Fwd: A typo of yours or a misunderstanding of mine?

[Tom de Vries]

On 11/18/21 6:19 PM, Z J Hu via Gdb wrote:
> I have been asked to forward my question about gdb documentation to you.
> 
> ---------- Forwarded message ---------
> From: Z J Hu <zihu88@gmail.com>
> Date: Wed, Nov 17, 2021 at 9:11 AM
> Subject: A typo of yours or a misunderstanding of mine?
> To: <sourcemaster@sourceware.org>
> 
> 
> Dear Sourceware Team,
> 
> Hope this email finds you well.
> 
> When I'm reading the content in "
> https://sourceware.org/gdb/onlinedocs/gdb/Memory.html", I find a statement
> for 'x/-3uh' command as: "You can also specify a negative repeat count to
> examine memory backward from the given address. For example, ‘x/-3uh 0x54320’
> prints three halfwords (h) at 0x54314, 0x54328, and 0x5431c." Should
> 0x54314 be 0x54324 instead? I don't know what machine word size would be
> for this statement, and it gives me some hard time to understand. Could you
> clear it for me?
> 

Hi,

I did an experiment, to understand the behaviour:
...
$ cat -n test.c
     1  #define N 256
     2
     3  unsigned short data[N];
     4
     5  int
     6  main (void)
     7  {
     8    int i;
     9    for (i = 0; i < N; ++i)
    10      data[i] = i;
    11
    12    return 0;
    13  }
$ gcc test.c -g
$ gdb -q -batch a.out -ex "set trace-commands on" -ex "b 12" -ex run -ex
"p &data[10]" -ex "x/-3uh &data[10]"
+b 12
Breakpoint 1 at 0x4004c3: file test.c, line 12.
+run

Breakpoint 1, main () at test.c:12
12        return 0;
+p &data[10]
$1 = (unsigned short *) 0x601074 <data+20>
+x/-3uh &data[10]
0x60106e <data+14>:     7       8       9
$
...

Based on this, I'd say you caught a typo in the docs, and the example in
the docs should mention the addresses 0x54314, 0x54318, and 0x5431c.

Thanks for bringing this to our attention, I'll commit a fix shortly.

Thanks,
- Tom