fix(vite): replace the browser preload helper before esbuild minifies it - #11425

PR Checklist

• The PR title follows our guidelines: https://github.com/NativeScript/NativeScript/blob/master/tools/notes/CONTRIBUTING.md#commit-messages.
• There is an issue for the bug/feature this PR is for. To avoid wasting your time, it's best to open a suggestion issue first and wait for approval before working on it.
• You have signed the CLA.
• All existing tests are passing: https://github.com/NativeScript/NativeScript/blob/master/tools/notes/DevelopmentWorkflow.md#running-unit-tests-application.
• Tests for the changes are included - https://github.com/NativeScript/NativeScript/blob/master/tools/notes/WritingUnitTests.md.

What is the current behavior?

dynamicImportPlugin replaces Vite's browser __vitePreload helper (which touches document and window) with a native one, but it does so from generateBundle using a literal regex on the identifier. With build.minify: 'esbuild' Vite's vite:esbuild-transpile plugin has already renamed __vitePreload in renderChunk, so the regex misses, the browser helper ships, and its window.dispatchEvent in the rejection path turns every failed dynamic import into ReferenceError: window is not defined with a misattributed stack. Details in #11423.

What is the new behavior?

The transform runs in renderChunk as an enforce: 'post' plugin. User post plugins run before buildPlugins.post (where the esbuild minifier lives), so the helper is still named when it is replaced. renderChunk returns null when nothing changed, so untouched chunks and already-transformed code are left alone. transformDynamicImports is unchanged.

Specs cover the plugin shape (post + renderChunk, no generateBundle), the replacement through renderChunk, and the no-op/idempotent return.

Verified with a real ns build ios of a blank Vue app with build.minify: 'esbuild': the emitted bundle contains the native helper and no vite:preloadError / document.createElement("link"); the published 8.0.5 bundle has the reverse. At runtime a deliberately failing import() now reports the real load error.

Fixes #11423.