Well, emacs lets you hit tab anywhere on a line, at which point it reindents the line to the appropriate depth, depending on the syntax of the current and preceding lines. I'm pretty sure that most of the modes I use don't look at lines after the current line when performing indentation, but I'm not sure there's any reason why they couldn't.
If you want to reformat an entire block of code in this way, you can set a region (the emacs equivalent of "selection") and then press Ctrl-Meta-\, which stands for "indent-region". Then either the whole region will magically become properly indented, or emacs will yell at you because your XHTML syntax is broken (at least, that's what just happened when I tried this in XHTML mode. Every major mode can be slightly different.)
I know that emacs supports all the indentation tricks shown in this guy's example ... except that I don't know about the end-of-line comments. I never use end-of-line comments.
That sounds like it is simply indenting the whole line... not reformatting the actual block... i.e. respacing internally within the line. Which is what the elastic tabstops is doing.
It's not simply realigning each line. It's reformatting the entire block, realigning ALL tabstops within the lines, so that blocks of text are all positioned within a grid.
Elastic tab stops is actually orthogonal to this. Aligning elastic tab stops is easy (minor complications caused by blank lines), knowing how many tab stops to put in is where those 27 000 lines of lisp come in.
In some modes (e.g. C mode, plain text mode, several others), yes, though it's usually turned off by default. (I find it really distracting.) Usually, it automatically indents when you type a newline, add an opening or closing brace, an if statement, etc. The re-indent-as-you-go behavior seems to be used most in C mode, because you can specifically configure your indent style (http://en.wikipedia.org/wiki/Indent_style).
Most modes also have a command such as c-fill-paragraph that re-indents the chunk of code (e.g. function body) that the cursor is in. (Emacs refers to this behavior as "filling", so that's what to search for.) I think people use this pretty often - it's alt-Q by default.
The automatic indentation is quite good for some modes (C, Lisps in general, Python, OCaml, Lua, probably many others), but not always perfect (haskell-mode comes to mind). I think the important thing is that Emacs provides a lot of hooks for building mode-specific indentation-control functions and is not terribly difficult to extend.
Honest question: Is that what emacs does?