Comment 6 for bug 777405

> Why would you run M-x slime-macroexpand-again manualy, it's bound to g only in the slime-macroexpansion mode.

What part of "It can easily happen by accident with minibuffer-complete" do you not understand?

Just because it doesn't happen to Stas or fit your usage pattern doesn't mean it isn't a bug.

`slime-macroexpand-again' relies on functions which set buffer-undo-list nil _and_ erase-buffer. When this happens in a non "*slime-macroexpansion*" buffer the end result is I loose the contents of whatever changes have been made since the last autosave/save/commit.

The proposed fix for this problem provided above is easy, clean, and does not break the semantics of existing slime features.

However, in so much as you will not consider this a bug I suggest we move the discussion there -- maybe others can provide additional insight. As such, I've forwarded a patch with more extensive changes related to the problem to slime-devel.

--
/s_P\