctly

Fix gitlab issue #5.

The parse_float() function has several bugs:

    1. For 1- and 2-word floats it will always write all 4 bytes if the value is
       exactly 0.
    2. It incorrectly rounds and normalizes 1- and 2-word floats.

To see the issue easily try the following inputs:

     .flt2  1.0
     .flt4  1.0

These will assemble as '040100 000000' and '040200 000000 000000 000000'.
They should both begin '040200'.

In fact the test file test/test-prec.mac is incorrect in its only floating point value:
	.word	^f 1.5		; 040140
should actually assemble as 040300, not 040140 (040140 is actually ^F0.875).
I confirmed this on RT-11 running MACRO V05.06.

I fixed the problem with the following deltas:

[the patch]

The most crucial change is the very last one. 0x200000000000 is actually (1 << 45)
and because ufrac is normalized it means that it will always downshift ufrac by 1.

tests/test-prec.lst.ok

@@ -45,7 +45,7 @@
       45 000100 070204                  	.word	^r<R D>		; bracketed strings are an extension
       46 000102 070250                  	.word	^r<RA >		; bracketed strings are an extension
       47 000104 157614                  	.word	^R50.		; 157614 there is no whitespace in between.
-      48 000106 040140                  	.word	^f 1.5		; 040140
+      48 000106 040300                  	.word	^f 1.5		; 040300
       49                                
       50                                ;;;;;
       51                                ;

tests/test-prec.mac

@@ -45,7 +45,7 @@
 	.word	^r<R D>		; bracketed strings are an extension
 	.word	^r<RA >		; bracketed strings are an extension
 	.word	^R50.		; 157614 there is no whitespace in between.
-	.word	^f 1.5		; 040140
+	.word	^f 1.5		; 040300
 
 ;;;;;
 ;