Re: [PATCH] Handle half-float in 'x' command

[Simon Marchi <simark@simark.ca>]

On 3/2/23 10:50, Tom Tromey wrote:
> Simon> Can you add a comment before the section you added, to indicate what it
> Simon> intends to test?
> 
> I mean to say, I did it.  Thanks for the review.

Oops, this causes:

    p/f *(short *)s^M
    $41 = 1.7345e-05^M
    (gdb) FAIL: gdb.base/long_long.exp: p/f *(short *)s

    x/2hf h^M
    0x555555558060 <h>:     1.7345e-05      -0.028046^M
    (gdb) FAIL: gdb.base/long_long.exp: x/2hf h

I think it's just the test outcomes that need to be updated.

It goes through that code:

https://gitlab.com/gnutools/binutils-gdb/-/blob/70728e1d396475e8e630bfdd3fb8e8c8211bdbbd/gdb/printcmd.c#L455-464


  /* Printing a non-float type as 'f' will interpret the data as if it were
     of a floating-point type of the same length, if that exists.  Otherwise,
     the data is printed as integer.  */
  char format = options->format;
  if (format == 'f' && type->code () != TYPE_CODE_FLT)
    {
      type = float_type_from_length (type);
      if (type->code () != TYPE_CODE_FLT)
	format = 0;
    }

Before, we didn't find a 2-byte float format, so it was printed as
integers, and now we do, so it's printed as a float.  It seems
consistent with the comment in the code.

Simon