Re: [PATCH 4/7] GCOV: add -j argument (human readable format).

[Nathan Sidwell <nathan@acm.org>]

On 10/26/2017 04:11 AM, marxin wrote:
> Human readable format is quite useful in my opinion. There's example:
> 
>          -:    1:unsigned
>     14.00K:    2:loop (unsigned n, int value)

My first thought is 'why 2 decimal places'?  That seems excessive.  Zero 
surely suffices?

> Question is do we want to do it by default, or a new option is fine?
> Note that all external tools using gcov should use intermediate format
> which is obviously unchanged.

I don't think other tools do it by default.


> +     [@option{-j}|@option{--human-numbers}]

man ls:
       -h, --human-readable
               with -l and/or -s, print human readable sizes (e.g., 1K 
234M 2G)

Sadly '-h' is help (IIRC).  but we could at least copy the long form.


> +  const char *units[] = {"", "K", "M", "G", "Y", "P", "E", "Z"};

Those aren't right  KMGTPEZY
http://www.npl.co.uk/reference/measurement-units/si-prefixes/


> +  for (unsigned i = 0; i < sizeof (units); i++)
> +    {
> +      if (v < 1000.0f)
> +	{
> +	  sprintf (buffer, "%3.2f%s", v, units[i]);
> +	  return buffer;
> +	}
> +
> +      v /= 1000.0f;
> +    }

that's going to fail on certain roundings.  You're doing multiple 
divisions by 1000, which itself will have roundings.  But more 
importantly, numbers after scaling like 999.999 will round to 1000, and 
you probably don't want that.  This kind of formatting is tricky.  My 
inclination is to keep it in the integer domain.  Determine a scaling 
factor.  Divide once by that and then explicitly round to nearest 
even[*] with a check for the 999.9 case.  Then print as an unsigned.

nathan

[*] I presume you're familiar with RNE?  That's probably overkill and 
plain RN would be adequate.

-- 
Nathan Sidwell