CharToInt

[Paulo Zemek]

My real problem here is that the autor of the code didn't test EVERY possible character. After all, using default to return an useless and invalid value is not a good idea. It will be much better to: case 'G': throw new Exception("G is an invalid value."); . . . case 'Z': throw new Exception("Z is an invalid value."); LOL

[OriginalGriff]

If you are going to be paid by the line, then you should continue this with lower case, punctuation and the other oddities available in the local character set... It would have been far to easy too write

default:
{
throw new Exception(string.Format("\"{0}\" is an invalid value", pcChar));
}

So we won't mention it. Or at least I won't. Oh bum.

modified on Sunday, April 19, 2009 5:44 AM

[OriginalGriff]

Except I just realized that to fit with his "coding style" I should have suggested

default:
{
throw new Exception(string.Format("\"{0}\" is an invalid value", pcChar));
return pcChar;
break;
}

I feel better now. Unclean, but better.

[Lutoslaw]

I'm doing A LOT of such stuff nowadays. In assembly. (AT&T syntax)

# READ (char* bufer, int bufer_len)

Reads a number from an ascii buffer and returns an actual number in eax

PARAMETERS

1. Address to the buffer

2. The string's length

Used registers:

esi - contains all read characters

ecx - index inside the buffer

ebx - used in a conversion process char -> int

edx - address to the buffer

eax - the result

.globl read
.type read, @function
read:
.equ DIGIT_0,'0'
.equ DIGIT_9,'9'
.equ CASE_A,'A'
.equ CASE_F,'F'
.equ CASE_a,'a'
.equ CASE_f,'f'
.equ ONE_DIGIT_MASK, 0x0000000F

mov 8(%esp), %esi # save number of read chars

mov $0, %ecx
mov $0, %eax # there was the zero at the beginnig of the Universe
cmp %ecx, %esi # empty string case
je read\_done

read_loop:
mov $0, %ebx

mov 4(%esp), %edx
#mov (%edx), %edx	
mov (%edx, %ecx, 1), %bl # take the next char
cmp $DIGIT\_9, %bl
jle read\_0to9

read_AtoF:
cmp $CASE_a, %bl
jge read_AtoF_lowercase
sub $CASE_A, %bl # substitute'A'
jmp read_AtoF_continue
read_AtoF_lowercase:
sub $CASE_a, %bl # substitute 'A'
read_AtoF_continue:
add $10, %bl # add 10 to get the correct value
jmp read_char_done

read_0to9:
cmp $DIGIT_0, %bl
jl read_done
sub $DIGIT_0, %bl # substitute '0'
jmp read_char_done

read_char_done:
# (bl is a lower part of ebx)
add %ebx, %eax # ebx is a read-to-eat int

inc %ecx # increment the index
cmp %ecx, %esi # check if we're done
je read\_done

shl $4, %eax # shift the number by one hex digit capacity.

jmp read\_loop

read_done:
ret

:omg:

Greetings - Gajatko Portable.NET is part of DotGNU, a project to build a complete Free Software replacement for .NET - a system that truly belongs to the developers.

[Paulo Zemek]

default: { throw new Exception(string.Format("\"{0}\" is an invalid value", pcChar)); return pcChar; break; } I must agree. This is a much better option. But I will keep the idea of checking each value, and ALSO use this pattern. So: case 'G': throw new Exception(string.Format("\"{0}\" is an invalid value", pcChar)); return pcChar; break; case 'H': throw new Exception(string.Format("\"{0}\" is an invalid value", pcChar)); return pcChar; break; As this is the original pattern from 0 to 8 (9 is NOT a number, LOL).

[riced]

Paulo Zemek wrote:

As this is the original pattern from 0 to 8 (9 is NOT a number, LOL)

I though 6 was not a number http://www.youtube.com/watch?v=29JewlGsYxs[^] :-D

Regards David R

[Jeroen De Dauw]

What about ... ? return ( "0123456789ABCDEF".IndexOf ( pcChar.ToUpper ) ) ;

My little forums: http://code.bn2vs.com 70 72 6F 67 72 61 6D 6D 69 6E 67 20 34 20 6C 69 66 65!

[Jeroen De Dauw]

Why not check if the char equal 10? And since your getting payed by line go on till you reach the max value of an int64? :D Cheers

My little forums: http://code.bn2vs.com 70 72 6F 67 72 61 6D 6D 69 6E 67 20 34 20 6C 69 66 65!

[Yusuf]

riced wrote:

What's he/she got against the number 9?

Oh, nothing. 9 will get flipped to 6 automagically

Yusuf Oh didn't you notice, analogous to square roots, they recently introduced rectangular, circular, and diamond roots to determine the size of the corresponding shapes when given the area. Luc Pattyn[^]

[bulg]

C#? how about some good ol'

ans = c_val > '9' ? (10 + c_val - 'A') : c_val-'0';

[Brady Kelly]

What's with all the 'break' statements after 8? He didn't need them before? Or does he hope that Convert.ToInt32 does some voodoo that never returns to the switch block?

[Rajesh R Subramanian]

Brilliant! Aren't we glad English has only 26 letters? I'd like to see the Chinese version of the code (thousands of letters).

It is a crappy thing, but it's life -^ Carlo Pallini