Add --remove-empty-dirs option to mkdwarfs

include/dwarfs/entry.h

@@ -134,6 +134,8 @@ class dir : public entry {
                   global_entry_data const& data) const;
   uint32_t inode_num() const override { return inode_; }
   void scan(os_access& os, progress& prog) override;
+  bool empty() const { return entries_.empty(); }
+  void remove_empty_dirs(progress& prog);
 
  private:
   using entry_ptr = std::shared_ptr<entry>;

include/dwarfs/options.h

@@ -49,10 +49,11 @@ struct filesystem_options {
 enum class file_order_mode { NONE, PATH, SCRIPT, SIMILARITY };
 
 struct scanner_options {
-  file_order_mode file_order;
+  file_order_mode file_order{file_order_mode::NONE};
   std::optional<uint16_t> uid;
   std::optional<uint16_t> gid;
   std::optional<uint64_t> timestamp;
+  bool remove_empty_dirs{false};
 };
 
 std::ostream& operator<<(std::ostream& os, file_order_mode mode);

src/dwarfs.cpp

@@ -501,18 +501,20 @@ int main(int argc, char* argv[]) {
 
   fuse_opt_parse(&args, &s_opts, dwarfs_opts, option_hdl);
 
-  s_opts.cachesize = s_opts.cachesize_str ? parse_size_with_unit(s_opts.cachesize_str)
+  s_opts.cachesize = s_opts.cachesize_str
-                                      : (static_cast<size_t>(512) << 20);
+                         ? parse_size_with_unit(s_opts.cachesize_str)
+                         : (static_cast<size_t>(512) << 20);
   // TODO: foreground mode, stderr vs. syslog?
   s_opts.debuglevel = s_opts.debuglevel_str
-                        ? logger::parse_level(s_opts.debuglevel_str)
+                          ? logger::parse_level(s_opts.debuglevel_str)
-                        : logger::INFO;
+                          : logger::INFO;
-  s_opts.workers = s_opts.workers_str ? folly::to<size_t>(s_opts.workers_str) : 2;
+  s_opts.workers =
+      s_opts.workers_str ? folly::to<size_t>(s_opts.workers_str) : 2;
   s_opts.lock_mode =
       s_opts.mlock_str ? parse_mlock_mode(s_opts.mlock_str) : mlock_mode::NONE;
   s_opts.decompress_ratio = s_opts.decompress_ratio_str
-                              ? folly::to<double>(s_opts.decompress_ratio_str)
+                                ? folly::to<double>(s_opts.decompress_ratio_str)
-                              : 0.8;
+                                : 0.8;
 
   s_lgr.set_threshold(s_opts.debuglevel);
   log_proxy<debug_logger_policy> log(s_lgr);

src/dwarfs/entry.cpp

@@ -211,6 +211,24 @@ void dir::pack(thrift::metadata::metadata& mv2,
   }
 }
 
+void dir::remove_empty_dirs(progress& prog) {
+  auto last = std::remove_if(entries_.begin(), entries_.end(),
+                             [&](std::shared_ptr<entry> const& e) {
+                               if (auto d = dynamic_cast<dir*>(e.get())) {
+                                 d->remove_empty_dirs(prog);
+                                 return d->empty();
+                               }
+                               return false;
+                             });
+
+  if (last != entries_.end()) {
+    auto num = std::distance(last, entries_.end());
+    prog.dirs_scanned -= num;
+    prog.dirs_found -= num;
+    entries_.erase(last, entries_.end());
+  }
+}
+
 entry::type_t link::type() const { return E_LINK; }
 
 const std::string& link::linkname() const { return link_; }

src/dwarfs/scanner.cpp

@@ -451,6 +451,12 @@ void scanner_<LoggerPolicy>::scan(filesystem_writer& fsw,
 
   auto root = scan_tree(path, prog);
 
+  if (options_.remove_empty_dirs) {
+    log_.info() << "removing empty directories...";
+    auto d = dynamic_cast<dir*>(root.get());
+    d->remove_empty_dirs(prog);
+  }
+
   // now scan all files
   scan_files_visitor sfv(wg_, *os_, prog);
   root->accept(sfv);

src/mkdwarfs.cpp

@@ -328,6 +328,9 @@ int mkdwarfs(int argc, char** argv) {
         po::value<unsigned>(&cfg.window_increment_shift)
             ->default_value(1),
         "window increment (as right shift of size)")
+    ("remove-empty-dirs",
+        po::value<bool>(&options.remove_empty_dirs)->zero_tokens(),
+        "remove empty directories in file system")
     ("log-level",
         po::value<std::string>(&log_level)->default_value("info"),
         "log level (error, warn, info, debug, trace)")