Issue with handling the degree character

I have a problem with the degree symbol.

For the record, the environment is: MacOS X 10.6.4 (Intel 
processor). The source file is edited in BBEdit 9.6.1 (2860) of 
Thu, 11 Nov 2010.

The symbol is used in a file the first two lines of which are 
(look for "Temp, °C"):
"Plot Title: 997111 - Cave "
"#","Time, GMT-07:00","Temp, °C","Intensity, Lux","Batt, 
V","Coupler Attached","Stopped","End Of File"

When I print the start of the file with od -ctxC, I get

0000000    "   P   l   o   t       T   i   t   l   e   :       
9   9   7
            22  50  6c  6f  74  20  54  69  74  6c  65  3a  20  
39  39  37
0000020    1   1   1       -       C   a   v   e       "  \r  
\n   "   #
            31  31  31  20  2d  20  43  61  76  65  20  22  0d  
0a  22  23
0000040    "   ,   "   T   i   m   e   ,       G   M   T   -   
0   7   :
            22  2c  22  54  69  6d  65  2c  20  47  4d  54  2d  
30  37  3a
0000060    0   0   "   ,   "   T   e   m   p   ,     241   C   
"   ,   "
            30  30  22  2c  22  54  65  6d  70  2c  20  a1  43  
22  2c  22
0000100

The degree symbol is represented as the single character 0xa1 in 
the file.

I read the file with:
   for line in file {
     if (line =~ `@_,@d1,@d2,@t1,@t2,@l1,@l2,@_`) {

I code a test for it in my program:
       if (t1 != "\"Temp" || t2 != " °C\"") {

Which results in the following disk version of the source code:
0004220    "   T   e   m   p   "       |   |       t   2       
!   =
            22  54  65  6d  70  22  20  7c  7c  20  74  32  20  
21  3d  20
0004240    "       °  **   C   \   "   "   )       {  \r  \t  \t
            22  20  c2  b0  43  5c  22  22  29  20  7b  0d  09  
09  20  20
Where the degree symbol is represented by the sequence 0xc2b0.


After the test fails, I added some code to print the constant 
and the parsed value:
def px {
   to printhex(value) {
     var str := ` value.size ${value.size()} `
     def albet := "0123456789ABCDEF"
     for ch in value {
       def hich := ch.asInteger() >> 8
       def loch := ch.asInteger() & 255
       str += albet[hich//16]
       str += albet[hich%16]
       str += albet[loch//16]
       str += albet[loch%16]
     }
     println(str)
   }
}

#...
             def con := " °C\""
             println(`Constant $con`)
             px.printhex(con)
             println(`t2 $t2`)
             px.printhex(t2)

Which prints:

Constant  °C"
  value.size 5 002000AC221E00430022
t2  ?C"
  value.size 4 002000B000430022

When I changed the text encoding in BBEdit from "Unicode 
(UTF-8)" to "Western (MacOS Roman)", the test succeeded and printed:
Constant  ¡C"
  value.size 4 002000B000430022
t2  ¡C"
  value.size 4 002000B000430022

The source file changed to:
0004760    !   =       "   \   "   T   e   m   p   "       |   
|       t
            21  3d  20  22  5c  22  54  65  6d  70  22  20  7c  
7c  20  74
0005000    2       !   =       "     241   C   \   "   "   
)       {  \r
            32  20  21  3d  20  22  20  a1  43  5c  22  22  29  
20  7b  0d
The degree symbol is represented as the single character 0xa1 in 
this source.


It seems to me that something in either reading the file or the 
quasi-parser is converting the degree symbol to a 0x00b0. Also 
something is happening to the Unicode value between the source 
file and the representation of the constant at run time 
expanding it to 5 characters.

I can live with this issue, but is this what is supposed to happen?

Cheers - Bill

-----------------------------------------------------------------------
Bill Frantz        | The nice thing about         | Periwinkle
(408)356-8506      | standards is there are so    | 16345 
Englewood Ave
www.pwpconsult.com | many to choose from. - anon  | Los Gatos, 
CA 95032


_______________________________________________
e-lang mailing list
[email protected]
http://www.eros-os.org/mailman/listinfo/e-lang