BUG: Check dim[0] before using it to index dim[] in the NIFTI-2 converter

[gdevenyi]

nifti_convert_n2hdr2nim() uses nhdr.dim[0], the number of dimensions,
as a loop bound over the eight-element dim[] array:

for( ii=2 ; ii <= nhdr.dim[0] ; ii++ )        ...
for( ii=nhdr.dim[0]+1 ; ii <= 7 ; ii++ )      ...
for( ii=1 ; ii <= nhdr.dim[0] ; ii++ )        ...

It never checks that dim[0] is in [0,7]. The NIFTI-1 converter gets
that check for free, because need_nhdr_swap() rejects a dim[0] outside
[1,7] in either byte order, but the NIFTI-2 path decides swapping from
sizeof_hdr and reaches the loops with whatever the file said.

A NIFTI-2 header with a large negative dim[0] therefore starts the
second loop at a wild negative index, reads far outside the header and
segfaults. This is not confined to the header API: nifti_image_read()
gets there for any file with a valid 540 byte NIFTI-2 header, so a
604 byte file crashes nifti_tool:

$ nifti_tool -disp_nim -infiles bad_n2_dim0.nii
nifti2_io.c:5079: runtime error: index -6727636073941130588 out of
    bounds for type 'int64_t[8]'
AddressSanitizer: SEGV ... in nifti_convert_n2hdr2nim

dim[0] is now range checked in the same place, and in the same style,
as the dim[1] check just below it. Zero stays acceptable, as it is on
the NIFTI-1 side. Valid headers are unaffected: dim[0] outside [0,7]
has no meaning in either format.

Found by fuzzing nifti_convert_n2hdr2nim() with clang's libFuzzer under
AddressSanitizer.

---

How to reproduce, and the sanitizer builds used, are in the commit message. Found with clang libFuzzer harnesses over nifti_image_read(), the two header converters, nifti_image_from_ascii() and axml_read_buf(), run under AddressSanitizer, UndefinedBehaviorSanitizer and MemorySanitizer.

One commit on master, independent of my other open PRs. ctest passes on master with this branch alone (343/343, excluding the nifti_c22_copy_image failure that #31 fixes), and on the union of all my branches under ASan+UBSan, TSan and MSan.

[hjmjohnson]

The commit messages in this range were rewritten to remove trailers that do not belong in permanent history: Co-Authored-By: naming an AI tool, and Claude-Session: URLs that resolve for nobody. Only messages changed — the tree at the tip of master is byte-identical, and author, committer, and dates are preserved.

This PR's commit on the rewritten master:

• 5e32002bd7 BUG: Check dim[0] before using it to index dim[] in the NIFTI-2 converter

The SHA recorded above by GitHub is from the pre-rewrite history and no longer resolves.