www.digitalmars.com         C & C++   DMDScript  

digitalmars.D - byte + single-bit-operations

reply "Alexander Panek" <alexander.panek brainsware.org> writes:
hello,

just wondered if there`s a common way to edit single bits in a byte? in C  
and C++ there were only crappy 'workarounds', but as D has a single-bit  
datatype it would be great to have an operator to read/write single bits  
in a byte-variable.

thanks for help,
alex

-- 
Using Opera's revolutionary e-mail client: http://www.opera.com/mail/
Mar 15 2005
parent reply =?ISO-8859-15?Q?Anders_F_Bj=F6rklund?= <afb algonet.se> writes:
Alexander Panek wrote:


 just wondered if there`s a common way to edit single bits in a byte? in 
 C  and C++ there were only crappy 'workarounds', but as D has a 
 single-bit  datatype it would be great to have an operator to read/write 
 single bits  in a byte-variable.



Use the bit[] type. Just be aware that it always aligns the memory used
to even powers of 32 bits (4 bytes), if unioned with for instance byte.

See http://www.prowiki.org/wiki4d/wiki.cgi?BitsAndBools

--anders
Mar 15 2005
parent reply "Alexander Panek" <alexander.panek brainsware.org> writes:
On Tue, 15 Mar 2005 21:23:23 +0100, Anders F Björklund <afb algonet.se>  
wrote:


 Alexander Panek wrote:


 just wondered if there`s a common way to edit single bits in a byte? in  
 C  and C++ there were only crappy 'workarounds', but as D has a  
 single-bit  datatype it would be great to have an operator to  
 read/write single bits  in a byte-variable.



 Use the bit[] type. Just be aware that it always aligns the memory used
 to even powers of 32 bits (4 bytes), if unioned with for instance byte.

 See http://www.prowiki.org/wiki4d/wiki.cgi?BitsAndBools

 --anders



Well, I use a bit-array. But it seems like the compiler does not want me  
to do this :o !

I use this code to copy a bit out of a byte into a bit-variable:

[code]
while(offset_ascii < 8) {
	result[offset_sms] = (b & (true << offset_ascii++)) << offset_sms++; //  
line 18
}
[/code]

DMD gives me this error:
" smscode.d(18): cannot implicitly convert expression (cast(int)(b) & 1 <<  
offset_ascii++) of type int to bit "

Thanks,
an hopefully Alex ;)

-- 
Using Opera's revolutionary e-mail client: http://www.opera.com/mail/
Mar 15 2005
next sibling parent reply =?ISO-8859-15?Q?Anders_F_Bj=F6rklund?= <afb algonet.se> writes:
Alexander Panek wrote:


 I use this code to copy a bit out of a byte into a bit-variable:
 
 [code]
 while(offset_ascii < 8) {
     result[offset_sms] = (b & (true << offset_ascii++)) << offset_sms++; 
 //  line 18
 }
 [/code]
 
 DMD gives me this error:
 " smscode.d(18): cannot implicitly convert expression (cast(int)(b) & 1 
 <<  offset_ascii++) of type int to bit "



Please pick either of bit-arrays or bit-shifting, but not both at once?

That is, either:

bit b; // value
int bitno; // index
bit[] array;

array[bitno] = b;

Or, alternatively:

bit b; // value
int bitno;  // index
byte[] array;

if (b)
   array[bitno >> 3] &= ~(1 << (bitno & 7); // set bit
else
   array[bitno >> 3] |= 1 << (bitno & 7); // clear bit

HTH,
--anders
Mar 15 2005
next sibling parent =?ISO-8859-15?Q?Anders_F_Bj=F6rklund?= <afb algonet.se> writes:
Anders F Björklund wrote:


 if (b)
   array[bitno >> 3] &= ~(1 << (bitno & 7); // set bit
 else
   array[bitno >> 3] |= 1 << (bitno & 7); // clear bit



Darnit. Seems to be typo-day today. :-(

   if (b)
     array[bitno >> 3] |= 1 << (bitno & 7); // set bit
   else
     array[bitno >> 3] &= ~(1 << (bitno & 7); // clear bit

--anders
Mar 15 2005
prev sibling parent reply "Alexander Panek" <alexander.panek brainsware.org> writes:
On Tue, 15 Mar 2005 22:14:34 +0100, Anders F Björklund <afb algonet.se>  
wrote:


 Alexander Panek wrote:


 I use this code to copy a bit out of a byte into a bit-variable:
  [code]
 while(offset_ascii < 8) {
     result[offset_sms] = (b & (true << offset_ascii++)) <<  
 offset_sms++; //  line 18
 }
 [/code]
  DMD gives me this error:
 " smscode.d(18): cannot implicitly convert expression (cast(int)(b) & 1  
 <<  offset_ascii++) of type int to bit "



 Please pick either of bit-arrays or bit-shifting, but not both at once?

 That is, either:

 bit b; // value
 int bitno; // index
 bit[] array;

 array[bitno] = b;

 Or, alternatively:

 bit b; // value
 int bitno;  // index
 byte[] array;

 if (b)
    array[bitno >> 3] &= ~(1 << (bitno & 7); // set bit
 else
    array[bitno >> 3] |= 1 << (bitno & 7); // clear bit

 HTH,
 --anders



I have to use both, because I get a byte-array ( byte[] ) and I don`t know  
any other way to get the first 7 bits out of every byte.

That`s the whole function:


bit [] toSMSCode(byte[] ascii)
{
	bit [] result;
	int offset_sms, offset_ascii;
	
	offset_sms = 0;
	offset_ascii = 0;
	
	foreach(byte b; ascii)
	{
		while(offset_ascii < 8)
		{
			result[offset_sms] = (b & (true << offset_ascii++)) << offset_sms++;
		}
		offset_ascii = 0;
	}
	
	return result;
}

It (should) converts ASCII-strings to SMS-coded-strings (only 7bits of 8  
used, so the string gets smaller).

Thanks for giving trouble, Alex

-- 
Using Opera's revolutionary e-mail client: http://www.opera.com/mail/
Mar 15 2005
parent =?ISO-8859-15?Q?Anders_F_Bj=F6rklund?= <afb algonet.se> writes:
Alexander Panek wrote:


 I have to use both, because I get a byte-array ( byte[] ) and I don`t 
 know  any other way to get the first 7 bits out of every byte.



OK, thanks for clearifying...


 That`s the whole function:




             result[offset_sms] = (b & (true << offset_ascii++)) << 
 offset_sms++;



Assuming the other logic is OK, this line should be:

   result[offset_sms++] = (b & (1 << offset_ascii++)) != 0;

I think that you can take it from here, if it's not...
(for instance, you need to set a result.length first)

--anders
Mar 15 2005
prev sibling next sibling parent reply "Regan Heath" <regan netwin.co.nz> writes:
On Tue, 15 Mar 2005 22:05:36 +0100, Alexander Panek  
<alexander.panek brainsware.org> wrote:

 On Tue, 15 Mar 2005 21:23:23 +0100, Anders F Björklund <afb algonet.se>  
 wrote:


 Alexander Panek wrote:


 just wondered if there`s a common way to edit single bits in a byte?  
 in C  and C++ there were only crappy 'workarounds', but as D has a  
 single-bit  datatype it would be great to have an operator to  
 read/write single bits  in a byte-variable.



 Use the bit[] type. Just be aware that it always aligns the memory used
 to even powers of 32 bits (4 bytes), if unioned with for instance byte.

 See http://www.prowiki.org/wiki4d/wiki.cgi?BitsAndBools

 --anders



 Well, I use a bit-array. But it seems like the compiler does not want me  
 to do this :o !

 I use this code to copy a bit out of a byte into a bit-variable:

 [code]
 while(offset_ascii < 8) {
 	result[offset_sms] = (b & (true << offset_ascii++)) << offset_sms++; //  
 line 18
 }
 [/code]

 DMD gives me this error:
 " smscode.d(18): cannot implicitly convert expression (cast(int)(b) & 1  
 << offset_ascii++) of type int to bit "

 Thanks,
 an hopefully Alex ;)



You might find this helpful.

# import std.stdio;
#
# void main()
# {
# 	byte b;
# 	short s;
# 	int i;
# 	long l;
# 	int[] ia;
# 	bit[] a;
# 	
# 	//slice a byte
# 	a = (cast(bit*)&b)[0..8];
# 	a[0] = 1; //2^0 == 1
# 	a[6] = 1; //2^6 == 64
# 	writefln("byte(",b,")"); //65
# 	
# 	//slice a short
# 	a = (cast(bit*)&s)[0..16];
# 	a[3] = 1;  //2^3  == 8
# 	a[14] = 1; //2^14 == 16384
# 	writefln("short(",s,")"); //16392
# 	
# 	//slice an int
# 	a = (cast(bit*)&i)[0..32];
# 	a[5] = 1;  //2^5  == 32
# 	a[22] = 1; //2^22 == 4194304
# 	writefln("int(",i,")"); //4194336
# 	
# 	//slice a long
# 	a = (cast(bit*)&l)[0..64];
# 	a[25] = 1; //2^25 == 33554432
# 	a[45] = 1; //2^45 == 35184372088832
# 	writefln("long(",l,")"); //35184405643264
# 	//WTF?! this prints 35184372088832
# 	writefln(a[25]);
# 	writefln(a[45]);
# 			
# 	//or even slice an array of 10 ints
# 	ia.length = 10;
# 	//set each int to it's index
# 	foreach(int i, inout int v; ia) v = i;
# 	
# 	a = (cast(bit*)ia.ptr)[0..32*ia.length];
# 	a[310] = 1; //ia[9] bit #22, 2^22 = 4194304
# 	foreach(int v; ia) writefln(v); //4194313 (2^22 + 9(it's index))
# }

Basically you can create a bit[] by slicing any location in memory, you  
just have to know it's starting location and how big the memory is, in  
bits.

Something I don't understand is happening with the 'long', can someone  
explain this to me please. :)

Re-writing your code, assuming 'result' is a bit[] and 'b' is a byte:
   result = (cast(bit*)b)[0..8];

Regan
Mar 15 2005
parent reply "Alexander Panek" <alexander.panek brainsware.org> writes:
On Wed, 16 Mar 2005 10:17:11 +1300, Regan Heath <regan netwin.co.nz> wrote:

 # 	//slice a byte
 # 	a = (cast(bit*)&b)[0..8];
 # 	a[0] = 1; //2^0 == 1
 # 	a[6] = 1; //2^6 == 64
 # 	writefln("byte(",b,")"); //65

 Basically you can create a bit[] by slicing any location in memory, you  
 just have to know it's starting location and how big the memory is, in  
 bits.

 Something I don't understand is happening with the 'long', can someone  
 explain this to me please. :)

 Re-writing your code, assuming 'result' is a bit[] and 'b' is a byte:
    result = (cast(bit*)b)[0..8];

 Regan



You`re my hero! Saved my day! :)

Thank you very much, Alex

-- 
Using Opera's revolutionary e-mail client: http://www.opera.com/mail/
Mar 15 2005
parent reply "Regan Heath" <regan netwin.co.nz> writes:
On Tue, 15 Mar 2005 22:27:44 +0100, Alexander Panek  
<alexander.panek brainsware.org> wrote:

 On Wed, 16 Mar 2005 10:17:11 +1300, Regan Heath <regan netwin.co.nz>  
 wrote:

 # 	//slice a byte
 # 	a = (cast(bit*)&b)[0..8];
 # 	a[0] = 1; //2^0 == 1
 # 	a[6] = 1; //2^6 == 64
 # 	writefln("byte(",b,")"); //65

 Basically you can create a bit[] by slicing any location in memory, you  
 just have to know it's starting location and how big the memory is, in  
 bits.

 Something I don't understand is happening with the 'long', can someone  
 explain this to me please. :)

 Re-writing your code, assuming 'result' is a bit[] and 'b' is a byte:
    result = (cast(bit*)b)[0..8];

 Regan



 You`re my hero! Saved my day! :)
 Thank you very much, Alex



After reading your other post, about "getting the first 7 bits" from each  
byte (I've written some sms decode/encode routines in C BTW) you could do  
something like...

bit[] result;
bit[] part;

foreach(byte b; raw_data) {
	part = (cast(bit*)byte)[0..7];
	result ~= part;
}

or similar to grab 7 bits from each byte (whether it's the correct 7 bits  
or not I haven't thought about, I'll leave that to you).

Regan
Mar 15 2005
parent reply "Regan Heath" <regan netwin.co.nz> writes:
On Wed, 16 Mar 2005 10:59:21 +1300, Regan Heath <regan netwin.co.nz> wrote:

 On Tue, 15 Mar 2005 22:27:44 +0100, Alexander Panek  
 <alexander.panek brainsware.org> wrote:

 On Wed, 16 Mar 2005 10:17:11 +1300, Regan Heath <regan netwin.co.nz>  
 wrote:

 # 	//slice a byte
 # 	a = (cast(bit*)&b)[0..8];
 # 	a[0] = 1; //2^0 == 1
 # 	a[6] = 1; //2^6 == 64
 # 	writefln("byte(",b,")"); //65

 Basically you can create a bit[] by slicing any location in memory,  
 you just have to know it's starting location and how big the memory  
 is, in bits.

 Something I don't understand is happening with the 'long', can someone  
 explain this to me please. :)

 Re-writing your code, assuming 'result' is a bit[] and 'b' is a byte:
    result = (cast(bit*)b)[0..8];

 Regan



 You`re my hero! Saved my day! :)
 Thank you very much, Alex



 After reading your other post, about "getting the first 7 bits" from  
 each byte (I've written some sms decode/encode routines in C BTW) you  
 could do something like...

 bit[] result;
 bit[] part;

 foreach(byte b; raw_data) {
 	part = (cast(bit*)byte)[0..7];
 	result ~= part;
 }

 or similar to grab 7 bits from each byte (whether it's the correct 7  
 bits or not I haven't thought about, I'll leave that to you).



To be honest though I haven't tried the code above. bit[] is a strange  
beast, not like other array types in some situations... one would hope the  
above code works though it may not.

Regan
Mar 15 2005
next sibling parent reply "Alexander Panek" <alexander.panek brainsware.org> writes:
On Wed, 16 Mar 2005 11:02:53 +1300, Regan Heath <regan netwin.co.nz> wrote:


 On Wed, 16 Mar 2005 10:59:21 +1300, Regan Heath <regan netwin.co.nz>  
 wrote:

 On Tue, 15 Mar 2005 22:27:44 +0100, Alexander Panek  
 <alexander.panek brainsware.org> wrote:

 On Wed, 16 Mar 2005 10:17:11 +1300, Regan Heath <regan netwin.co.nz>  
 wrote:

 # 	//slice a byte
 # 	a = (cast(bit*)&b)[0..8];
 # 	a[0] = 1; //2^0 == 1
 # 	a[6] = 1; //2^6 == 64
 # 	writefln("byte(",b,")"); //65

 Basically you can create a bit[] by slicing any location in memory,  
 you just have to know it's starting location and how big the memory  
 is, in bits.

 Something I don't understand is happening with the 'long', can  
 someone explain this to me please. :)

 Re-writing your code, assuming 'result' is a bit[] and 'b' is a byte:
    result = (cast(bit*)b)[0..8];

 Regan



 You`re my hero! Saved my day! :)
 Thank you very much, Alex



 After reading your other post, about "getting the first 7 bits" from  
 each byte (I've written some sms decode/encode routines in C BTW) you  
 could do something like...

 bit[] result;
 bit[] part;

 foreach(byte b; raw_data) {
 	part = (cast(bit*)byte)[0..7];
 	result ~= part;
 }

 or similar to grab 7 bits from each byte (whether it's the correct 7  
 bits or not I haven't thought about, I'll leave that to you).



 To be honest though I haven't tried the code above. bit[] is a strange  
 beast, not like other array types in some situations... one would hope  
 the above code works though it may not.

 Regan



My code is quite similar to yours:


 	foreach(byte b; ascii)
 	{
 		sms[offset_sms..(offset_sms + 6)] = (cast(bit*)&b)[0..6];
 		offset_sms += 7;
 	}



.. it *should* copy the first 7 bits of a byte into the bit-array, but it  
allways ends up either with an ArrayBoundsError, or something like that:  
"Error: lengths don't match for array copy".
I already tried to modify the length of "bit [] sms" dynamically, so it`s  
allways big enough to get the next 7 bits -> same errors. (Well, the first  
2 steps of the loop are working, but then it gives the ArrayBoundsError.)

Alex

-- 
Using Opera's revolutionary e-mail client: http://www.opera.com/mail/
Mar 15 2005
parent reply "Alexander Panek" <alexander.panek brainsware.org> writes:
On Wed, 16 Mar 2005 07:16:00 +0100, Alexander Panek  
<alexander.panek brainsware.org> wrote:


 On Wed, 16 Mar 2005 11:02:53 +1300, Regan Heath <regan netwin.co.nz>  
 wrote:


 On Wed, 16 Mar 2005 10:59:21 +1300, Regan Heath <regan netwin.co.nz>  
 wrote:

 On Tue, 15 Mar 2005 22:27:44 +0100, Alexander Panek  
 <alexander.panek brainsware.org> wrote:

 On Wed, 16 Mar 2005 10:17:11 +1300, Regan Heath <regan netwin.co.nz>  
 wrote:

 # 	//slice a byte
 # 	a = (cast(bit*)&b)[0..8];
 # 	a[0] = 1; //2^0 == 1
 # 	a[6] = 1; //2^6 == 64
 # 	writefln("byte(",b,")"); //65

 Basically you can create a bit[] by slicing any location in memory,  
 you just have to know it's starting location and how big the memory  
 is, in bits.

 Something I don't understand is happening with the 'long', can  
 someone explain this to me please. :)

 Re-writing your code, assuming 'result' is a bit[] and 'b' is a byte:
    result = (cast(bit*)b)[0..8];

 Regan



 You`re my hero! Saved my day! :)
 Thank you very much, Alex



 After reading your other post, about "getting the first 7 bits" from  
 each byte (I've written some sms decode/encode routines in C BTW) you  
 could do something like...

 bit[] result;
 bit[] part;

 foreach(byte b; raw_data) {
 	part = (cast(bit*)byte)[0..7];
 	result ~= part;
 }

 or similar to grab 7 bits from each byte (whether it's the correct 7  
 bits or not I haven't thought about, I'll leave that to you).



 To be honest though I haven't tried the code above. bit[] is a strange  
 beast, not like other array types in some situations... one would hope  
 the above code works though it may not.

 Regan



 My code is quite similar to yours:


 	foreach(byte b; ascii)
 	{
 		sms[offset_sms..(offset_sms + 6)] = (cast(bit*)&b)[0..6];
 		offset_sms += 7;
 	}



 .. it *should* copy the first 7 bits of a byte into the bit-array, but  
 it allways ends up either with an ArrayBoundsError, or something like  
 that: "Error: lengths don't match for array copy".
 I already tried to modify the length of "bit [] sms" dynamically, so  
 it`s allways big enough to get the next 7 bits -> same errors. (Well,  
 the first 2 steps of the loop are working, but then it gives the  
 ArrayBoundsError.)

 Alex



Rearranged my code:


 bit [] enSMS(byte [] ascii)
 {
 	bit [] sms;
 	uint i = 1;
 	//sms.length = (ascii.length * 7);

 	foreach(byte b; ascii)
 	{
 		writefln("encoding .. byte:", i);
 		if(i == ascii.length) {
 			writefln("last byte!");
 			return sms[];
 		}
 		sms ~= (cast(bit*)&b)[0..6];
 		i++;
 	}
 	
 	return sms[];
 }



Works fine .. until the loop is finished, or the return is made - I don`t  
know.

Output:

Y:\Programming\D\SMS Codierung>smscode
encoding .. byte:1
encoding .. byte:2
encoding .. byte:3
encoding .. byte:4
encoding .. byte:5
encoding .. byte:6
encoding .. byte:7
encoding .. byte:8
encoding .. byte:9
encoding .. byte:10
encoding .. byte:11
encoding .. byte:12
encoding .. byte:13
encoding .. byte:14
encoding .. byte:15
encoding .. byte:16
encoding .. byte:17
encoding .. byte:18
encoding .. byte:19
encoding .. byte:20
encoding .. byte:21
encoding .. byte:22
encoding .. byte:23
encoding .. byte:24
encoding .. byte:25
last byte!
Error: lengths don't match for array copy


..well, I`m helpless :) .

Thanks, Alex



-- 
Using Opera's revolutionary e-mail client: http://www.opera.com/mail/
Mar 15 2005
parent reply "Regan Heath" <regan netwin.co.nz> writes:
On Wed, 16 Mar 2005 07:33:18 +0100, Alexander Panek  
<alexander.panek brainsware.org> wrote:

 On Wed, 16 Mar 2005 07:16:00 +0100, Alexander Panek  
 <alexander.panek brainsware.org> wrote:


 On Wed, 16 Mar 2005 11:02:53 +1300, Regan Heath <regan netwin.co.nz>  
 wrote:


 On Wed, 16 Mar 2005 10:59:21 +1300, Regan Heath <regan netwin.co.nz>  
 wrote:

 On Tue, 15 Mar 2005 22:27:44 +0100, Alexander Panek  
 <alexander.panek brainsware.org> wrote:

 On Wed, 16 Mar 2005 10:17:11 +1300, Regan Heath <regan netwin.co.nz>  
 wrote:

 # 	//slice a byte
 # 	a = (cast(bit*)&b)[0..8];
 # 	a[0] = 1; //2^0 == 1
 # 	a[6] = 1; //2^6 == 64
 # 	writefln("byte(",b,")"); //65

 Basically you can create a bit[] by slicing any location in memory,  
 you just have to know it's starting location and how big the memory  
 is, in bits.

 Something I don't understand is happening with the 'long', can  
 someone explain this to me please. :)

 Re-writing your code, assuming 'result' is a bit[] and 'b' is a  
 byte:
    result = (cast(bit*)b)[0..8];

 Regan



 You`re my hero! Saved my day! :)
 Thank you very much, Alex



 After reading your other post, about "getting the first 7 bits" from  
 each byte (I've written some sms decode/encode routines in C BTW) you  
 could do something like...

 bit[] result;
 bit[] part;

 foreach(byte b; raw_data) {
 	part = (cast(bit*)byte)[0..7];
 	result ~= part;
 }

 or similar to grab 7 bits from each byte (whether it's the correct 7  
 bits or not I haven't thought about, I'll leave that to you).



 To be honest though I haven't tried the code above. bit[] is a strange  
 beast, not like other array types in some situations... one would hope  
 the above code works though it may not.

 Regan



 My code is quite similar to yours:


 	foreach(byte b; ascii)
 	{
 		sms[offset_sms..(offset_sms + 6)] = (cast(bit*)&b)[0..6];
 		offset_sms += 7;
 	}



 .. it *should* copy the first 7 bits of a byte into the bit-array, but  
 it allways ends up either with an ArrayBoundsError, or something like  
 that: "Error: lengths don't match for array copy".
 I already tried to modify the length of "bit [] sms" dynamically, so  
 it`s allways big enough to get the next 7 bits -> same errors. (Well,  
 the first 2 steps of the loop are working, but then it gives the  
 ArrayBoundsError.)

 Alex



 Rearranged my code:


 bit [] enSMS(byte [] ascii)
 {
 	bit [] sms;
 	uint i = 1;
 	//sms.length = (ascii.length * 7);

 	foreach(byte b; ascii)
 	{
 		writefln("encoding .. byte:", i);
 		if(i == ascii.length) {
 			writefln("last byte!");
 			return sms[];
 		}
 		sms ~= (cast(bit*)&b)[0..6];
 		i++;
 	}
 	
 	return sms[];
 }



 Works fine .. until the loop is finished, or the return is made - I  
 don`t know.

 Output:

 Y:\Programming\D\SMS Codierung>smscode
 encoding .. byte:1
 encoding .. byte:2
 encoding .. byte:3
 encoding .. byte:4
 encoding .. byte:5
 encoding .. byte:6
 encoding .. byte:7
 encoding .. byte:8
 encoding .. byte:9
 encoding .. byte:10
 encoding .. byte:11
 encoding .. byte:12
 encoding .. byte:13
 encoding . byte:14
 encoding .. byte:15
 encoding .. byte:16
 encoding .. byte:17
 encoding .. byte:18
 encoding .. byte:19
 encoding .. byte:20
 encoding .. byte:21
 encoding .. byte:22
 encoding .. byte:23
 encoding .. byte:24
 encoding .. byte:25
 last byte!
 Error: lengths don't match for array copy


 ..well, I`m helpless :) .



Odd. What happes if you change "return sms[];" to "return sms;" ?

I am having different trouble, no array bounds errors for me, but...

import std.stdio;

void main()
{
	byte[] ascii = cast(byte[])"0000";
	bit[] sms;
	bit[] tmp;
	
	writef("ASC:");
	foreach(int i, bit bb; (cast(bit*)ascii.ptr)[0..32]) {
		if (i && i%8 == 0) writef(" ");
		writef("%d",bb);
	}
	writefln("");
	
	foreach(int j, byte b; ascii) {
		tmp = (cast(bit*)&b)[0..7];		
		
		writef("TMP:");
		for(int k = 0; k < j; k++) writef("         ");
		foreach(int i, bit b; tmp) writef("%d",b);
		writefln("");
		
		sms ~= tmp;
		
		writef("SMS:");
		foreach(int i, bit b; sms) {
			if (i && i%7 == 0) writef("  ");
			writef("%d",b);
		}
		writefln("");
	}
	writefln("");
}

output:

ASC:00001100 00001100 00001100 00001100
TMP:0000110
SMS:0000110
TMP:         0000110
SMS:0000110  0010111
TMP:                  0000110
SMS:0000110  0010111  0010000
TMP:                           0000110
SMS:0000110  0010111  0010000  0100000

'tmp' is appended to 'sms' each time round the loop, but something is  
going very wrong.

Anyone got any ideas?

Regan
Mar 16 2005
parent "Alexander Panek" <alexander.panek brainsware.org> writes:
On Wed, 16 Mar 2005 22:10:30 +1300, Regan Heath <regan netwin.co.nz> wrote:


 Odd. What happes if you change "return sms[];" to "return sms;" ?

 I am having different trouble, no array bounds errors for me, but...

 import std.stdio;

 void main()
 {
 	byte[] ascii = cast(byte[])"0000";
 	bit[] sms;
 	bit[] tmp;
 	
 	writef("ASC:");
 	foreach(int i, bit bb; (cast(bit*)ascii.ptr)[0..32]) {
 		if (i && i%8 == 0) writef(" ");
 		writef("%d",bb);
 	}
 	writefln("");
 	
 	foreach(int j, byte b; ascii) {
 		tmp = (cast(bit*)&b)[0..7];		
 		
 		writef("TMP:");
 		for(int k = 0; k < j; k++) writef("         ");
 		foreach(int i, bit b; tmp) writef("%d",b);
 		writefln("");
 		
 		sms ~= tmp;
 		
 		writef("SMS:");
 		foreach(int i, bit b; sms) {
 			if (i && i%7 == 0) writef("  ");
 			writef("%d",b);
 		}
 		writefln("");
 	}
 	writefln("");
 }

 output:

 ASC:00001100 00001100 00001100 00001100
 TMP:0000110
 SMS:0000110
 TMP:         0000110
 SMS:0000110  0010111
 TMP:                  0000110
 SMS:0000110  0010111  0010000
 TMP:                           0000110
 SMS:0000110  0010111  0010000  0100000

 'tmp' is appended to 'sms' each time round the loop, but something is  
 going very wrong.

 Anyone got any ideas?

 Regan



It was "return sms;" before, just tried out if it helps. ;)

For some reason the encoded string seems to be bigger than the  
ascii-string here:

string has 6bytes
encoding .. byte:1
0000110\
  0000110
encoding .. byte:2
1000110\
  00001100 000000
encoding .. byte:3
0000110\
  00001100 00000000 00000
encoding .. byte:4
1000110\
  00001100 00000000 00000000 0000
encoding .. byte:5
0000110\
  00001100 00000000 00000000 00000000 001
encoding .. byte:6 (last byte!)
1000110\
  00001100 00000000 00000000 00000000 00111111 01Error: lengths don't match  
for array copy
  <--------------------------------------------->
                  = 5bytes + 2 bits

current code:

//  
=========================================================================
bit [] enSMS(byte [] ascii)
{
	bit [] sms, tmp;
	uint i = 1;
	//sms.length = (ascii.length * 7);
	
	writef("string(", cast(char[])ascii, ") has ", ascii.length, "bytes");

	foreach(byte by; ascii)
	{
		writef("\nencoding .. byte:", i);
		if(i == ascii.length)
			writefln(" (last byte!)");
		else
			writefln("");

		tmp = (cast(bit*)&by)[0..7];

		foreach(bit b; tmp) {
			writef("%d",b);
		}

		sms ~= tmp;

		foreach(int j, bit b; sms) {
			if(j%50 == 0) writefln("\\");
			if(j%8 == 0)writef(" ");
			writef("%d",b);
		}
		i++;
	}

	return sms;
}
//  
=========================================================================

So, as you see, I`ve got the same problem.

Alex
-- 
Using Opera's revolutionary e-mail client: http://www.opera.com/mail/
Mar 16 2005
prev sibling parent Derek Parnell <derek psych.ward> writes:
On Wed, 16 Mar 2005 11:02:53 +1300, Regan Heath wrote:


 On Wed, 16 Mar 2005 10:59:21 +1300, Regan Heath <regan netwin.co.nz> wrote:

 On Tue, 15 Mar 2005 22:27:44 +0100, Alexander Panek  
 <alexander.panek brainsware.org> wrote:

 On Wed, 16 Mar 2005 10:17:11 +1300, Regan Heath <regan netwin.co.nz>  
 wrote:

 # 	//slice a byte
 # 	a = (cast(bit*)&b)[0..8];
 # 	a[0] = 1; //2^0 == 1
 # 	a[6] = 1; //2^6 == 64
 # 	writefln("byte(",b,")"); //65

 Basically you can create a bit[] by slicing any location in memory,  
 you just have to know it's starting location and how big the memory  
 is, in bits.

 Something I don't understand is happening with the 'long', can someone  
 explain this to me please. :)

 Re-writing your code, assuming 'result' is a bit[] and 'b' is a byte:
    result = (cast(bit*)b)[0..8];

 Regan



 You`re my hero! Saved my day! :)
 Thank you very much, Alex



 After reading your other post, about "getting the first 7 bits" from  
 each byte (I've written some sms decode/encode routines in C BTW) you  
 could do something like...

 bit[] result;
 bit[] part;

 foreach(byte b; raw_data) {
 	part = (cast(bit*)byte)[0..7];
 	result ~= part;
 }

 or similar to grab 7 bits from each byte (whether it's the correct 7  
 bits or not I haven't thought about, I'll leave that to you).



 To be honest though I haven't tried the code above. bit[] is a strange  
 beast, not like other array types in some situations... one would hope the  
 above code works though it may not.
 
 Regan



I think that one difference is that with non-bit arrays, element[x] has a
lower RAM address than element[x+1], but it is the reverse with bit arrays.
For example, if you have a bit array of 32 bits, the elements 0-7 inclusive
might have an address of 0xFE78 (for one byte), elements 8-15 inclusive
will have an address of 0xFE77, elements 16-23 will have the address
0xFE76, and the last octet will have the address 0xFE75. Such that if you
set "element[31]=1", it means that the left-most bit at address 0xFE75 will
be set, and "element[0]=1" will set the rightmost bit at address 0xFE78.

-- 
Derek Parnell
Melbourne, Australia
16/03/2005 9:43:04 PM
Mar 16 2005
prev sibling parent Nrgyzer <nrgyzer gmail.com> writes:
Alexander Panek Wrote:


 On Tue, 15 Mar 2005 21:23:23 +0100, Anders F Björklund <afb algonet.se>  
 wrote:
 

 Alexander Panek wrote:


 just wondered if there`s a common way to edit single bits in a byte? in  
 C  and C++ there were only crappy 'workarounds', but as D has a  
 single-bit  datatype it would be great to have an operator to  
 read/write single bits  in a byte-variable.



 Use the bit[] type. Just be aware that it always aligns the memory used
 to even powers of 32 bits (4 bytes), if unioned with for instance byte.

 See http://www.prowiki.org/wiki4d/wiki.cgi?BitsAndBools

 --anders



 Well, I use a bit-array. But it seems like the compiler does not want me  
 to do this :o !
 
 I use this code to copy a bit out of a byte into a bit-variable:
 
 [code]
 while(offset_ascii < 8) {
 	result[offset_sms] = (b & (true << offset_ascii++)) << offset_sms++; //  
 line 18
 }
 [/code]
 
 DMD gives me this error:
 " smscode.d(18): cannot implicitly convert expression (cast(int)(b) & 1 <<  
 offset_ascii++) of type int to bit "
 
 Thanks,
 an hopefully Alex ;)
 
 -- 
 Using Opera's revolutionary e-mail client: http://www.opera.com/mail/



I tried a bit... I can't get it to work with bud... only dmd will compile the
source. When I first compile with dmd and then with bud, it works with bud, too
- but only until I didn't use the -clean flag when I build the exe-file.
May 07 2010