Another "hg diff" annoyance

[TK Soh]

On Fri, May 22, 2009 at 5:02 AM, Greg Lindahl <greg at blekko.com> wrote:
>> > As I said, he didn't untrack the file. That's a user error. But the
>> > consequences of this error are too hard to find... he reasonably
>> > expects that if there are uncommitted changes, "hg diff" will tell him
>> > about it.
>>
>> IMHO, hg status is _the_ command for this. He simply use the 'wrong' command.
>
> I thought it was git that's famous for its obscure and unforgiving
> interface? ;-)
>
> From the number of people who have commented about "hg diff" not
> showing them as much as "hg diff --git" when there are changes, it's
> clear that a lot of people want to use "hg diff" in this sort of
> situation.
>
> Chatting with my other hg users, it seems that this particular error
> (removing a file without hg rm, then getting stuck with "uncommitted

Do I understand correctly that you want hg to treat the missing file as removed?

> changes") has happened several times to other people, and they've
> ended up recloning their repos to fix it.

I'm missing the picture. Why is recloning needed here?