bookmarks and branches again: files

[Uwe Brauer]

    > On Tue, Jul 12, 2016 at 10:43 AM, Uwe Brauer <oub at mat.ucm.es> wrote:

    > Obviously, because you updated to the bookmark that doesn't contain it. Why
    > did you expect it to be there?

Misunderstanding, I expected it *not* to be there. In my first example
however, I only used one bookmark instead of two, and then when I
updated to «default» the file was there.

So I repeated my example with two bookmakrs and I sent it, because I
thought Simon was questioning such a behavior (but most likely I
misunderstood him, since he knows more than me about mercurial).

    > Why did you expect this? You updated to a changeset which does not contain
    > the file, so why should it be present in your working copy?


    > It may not be in your working copy, but it's still in a head in your
    > repository graph. You can still update to it by passing an explicit
    > revision. The file is still there in your repository, and unlike git
    > nothing will ever be deleted from your history until you explicitly tell
    > Mercurial to do so using an extension which you must enable beforehand.

The problem I have is this: suppose I accidentally deleted one (or more)
bookmarks.

How can I know about these file which correspond to the bookmark
deleted?

(That issue seems to be absent if I used branches, since I cannot delete branches)