Combine smaller bundles into larger ones (closes #15)

docs/man/zvault-vacuum.1.md

@@ -43,6 +43,11 @@ should be avoided when the storage space permits it.
 
 ## OPTIONS
 
+  * `--combine`:
+
+    Also combine small bundles into larger ones.
+
+
   * `-r`, `--ratio <NUM>`:
 
     Do not rewrite bundles with more than `NUM`% of used chunks.

src/cli/args.rs

@@ -52,7 +52,8 @@ pub enum Arguments {
     Vacuum {
         repo_path: String,
         ratio: f32,
-        force: bool
+        force: bool,
+        combine: bool
     },
     Check {
         repo_path: String,
@@ -337,6 +338,7 @@ pub fn parse() -> Result<(LogLevel, Arguments), ErrorCode> {
         .subcommand(SubCommand::with_name("vacuum").about("Reclaim space by rewriting bundles")
             .arg(Arg::from_usage("-r --ratio [NUM] 'Ratio in % of unused space in a bundle to rewrite that bundle'")
                 .default_value(DEFAULT_VACUUM_RATIO_STR).validator(validate_num))
+            .arg(Arg::from_usage("--combine 'Combine small bundles into larger ones'"))
             .arg(Arg::from_usage("-f --force 'Actually run the vacuum instead of simulating it'"))
             .arg(Arg::from_usage("<REPO> 'Path of the repository'")
                 .validator(|val| validate_repo_path(val, true, Some(false), Some(false)))))
@@ -490,6 +492,7 @@ pub fn parse() -> Result<(LogLevel, Arguments), ErrorCode> {
             Arguments::Vacuum {
                 repo_path: repository.to_string(),
                 force: args.is_present("force"),
+                combine: args.is_present("combine"),
                 ratio: parse_num(args.value_of("ratio").unwrap()).unwrap() as f32 / 100.0
             }
         },

src/cli/mod.rs

@@ -435,10 +435,10 @@ pub fn run() -> Result<(), ErrorCode> {
                 info!("Run with --force to actually execute this command");
             }
         },
-        Arguments::Vacuum{repo_path, ratio, force} => {
+        Arguments::Vacuum{repo_path, ratio, force, combine} => {
             let mut repo = try!(open_repository(&repo_path));
             let info_before = repo.info();
-            checked!(repo.vacuum(ratio, force), "vacuum", ErrorCode::VacuumRun);
+            checked!(repo.vacuum(ratio, combine, force), "vacuum", ErrorCode::VacuumRun);
             if !force {
                 info!("Run with --force to actually execute this command");
             } else {

src/repository/integrity.rs

@@ -284,7 +284,7 @@ impl Repository {
             },
             Err(err) => return Err(err)
         };
-        for (name, mut backup) in ProgressIter::new("ckecking backups", backup_map.len(), backup_map.into_iter()) {
+        for (name, mut backup) in ProgressIter::new("checking backups", backup_map.len(), backup_map.into_iter()) {
             let path = format!("{}::", name);
             match self.check_subtree(Path::new(&path).to_path_buf(), &backup.root, &mut checked, repair) {
                 Ok(None) => (),

src/repository/vacuum.rs

@@ -13,7 +13,7 @@ impl Repository {
         }
     }
 
-    pub fn vacuum(&mut self, ratio: f32, force: bool) -> Result<(), RepositoryError> {
+    pub fn vacuum(&mut self, ratio: f32, combine: bool, force: bool) -> Result<(), RepositoryError> {
         try!(self.flush());
         info!("Locking repository");
         try!(self.write_mode());
@@ -36,6 +36,28 @@ impl Repository {
                 reclaim_space += bundle.get_unused_size();
             }
         }
+        if combine {
+            let mut small_meta = vec![];
+            let mut small_data = vec![];
+            for (id, bundle) in &usage {
+                if bundle.info.encoded_size * 4 < self.config.bundle_size {
+                    match bundle.info.mode {
+                        BundleMode::Meta => small_meta.push(*id),
+                        BundleMode::Data => small_data.push(*id),
+                    }
+                }
+            }
+            if small_meta.len() >= 2 {
+                for bundle in small_meta {
+                    rewrite_bundles.insert(bundle);
+                }
+            }
+            if small_data.len() >= 2 {
+                for bundle in small_data {
+                    rewrite_bundles.insert(bundle);
+                }
+            }
+        }
         info!("Reclaiming {} by rewriting {} bundles", to_file_size(reclaim_space as u64), rewrite_bundles.len());
         if !force {
             self.dirty = false;