Shouldn't need access to private bug report to reverse a public duplicate marking

I'm not entirely sure. First question: how do you know that bug 126801 is the same issue as bug 122971, given it's private?

Well, I know that 122971 is a dupe of 124233. It is a sure thing.

But as 126801 is private, I have to trust the one that declared 124233 a dupe of 126801.

In the end, it shouldn't be possible to mark a bug as a duplicate of a private bug because we can't verify anything.

Hmmm. That bug is only private because it contains a private dump of data.

A quick solution is that as you have access to this bug, make 122971 a dupe of 126801.
In case I encounter such a problem later, I'll poke someone with higher privileges

If the only reason for bug 126801 being private is that it contains private data, and there is no reason for the other reports being private, one of the public reports should become the canonical version, because that results in more people who can learn about (and therefore fix) the bug. So someone with access should mark bug 126801 as a duplicate of either bug 122971 or bug 124233, whichever is most useful.

So I guess the only bug in Launchpad here is that you should be able to reverse "public X is duplicate of private Y" to "private Y is duplicate of public X" without having special access to Y. Stéphane, does that sound reasonable? If so I'll update this report to reflect that request.

This solution sounds ok for me.
Thanks

Is there any value in being able to link bug reports to related bug reports, in the way that we can presently link bugs<-->questions?

Isn't that relatively orthogonal to the question of being able to un-dupe a
bug?

So, unless you can see the private bug, you cannot actually judge whether:
 - it really is a duplicate
 - the privacy was due to a need for private *discussion* or because sensitive *data* was visible or it was *security* related.

Its possible that the public duplicate actually should be privatised, for instance.

Anyhow - I don't see letting folk without access to the private bug control that bug being sensible; it seems far to likely to have significant negative impact, so I'm going to close this wontfix.

However - we have some work coming up on disclosure and visibility and hopefully we'll improve the interactions around duplicating bugs and bug privacy as part of that work.