I'm having trouble understanding why these were removed.

@DanielRosenwasser The SpaceAfterAwaitKeyword and SpaceAfterTypeKeyword rules are merged into SpaceAfterCertainKeywords and SpaceAfterCertainTypeScriptKeywords rules respectively.

What about the NoSpaceAfterAwaitKeyword one? Is that just encompassed by SpaceAfterCertainKeywords?

The other NoSpace ones should not exist. I once thought they will work together with Space rules but then found they should only exist when there are some valid formatting options.

I once thought NoSpace rules would first remove those whitespaces and after that Space rules would add one. However, a Space rule can remove excessive whitespaces by itself.

I once thought NoSpace rules would first remove those whitespaces and after that Space rules would add one. However, a Space rule can remove excessive whitespaces by itself.

That's what I thought as well - it does make sense that one doesn't necessarily need the other though.

Not sure why this is needed? Can we just do what is done for all other kinds of blocks, namely:

• add SyntaxKind.NamedImports, SyntaxKind.NamedExports to the list in shouldIndentChildNode function so its content will be indented
• update isCompletedNode function to correctly process case when cursor is in incompleted NamesImport / NamedExport node by using hasChildOfKind(n, SyntaxKind.CloseBraceToken, sourceFile);

hasBraceShell function is not to check completeness as ImportClause node can omit its optional braces. I think isCompletedNode function is to check a node has its required children.

@vladima I was wrong and the braces are not optional. I just removed hasBraceShell and added a simpler check.

@saschanaz currently nodeIsCompleted handles cases when code is incomplete i.e.

import {$ // hit enter at $, caret on the next line should be indented
```.

I would expect imports and export to follow the same technique that we already use for all other blocks\block like constructions

Can you move this comment down? Also, is it correct?

I later changed it to // NamedImports has its own braces as Block does (so do not give it additional indentation)

@vladima Former hasBraceShell (and current namedBindings check) is needed to prevent unwanted excessive indentation. Consider this code:

import
{
    x as xx
}
from "foo"

import
    * as Foo from "foo"

ImportDeclaration has ImportClause node and then ImportClause node can have two different kind of nodes: NamedImports and NamespaceImport.

NamedImports, as you noted, looks like a block, but NamespaceImport doesn't. If I decide to always indent ImportClause or always not to do it, formatting will look bad:

import
    {
        x as xx
    } // when ImportClause is always indented
from "foo"

import
* as Foo from "foo" // when always not

Thus I added this check.

Anyway I agree that isCompletedNode should be updated. I'll push a change for it.

we don't use nulls - use undefined instead