Fix global fetch this issue

[NorOldBurden]

The Bug
In FetchAdapter, the generated code does this:
ts
this.fetchFn = fetchFn ?? createFetch(); // step 1: store fetch as property
// ...
const response = await this.fetchFn(url, opts); // step 2: call it
When createFetch() runs in the browser, it returns globalThis.fetch — which is window.fetch. The problem is step 2: calling this.fetchFn(...) invokes fetch with this === the FetchAdapter instance, but native window.fetch requires this === window. That's what "Illegal invocation" means — the browser rejects the call because this is wrong.In plain terms: storing a native function as an object property and then calling it detaches it from its original owner. Same reason "hello".trim works but const fn = "hello".trim; fn() throws.

[yyyyaaa]

1. sdk/constructive-sdk/src/public/orm/client.ts:61 and the other 12 checked-in generated SDK clients still contain this.fetchFn = fetchFn ?? createFetch();.
   .ie need to regenerate all sdks in this monorepo

2. The added regression tests are weak. The substring assertion at PR test lines 80-87 only proves .bind(globalThis) appears somewhere in generated text, not that FetchAdapter binds the selected fetch function correctly. It would pass with .bind(globalThis) in a comment or
   with the precedence bug fetchFn ?? createFetch().bind(globalThis). The second test at lines 89-105 does not instantiate FetchAdapter; it tests JavaScript bind() behavior in isolation.

3. The second test also does not match native browser timing exactly. In Chrome, detached window.fetch returns a Promise and then rejects with TypeError: Failed to execute 'fetch' on 'Window': Illegal invocation; it does not synchronously throw. The production await
   this.fetchFn(...) still observes it as a failure, but the test is a simplified demo, not a browser-faithful regression.