metadata: filter the `removed` set to only contains relevant data
The `files` entry can be bogus and contains too many entries. This can badly
combines with the computation of `removed` inflating the set size. The can lead
to the changesets centric rename computation to process much more data than
needed, slowing it down (and increasing space taken by data storage).
In practice newer commits already that reduced set, this applies this "fix" to
older changeset.
Differential Revision: https://phab.mercurial-scm.org/D8589
$ hg init base
$ cd base
$ echo 'alpha' > alpha
$ hg ci -A -m 'add alpha'
adding alpha
$ cd ..
$ hg clone base work
updating to branch default
1 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ cd work
$ echo 'beta' > beta
$ hg ci -A -m 'add beta'
adding beta
$ cd ..
$ cd base
$ echo 'gamma' > gamma
$ hg ci -A -m 'add gamma'
adding gamma
$ cd ..
$ cd work
$ hg pull -q
$ hg merge
1 files updated, 0 files merged, 0 files removed, 0 files unresolved
(branch merge, don't forget to commit)
Update --clean to revision 1 to simulate a failed merge:
$ rm alpha beta gamma
$ hg update --clean 1
2 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ cd ..