refactor(ssr): reduce duplicated text rendering logic

[nolanlawson]

Details

I got bugged by seeing so much repeated code in the generated JS. So I extracted some of it into a shared runtime helper.

Before:

didBufferTextContent = true;
{
  const value = instance.foo;
  textContentBuffer += value == null ? '' : String(value);
}
didBufferTextContent = true;
{
  const value = instance.bar;
  textContentBuffer += value == null ? '' : String(value);
}
didBufferTextContent = true;
{
  const value = instance.baz;
  textContentBuffer += value == null ? '' : String(value);
}

After:

didBufferTextContent = true;
textContentBuffer += renderTextContent(instance.foo);
didBufferTextContent = true;
textContentBuffer += renderTextContent(instance.bar);
didBufferTextContent = true;
textContentBuffer += renderTextContent(instance.baz);

No we just need to have optimizeAdjacentYieldStmts also optimize for adjacent text content boilerplate, since Terser doesn't seem smart enough to collapse this.

Does this pull request introduce a breaking change?

• 😮‍💨 No, it does not introduce a breaking change.

Does this pull request introduce an observable change?

• 🤞 No, it does not introduce an observable change.

[wjhsf]

🤔 Could we make the buffer state part of the context object that we pass around? Might help address this repetition.

[wjhsf]

Can't be fully at compile time, because directives like lwc:if are unknown until runtime. But maybe it could mitigate it?