Gemini Lasswell <gazally@runbox.com> writes:

> This could be fixed by updating the list in edebug-read-function of
> characters that might possibly follow # to keep up with the changes to
> read1 since the last time that list was updated (in 2001). But I think
> it's better to remove the test entirely. If there is an invalid
> character then read1 will report a syntax error.

Ooops, attached the wrong patch. Here's the right one: