seen couple of use cases where the model usage pattern is not as
            expected and would not work with this change.
    *   output tensors of one layer is used in the constructor of another.
    *   symbolic tensors outside the scope of the model are used in custom loss
        functions. The flag can be disabled for these cases and ideally the
        usage pattern will need to be fixed.

and returns a new [cipher.Block] implementation. // The key argument should be the AES key, either 16, 24, or 32 bytes to select // AES-128, AES-192, or AES-256. func New(key []byte) (*Block, error) { // This call is outline to let the allocation happen on the parent stack. return newOutlined(&Block{}, key) } // newOutlined is marked go:noinline to avoid it inlining into New, and making New // too complex to inline itself. // //go:noinline func newOutlined(b *Block, key []byte) (*Block, error) {...

and returns a new [cipher.Block] implementation. // The key argument should be the AES key, either 16, 24, or 32 bytes to select // AES-128, AES-192, or AES-256. func New(key []byte) (*Block, error) { // This call is outline to let the allocation happen on the parent stack. return newOutlined(&Block{}, key) } // newOutlined is marked go:noinline to avoid it inlining into New, and making New // too complex to inline itself. // //go:noinline func newOutlined(b *Block, key []byte) (*Block, error) {...

### Docs

* ✏️ Fix validation parameter name in docs, from `regex` to `pattern`. PR [#10085](https://github.com/tiangolo/fastapi/pull/10085) by [@pablodorrio](https://github.com/pablodorrio).