Thread: table of functions

Hi,
I'm having difficulty visualising how to solve this.

Currently I have a function that looks up a value inside a table like so:

value = index_table[i];

However **without changing** the above line I'd like to call a function instead with the param i. In effect swapping the index_table for a function.

I have two ideas:
1: Swap the table completely by using a #define to remap the table to a function.
2: Keep the table but for each element in the table (table is not so big) call the function with a corresponding param.

I can not seem to manage for either way currently.
Is one or both of the ideas suitable or I'm I missing something more obvious?
Can someone point me to an example or show me a simple example?

Please remember I can not change value = index_table[i];

It'd be appreciated!!

Thanks for the quick reply.

>Why can't you change the line?
This line of code is in another module which is legacy code and is used in many different programs. I'm trying to not touch this for my circumstance. Of course if it not possible to do this then I will have to change this but really that is my last resort. However I believe it can be done without changing the line (easy said when I can't get it to work, i know!)

>Are you trying to have it call a different function for each value of i?
>Does it call the same function, but just pass it 0 or 1?
The same function but with 0 or 1. However wouldn't it be the same method in either case?

Also re-reading my original post, I should also clarify that the function must also return a value. This can be seen in the one line of code but I didn't outline it explicitly.

Thanks

Your constraint not to change the line "value = index_table[i];" makes what you're trying to do unachievable in C.

You could get a partial solution in C++ by making index_table an object of some type that has an operator[] defined. The catch with that, naturally enough, is that it won't work with a vanilla C compiler.

@quzah
Of course everything must be re-compiled, this is not the problem. The problem is that the module with

Code:

value = index_table[i];

is used in other programs. If i change this module then all other programs which make use of this module also have to change. Alternatively they must branch which is also not desireable. I don't want to get into a discsussion about SW reuse but I do not have the freedom to change/edit anything I like.

@Deckard
Your response did make me think about what new functionality I wanted and if it could not be done inside a table.
Below is a simple example:

Orginal table

Code:

static const u8 index_table[] =
{
  0x01,
  0x02,
  0x03,
}

However now I'd like to index into another table from this table based on the value of [i]. Something like:

Code:

static const u8* index_table[] =
{
  new_index_table,
  new_index_table,
  new_index_table,
}

static const u8 new_index_table[] =
{
  0x05,
  0x06,
  0x07,
}

so using

Code:

value = index_table[i]; //when i=2, value will equal 0x06

Hope this clarifies what I would like to do.

Thanks for the help.

Deckard,
I'm a little confused by your reply. I'm afraid I'm not doing a good job of explainaing the situation.

The code

Code:

value = index_table[i];

is inside a function which i can 'not' touch. However the actual index_table[] is in my module and I have complete control over this.
The variable value is used by the legacy code and this code expects a u8 value and that is all.
I initially wanted to change this as I need to process the data differently before it is returned to the legacy function. My idea was to perform this processing in a function hence why I was trying to use the table to point to a function passing differnt values into the function to process and also return the value.
However afer looking at the function which i wrote, it is not doing anything too clever. In fact it's only really looking up an entry in one table based on a value in another table (plus some masking). Therefore the solution is much simpler than my initial post was asking.
Thanks for the help anyway!