Current (bad) behaviour: If I do darcs add or remove in a repository with a
format darcs does not recognise (in this case darcs-2-experimental), it
complains that it is not in a repository.

Desired behaviour: it should complain that it does not recognise the format
darcs-2-experimental

Note that other commands do this just fine, e.g. darcs changes

[Note: this is what I would normally mark as a bug, but I'm trying to follow the
new convention proposed by Thorkil,
http://wiki.darcs.net/index.html/BugTrackerIssueManagement so I'm putting this
down as a wishlist item... I hope I got it right]

It seems to me that the first step for this is to produce a
test script for bugs/.  Eric, would you like to do this?

Assigning to myself to write a test for this.  It's important to me that we take
care of this fairly early to better accommodate future changes to the format file.

Just to note that because we are planning on adding a new format for 
rebase, which itself has probably slipped to 2.6, we really should do this 
one for 2.5.

See also issue1647

Kowey, if it's important given the oncoming format work and it's
probably easy, maybe you can do this one quickly before the freeze?

Test submitted (need-implementation); code review under way (in-progress)

The following patch updated issue issue1277 with status=resolved;resolvedin=2.6.0 HEAD

* Resolve issue1277: percolate repository format errors correctly. 
Ignore-this: b541efa39c3b55b67479b209f55ffd1d
The problem is that we do not distinguish between bad repos and
non-repositories so we keep seeking upwards.

The following patch updated issue issue1277 with status=resolved;resolvedin=2.5.0 CURRENT

* Resolve issue1277: percolate repository format errors correctly. 
Ignore-this: b541efa39c3b55b67479b209f55ffd1d
The problem is that we do not distinguish between bad repos and
non-repositories so we keep seeking upwards.