[clang-format] Lambda body left unformatted when a sibling field is a nested designated initializer

### Summary

When a brace-initializer list contains **both** (a) a field whose value is a
*nested designated initializer* and (b) a sibling field whose value is a *lambda
with a multi-statement body*, clang-format silently skips formatting the entire
lambda body. The body is left exactly as written and wrong indentation is not
corrected, `==` spacing is not applied, brace wrapping is not applied. The rest
of the file formats normally.

This is order-independent (the lambda may come before or after the nested
designated initializer) and reproduces on well-formed, valid C++ (it is not a
"fixing invalid code" edge case).

### Version

clang-format version 22.1.2.

### `.clang-format` (minimal)

```yaml
BasedOnStyle: WebKit
ColumnLimit: 120
```

`ColumnLimit` must be non-zero. WebKit's default `ColumnLimit: 0` masks the bug
(no line ever wraps). Any non-zero value (80, 120, 200) triggers it.

### Input (well-formed; only the lambda body is mis-indented)

```cpp
struct Layout {
int a;
int b;
};
struct Source {
Layout layout;
int count;
void (*fn)();
};
Source make()
{
return Source{
.layout = Layout{
.a = 1,
.b = 2,
},
.count = 3,
.fn = []() {
int x = 0;
x = x + 1;
},
};
}
```

### Command

```
clang-format --style=file repro.cpp
```

### Actual output (lambda body unchanged: bug)

```cpp
.fn = []() {
int x = 0;
x = x + 1;
},
```

Running clang-format a second time does not fix it: the wrong output is stable
(idempotent), so it is not a convergence/pass-ordering artifact.

### Expected output

```cpp
.fn = []() {
int x = 0;
x = x + 1;
},
```

### Minimal trigger (both elements required)

| Case | Nested designated-init sibling | Lambda sibling | Result |
| ---- | ---------------------------------------- | -------------- | ------- |
| A | `.layout = Layout{.a=1, .b=2}` | yes | **bug** |
| B | none (`.layout = 9`) | yes | ok |
| C | `.layout = Layout{.a=1, .b=2}` | no | ok |
| F | nested **non**-designated `Layout{1, 2}` | yes | ok |
| G | lambda first, nested designated second | yes | **bug** |

Removing either element, or making the nested initializer non-designated, makes
the lambda body format correctly (rows B/C/F are direct controls).

### Tracker search

I searched the issue tracker and found no matching report. The nearest existing
issue is #119688 ("Misaligned formatting of array of objects with designated
initializers"), but that is about column *alignment* of designated initializers,
not about an adjacent lambda body being skipped entirely.