Only print the pruning note if the trace was definitely pruned

[saethlin]

Per #1987 (comment)

The only problem with this is that it looks like we can't emit the note about -Zmiri-backtrace=full if we also emit a tag-tracking diagnostic. The final note about the setting simply doesn't appear on the final error if it is printed before. I feel like we're running up against some diagnostic deduplication logic?

[RalfJung]

I feel like we're running up against some diagnostic deduplication logic?

Yeah, sounds plausible.
An alternative might be to not emit this as a separate diagnostic, but attach it to the backtrace diagnostic instead? Doesn't seem very important though.

@bors r+

[bors]

📌 Commit f1c6498 has been approved by RalfJung

[bors]

⌛ Testing commit f1c6498 with merge b490e6b...

[saethlin]

attach it to the backtrace diagnostic instead?

I don't like how this tucks it away. And the notes we display like this are mostly relevant to particular frames, not the backtrace overall, and they are formatted as such.

[bors]

☀️ Test successful - checks-actions
Approved by: RalfJung
Pushing b490e6b to master...