H5LS-tintascii fails on big-endian ppc64: h5ls prints "8-bit integer" instead of "native signed char"

### Describe the bug

The `H5LS-tintascii` CTest fails on a **big-endian PowerPC (ppc64)** machine because `h5ls` reports the datatype of an 8-bit signed integer as a generic `8-bit integer` instead of the expected `native signed char`.

On PowerPC, the default `char` is *unsigned*, so `H5T_NATIVE_SCHAR` no longer matches the file's signed 8-bit integer type. `h5ls` falls back to printing the generic `8-bit integer` description, which does not match the portable reference output `tintascii.ls`.

### Expected behavior

`h5ls -w80 -vldrs tintascii.h5` prints:

```
Type: native signed char
```

### Actual behavior

On ppc64 (big endian), it prints:

```
Type: 8-bit integer
```

### CDash test result

https://my.cdash.org/tests/444670969

Relevant diff from the test output:

```
-- line = 7
***ACTUAL: Type: 8-bit integer
****REFER: Type: native signed char

-- line = 15
***ACTUAL: Type: 8-bit integer
****REFER: Type: native signed char
```

Command run:

```
h5ls -w80 -vldrs tintascii.h5
```

### Platform (please complete the following information)

- **Site:** echidna (Continuous)
- **OS:** Linux 3.10.0-1160.36.2.el7.ppc64 (RHEL 7.9), **ppc64, big endian**
- **Compiler:** gcc 4.8
- **MPI:** mpich 3.3
- **Build config:** `ppc64/rh-7.9/mpich-3.3/gcc-4.8/-sh/f/mp/nr/np=4`
- **CMake/CTest:** ctest 3.29.0-rc2

### Additional context

The root cause is that `char` signedness is platform-dependent (unsigned by default on PowerPC). The reference expected output assumes `signed char` maps to `H5T_NATIVE_SCHAR`. Either the reference output needs a big-endian / unsigned-char variant, or the test needs to account for platforms where the native `char` is unsigned.