fix: reject invalid n_dims in tensor header to prevent stack-buffer-overflow

[shafiuzzaman-md]

Problem

encodec_load_model_weights() (encodec.cpp:447) reads a tensor header's n_dims field as a raw int32_t directly from the model file and uses it, unbounded, as the loop count writing into a fixed-size 3-element stack array int32_t ne[3] = {1, 1, 1}:

int32_t nelements = 1;
int32_t ne[3] = {1, 1, 1};
for (int i = 0; i < n_dims; i++) {
    read_safe(infile, ne[i]);   // n_dims is fully attacker-controlled, no upper bound
    nelements *= ne[i];
}

A model file with n_dims > 3 writes past the ne[3] array on the stack. Reproduced with a crafted 132-byte model file (n_dims = 20) through the real public encodec_load_model() API, under an AddressSanitizer build:

==...==ERROR: AddressSanitizer: stack-buffer-overflow
WRITE of size 4 at ... thread T0
    #0 memcpy
    #1 read_safe<int>                       utils.h:12
    #2 encodec_load_model_weights           encodec.cpp:447
    #3 encodec_load_model                   encodec.cpp:949
    #4 main

Fix

Add an explicit n_dims range check immediately after it is read, before the loop that populates ne[]:

if (n_dims < 1 || n_dims > 3) {
    fprintf(stderr, "%s: invalid n_dims %d in model file (expected 1 <= n_dims <= 3)\n", __func__, n_dims);
    return false;
}

For normal, well-formed model files this is a no-op; a malformed file is now rejected with a clear error message instead of corrupting the stack.

Testing

• Before: the crafted model file (n_dims=20) crashes with an ASan stack-buffer-overflow WRITE at encodec_load_model_weights (encodec.cpp:447).
• After: the same file is rejected cleanly ("invalid n_dims 20 in model file"), no crash.
• Regression check: a file with a valid n_dims (3) passes the new check unaffected (fails later
  only because the synthetic test file has no real tensor data, unrelated to this fix).
• Both the library and a standalone harness calling only the public encodec_load_model() API
  build and run cleanly with the change.