Integer and character manipulation

[ruroma]

Hello,

I need some help with the following:

1) I need to split a 16bit INT into two 8bit characters, and then be
able to join these two characters to form again the original 16bit
integer.

For example with values expressed in binary:

int n= 000000001111111;
--> char a=00000000;
--> char b=11111111;

and then be able to join a and b to form back n.
2) And last, I need to convert an integer into a string.
For example,

int n=3.142 ---> char string[5] = ['3','.','1','4','2']
Thank you very much!! I've tried working with casting but as far as I
know they don't seem to be the answer. I've also searched numerous
tutorials and web pages, but nothing productive. Any help or
suggestion will be highly appreciated. Thank you very much!!

[aegis]

ru****@gmail.com wrote:

Hello,

I need some help with the following:

1) I need to split a 16bit INT into two 8bit characters, and then be
able to join these two characters to form again the original 16bit
integer.

For example with values expressed in binary:

int n= 000000001111111;
--> char a=00000000;
--> char b=11111111;

and then be able to join a and b to form back n.

You can use a mask for that.

a = n & 0xff00;
b = n & 0x00ff;

And use 'unsigned char' instead of 'char'. 'char' can be either
equivalent to 'unsigned char' or 'signed char'. If you rely on signed
char and use some value that cannot be represented with signed char,
then you may invoke implementation defined behavior or raise an
implementation defined signal.

2) And last, I need to convert an integer into a string.
For example,

int n=3.142 ---> char string[5] = ['3','.','1','4','2']

'int' is a part of the standard integer types and can't be used to
represent rationals. So you should either use an object type of float
or double. To produce a string from a given value of type float or
double, use sprintf/snprintf

Thank you very much!! I've tried working with casting but as far as I
know they don't seem to be the answer. I've also searched numerous
tutorials and web pages, but nothing productive. Any help or
suggestion will be highly appreciated. Thank you very much!!

casting is not a magical thing. The reason casting exists in the
language is for a few corner cases, namely: %p expecting a (void *)
with *printf family of functions. To compare the return value of mktime
to (time_t)-1 in case of error. For the is*, to* functions/macros.

There may be one or two other places that escape me at the moment but
not much else beyond that. The reason it doesn't produce a string is
because a cast yields a value of a given type, given a value. As there
does not exist a string type in C, then you can't expect that there
will be a string produced by a cast.
--
aegis

[Barry Schwarz]

On 5 Sep 2005 12:03:12 -0700, ru****@gmail.com wrote:

Hello,

I need some help with the following:

1) I need to split a 16bit INT into two 8bit characters, and then be
able to join these two characters to form again the original 16bit
integer.

For example with values expressed in binary:

int n= 000000001111111;
--> char a=00000000;
--> char b=11111111;

and then be able to join a and b to form back n.
Look at the bitwise operators, especially "and" and shift.


2) And last, I need to convert an integer into a string.
For example,

int n=3.142 ---> char string[5] = ['3','.','1','4','2']
3.142 is not an integer value of any kind. Your array is not a
string. Perhaps you'd like to tell us what you really need rather
than provide an incorrect example.


Thank you very much!! I've tried working with casting but as far as I
know they don't seem to be the answer. I've also searched numerous
tutorials and web pages, but nothing productive. Any help or
suggestion will be highly appreciated. Thank you very much!!

You will get a lot more help if you post your code. We won't do you
homework for you.
<<Remove the del for email>>

[Barry Schwarz]

On 5 Sep 2005 13:49:26 -0700, "aegis" <ae***@mad.scientist.com> wrote:

ru****@gmail.com wrote:

Hello,

I need some help with the following:

1) I need to split a 16bit INT into two 8bit characters, and then be
able to join these two characters to form again the original 16bit
integer.

For example with values expressed in binary:

int n= 000000001111111;
--> char a=00000000;
--> char b=11111111;

and then be able to join a and b to form back n.

You can use a mask for that.

a = n & 0xff00;

You are missing a shift. On most systems, the result of your
expression will not fit in a char.
b = n & 0x00ff;

And use 'unsigned char' instead of 'char'. 'char' can be either
equivalent to 'unsigned char' or 'signed char'. If you rely on signed
char and use some value that cannot be represented with signed char,
then you may invoke implementation defined behavior or raise an
implementation defined signal.

2) And last, I need to convert an integer into a string.
For example,

int n=3.142 ---> char string[5] = ['3','.','1','4','2']

'int' is a part of the standard integer types and can't be used to
represent rationals. So you should either use an object type of float
or double. To produce a string from a given value of type float or
double, use sprintf/snprintf

Thank you very much!! I've tried working with casting but as far as I
know they don't seem to be the answer. I've also searched numerous
tutorials and web pages, but nothing productive. Any help or
suggestion will be highly appreciated. Thank you very much!!

casting is not a magical thing. The reason casting exists in the
language is for a few corner cases, namely: %p expecting a (void *)
with *printf family of functions. To compare the return value of mktime
to (time_t)-1 in case of error. For the is*, to* functions/macros.

There may be one or two other places that escape me at the moment but
not much else beyond that. The reason it doesn't produce a string is
because a cast yields a value of a given type, given a value. As there
does not exist a string type in C, then you can't expect that there
will be a string produced by a cast.

<<Remove the del for email>>

[Ian]

ru****@gmail.com wrote:

Hello,

I need some help with the following:

1) I need to split a 16bit INT into two 8bit characters, and then be
able to join these two characters to form again the original 16bit
integer.

For example with values expressed in binary:

int n= 000000001111111;
--> char a=00000000;
--> char b=11111111;

and then be able to join a and b to form back n.

Sounds like homework! Mug up on shifting and masking.
2) And last, I need to convert an integer into a string.
For example,

int n=3.142 ---> char string[5] = ['3','.','1','4','2']

3.142.isn't an int, its a floating point value. Look up formatted output.

Ian

[aegis]

Barry Schwarz wrote:

On 5 Sep 2005 13:49:26 -0700, "aegis" <ae***@mad.scientist.com> wrote:

ru****@gmail.com wrote:

Hello,

I need some help with the following:

1) I need to split a 16bit INT into two 8bit characters, and then be
able to join these two characters to form again the original 16bit
integer.

For example with values expressed in binary:

int n= 000000001111111;
--> char a=00000000;
--> char b=11111111;

and then be able to join a and b to form back n.

You can use a mask for that.

a = n & 0xff00;

You are missing a shift. On most systems, the result of your
expression will not fit in a char.

Oops. Indeed. That should have been a = (n & 0xff00) >> 8;

b = n & 0x00ff;

And use 'unsigned char' instead of 'char'. 'char' can be either
equivalent to 'unsigned char' or 'signed char'. If you rely on signed
char and use some value that cannot be represented with signed char,
then you may invoke implementation defined behavior or raise an
implementation defined signal.

2) And last, I need to convert an integer into a string.
For example,

int n=3.142 ---> char string[5] = ['3','.','1','4','2']

'int' is a part of the standard integer types and can't be used to
represent rationals. So you should either use an object type of float
or double. To produce a string from a given value of type float or
double, use sprintf/snprintf

Thank you very much!! I've tried working with casting but as far as I
know they don't seem to be the answer. I've also searched numerous
tutorials and web pages, but nothing productive. Any help or
suggestion will be highly appreciated. Thank you very much!!

casting is not a magical thing. The reason casting exists in the
language is for a few corner cases, namely: %p expecting a (void *)
with *printf family of functions. To compare the return value of mktime
to (time_t)-1 in case of error. For the is*, to* functions/macros.

There may be one or two other places that escape me at the moment but
not much else beyond that. The reason it doesn't produce a string is
because a cast yields a value of a given type, given a value. As there
does not exist a string type in C, then you can't expect that there
will be a string produced by a cast.

<<Remove the del for email>>

[ruroma]

Hi to all,

Firstly thank you very much all of you for such a quick answer.

Using the bitwise operators was the important hint, and yes, after
using the masks and shifting 8 positions I've finally got the high byte
and low byte separated as two 8 bit characters as I needed.

As for the second part, sorry, of course 3.142 isn't an int, just quick
writing and bad example. I'havent tried that part yet, but after more
searching I think that fcvtf will probably do the job.

Posting the whole code wont probably help, as these are just two little
parts I need for doing the big task. It is not actually homework, but
something I need to do after some long time without coding (and wasn't
too good either when I used to code more often).

Thank you once again :) I'll let you know If I encounter problems with
the second part, though I hope I don't ;)

Ian wrote:

ru****@gmail.com wrote:

Hello,

I need some help with the following:

1) I need to split a 16bit INT into two 8bit characters, and then be
able to join these two characters to form again the original 16bit
integer.

For example with values expressed in binary:

int n= 000000001111111;
--> char a=00000000;
--> char b=11111111;

and then be able to join a and b to form back n.

Sounds like homework! Mug up on shifting and masking.

2) And last, I need to convert an integer into a string.
For example,

int n=3.142 ---> char string[5] = ['3','.','1','4','2']

3.142.isn't an int, its a floating point value. Look up formatted output.

Ian

[Barry Schwarz]

On 5 Sep 2005 14:38:26 -0700, ru****@gmail.com wrote:

Hi to all,

Firstly thank you very much all of you for such a quick answer.

Using the bitwise operators was the important hint, and yes, after
using the masks and shifting 8 positions I've finally got the high byte
and low byte separated as two 8 bit characters as I needed.

As for the second part, sorry, of course 3.142 isn't an int, just quick
writing and bad example. I'havent tried that part yet, but after more
searching I think that fcvtf will probably do the job.

Is fcvtf a standard function?
<<Remove the del for email>>

[Lawrence Kirby]

On Mon, 05 Sep 2005 14:30:25 -0700, aegis wrote:

Barry Schwarz wrote:

On 5 Sep 2005 13:49:26 -0700, "aegis" <ae***@mad.scientist.com> wrote:
....

>You can use a mask for that.
>
>a = n & 0xff00;

You are missing a shift. On most systems, the result of your
expression will not fit in a char.

Oops. Indeed. That should have been a = (n & 0xff00) >> 8;

That can have problems on systems with 16 bit ints because you could be
shifting a negative value.

Lawrence

[Tim Rentsch]

Lawrence Kirby <lk****@netactive.co.uk> writes:

On Mon, 05 Sep 2005 14:30:25 -0700, aegis wrote:

Barry Schwarz wrote:

On 5 Sep 2005 13:49:26 -0700, "aegis" <ae***@mad.scientist.com> wrote:
...
You can use a mask for that.
>
>a = n & 0xff00;

You are missing a shift. On most systems, the result of your
expression will not fit in a char.

Oops. Indeed. That should have been a = (n & 0xff00) >> 8;

That can have problems on systems with 16 bit ints because you could be
shifting a negative value.

If ints were 16 bits, I'd be inclined to think 0xff00 would
be an unsigned int, which means n would be converted to unsigned
int before the '&' were done; the shift would then be working
on an unsigned int.

[Lawrence Kirby]

On Tue, 06 Sep 2005 03:14:13 -0700, Tim Rentsch wrote:

....

If ints were 16 bits, I'd be inclined to think 0xff00 would
be an unsigned int, which means n would be converted to unsigned
int before the '&' were done; the shift would then be working
on an unsigned int.

You're right, the problem in this case isn't in shifting a negative value,
is is converting a negative value to an unsigned type which won't preserve
the bit pattern on 1's complement and sign-magnitude based implementations.

Lawrence