Re: hard to understand code for the representation of tab

[Gavin Smith]

On Sat, Oct 12, 2024 at 09:15:10PM +0200, Patrice Dumas wrote:
> Hello,
> 
> This is not important, but in the info code, in info/util.c in the
> printed_representation function, which returns a pointer to string that
> is the printed representation of character (or other logical unit) if it
> were printed at a given screen column.
> 
> l 226 for tab there is a code that I do not understand at all,
>   *pchars = ((pl_chars + 8) & 0xf8) - pl_chars;
> 
> A bit more context for the variables:
> pl_chars: size_t the string column the printed representation of character
>    should be printed at
> rep: text buffer corresponding to the returned value
> pchars: int gets the number of screen columns taken up by the return value
> pbytes: int gets the number of bytes in returned string

Here's an attempt at explaining it:

pl_chars is the current column.  We want to find the next multiple of 8
after the current column, assuming a tab is 8 spaces.  We add 8 to the
current column, and then subtract the remainder when this result is
divided by 8.

For example 

A.  21 + 8 = 28, 
B.  Then 28 mod 8 = 4
C.  28 - 4 = 24.
D.  Finally, we take away the starting point: 24 - 21 = 3.

So we output 3 spaces to get up to 24 columns.

(n & 0xf8) is supposed to do steps B and C combined
as the binary representation of 0xf8 is 1111 1000 so the result has
to be a multiple of 8.

However, this doesn't work if there are more than 255 columns (which is
unlikely, though).

> here is more context for the code:
> 
>       else if (*cur_ptr == '\t')
>         {
>           int i = 0;
> 
>           *pchars = ((pl_chars + 8) & 0xf8) - pl_chars;
>           *pbytes = *pchars;
> 
>           /* We must output spaces instead of the tab because a tab may
>              not clear characters already on the screen. */
>           for (i = 0; i < *pbytes; i++)
>             text_buffer_add_char (rep, ' ');
>           return text_buffer_base (rep);
>         }
> 
> -- 
> Pat
>