Anonymizing text using emacs (replace all non-whitespace characters by x)

Question

Suppose I select some text in an emacs buffer using point and mark. How do I tell emacs to replace each non-whitespace character with x, say? For concreteness, that I select the text using point and mark. Also, for clarity, I consider newlines to be whitespace for the purpose of this question.

The idea is to keep the structure of the text intact, but remove all information. My current use case is a LaTeX MWE document.

Answer 1

Use M-x replace-regexp to replace all non-whitespace characters by x. You can use C-M-% to run query-replace-regexp and, once a match is found, press ! to replace all matches.

Now for a regexp that matches non-whitespace characters. You'd think that the character class [:space:] would match whitespace characters, and so you'd use [^[:space:]] as the regexp to mean “any non-whitespace character”. But in LaTeX mode, newlines have the “comment end” syntax instead of the “whitespace” syntax, so they don't count as whitespace for the purpose of [:space:]. So you need to use the regexp

[^[:space:]
]

To enter a newline in the regexp, press C-q C-j: the command C-q means to enter the next character literally instead of interpreting it as a command, and C-j (not the C-m character that the Enter key sends) is the newline character.

That is, you type:

M-x replace-regexp RET
[^[:space:] C-q C-j ] RET
x RET

Note that this replaces all non-whitespace, including LaTeX commands, comments, etc.

Why not use lipsum instead?