cli rebase: do not allow -r --skip-empty to empty descendants

This follows up on @matts1 's jj-vcs#2609.

We still allow the `-r` commit to become empty. I would be more comfortable if
there was a test for that, but I haven't done that (yet?) and it seems pretty
safe. If that's a problem, I'm happy to forbid `-r --skip-empty` entirely,
since it is far less useful than `-s --skip-empty` or `-b --skip-empty`.

I think it is undesired to abandon emptied descendants, since as far as
descendants of `A` are concerned, `jj rebase -r A` should be equivalent to `jj
abandon A`, and `jj abandon` does not remove emptied commits. It also doesn't
seem very useful to do that to descendant commits, since I think they'd only
become empty in pathological cases.

Additionally, if we did want -r to empty descendants of `A`, we'd have to add
thorough tests and possibly improve the algorithm. I want to refactor `rebase -r`
and add features to it, and considering these cases makes everything harder.

For example, if we have

```
root -> A -> B -> C
```

and `jj rebase -r A -d C` empties commit `B` (or `C`), I do not know whether
the current algorithm will work correctly. It seems possible that it would, but
that depends on the fact that empty merge commits are not abandoned for
descendants. That seems dangerous to rely on without tests.

Making DescendantRebaser return more information (which I am likely to do as part
of implementing `-r --after`) should help make such code more obviously correct.

cli/src/commands/rebase.rs

@@ -22,7 +22,7 @@ use jj_lib::backend::{CommitId, ObjectId};
 use jj_lib::commit::Commit;
 use jj_lib::repo::{ReadonlyRepo, Repo};
 use jj_lib::revset::{RevsetExpression, RevsetIteratorExt};
-use jj_lib::rewrite::{rebase_commit_with_options, EmptyBehaviour, RebaseOptions};
+use jj_lib::rewrite::{rebase_commit, rebase_commit_with_options, EmptyBehaviour, RebaseOptions};
 use jj_lib::settings::UserSettings;
 use tracing::instrument;
 
@@ -185,7 +185,7 @@ Please use `jj rebase -d 'all:x|y'` instead of `jj rebase --allow-large-revsets
             &mut workspace_command,
             &new_parents,
             rev_str,
-            &rebase_options,
+            &RebaseOptions::default(),
         )?;
     } else if !args.source.is_empty() {
         let source_commits =
@@ -357,22 +357,15 @@ fn rebase_revision(
 
         rebased_commit_ids.insert(
             child_commit.id().clone(),
-            rebase_commit_with_options(
-                settings,
-                tx.mut_repo(),
-                child_commit,
-                &new_child_parents,
-                rebase_options,
-            )?
-            .id()
-            .clone(),
+            rebase_commit(settings, tx.mut_repo(), child_commit, &new_child_parents)?
+                .id()
+                .clone(),
         );
     }
     // Now, rebase the descendants of the children.
-    rebased_commit_ids.extend(
-        tx.mut_repo()
-            .rebase_descendants_return_map(settings, rebase_options.clone())?,
-    );
+    // TODO(ilyagr): Consider making it possible for these decendants to become
+    // emptied, like --skip_empty. This would require writing careful tests.
+    rebased_commit_ids.extend(tx.mut_repo().rebase_descendants_return_map(settings)?);
     let num_rebased_descendants = rebased_commit_ids.len();
 
     // We now update `new_parents` to account for the rebase of all of
@@ -405,11 +398,7 @@ fn rebase_revision(
         &new_parents,
         rebase_options,
     )?;
-    debug_assert_eq!(
-        tx.mut_repo()
-            .rebase_descendants_with_options(settings, rebase_options.clone())?,
-        0
-    );
+    debug_assert_eq!(tx.mut_repo().rebase_descendants(settings)?, 0);
 
     if num_rebased_descendants > 0 {
         writeln!(

lib/src/repo.rs

@@ -881,10 +881,12 @@ impl MutableRepo {
     pub fn rebase_descendants_return_map(
         &mut self,
         settings: &UserSettings,
-        options: RebaseOptions,
     ) -> Result<HashMap<CommitId, CommitId>, TreeMergeError> {
         Ok(self
-            .rebase_descendants_return_rebaser(settings, options)?
+            // We do not set RebaseOptions here, since this function does not currently return
+            // enough information to describe the results of a rebase if some commits got
+            // abandoned
+            .rebase_descendants_return_rebaser(settings, Default::default())?
             .map_or(HashMap::new(), |rebaser| rebaser.rebased().clone()))
     }