Switch to a command-line API closer to TaskWarrior

* Use a parser (rather than clap) to process the command line
* Outline some generic support for filtering, reporting, modifying, etc.
* Break argument parsing strictly from invocation, to allow independent testing

cli/src/invocation/cmd/modify.rs

@@ -0,0 +1,49 @@
+use crate::argparse::{Filter, Modification};
+use crate::invocation::{apply_modification, filtered_tasks};
+use failure::Fallible;
+use taskchampion::Replica;
+
+pub(crate) fn execute(
+    replica: &mut Replica,
+    filter: Filter,
+    modification: Modification,
+) -> Fallible<()> {
+    for task in filtered_tasks(replica, &filter)? {
+        let mut task = task.into_mut(replica);
+
+        apply_modification(&mut task, &modification)?;
+    }
+
+    Ok(())
+}
+
+#[cfg(test)]
+mod test {
+    use super::*;
+    use crate::argparse::DescriptionMod;
+    use crate::invocation::cmd::test::test_replica;
+    use taskchampion::Status;
+
+    #[test]
+    fn test_modify() {
+        let mut replica = test_replica();
+
+        let task = replica
+            .new_task(Status::Pending, "old description".to_owned())
+            .unwrap();
+
+        let filter = Filter {
+            ..Default::default()
+        };
+        let modification = Modification {
+            description: DescriptionMod::Set("new description".to_owned()),
+            ..Default::default()
+        };
+        execute(&mut replica, filter, modification).unwrap();
+
+        // check that the task appeared..
+        let task = replica.get_task(task.get_uuid()).unwrap().unwrap();
+        assert_eq!(task.get_description(), "new description");
+        assert_eq!(task.get_status(), Status::Pending);
+    }
+}