From eb8803d6dfbeaf01a34a5ac5653cb48c369f96f2 Mon Sep 17 00:00:00 2001
From: Marcus Holland-Moritz <github@mhxnet.de>
Date: Sun, 6 Nov 2022 21:03:01 +0100
Subject: [PATCH] Add --input-list option to pass in a list of files

---
 doc/mkdwarfs.md          |  10 +-
 include/dwarfs/entry.h   |   6 +
 include/dwarfs/scanner.h |  12 +-
 src/dwarfs/entry.cpp     |  42 +++++-
 src/dwarfs/scanner.cpp   | 302 ++++++++++++++++++++++++++-------------
 src/mkdwarfs.cpp         |  49 ++++++-
 6 files changed, 315 insertions(+), 106 deletions(-)

diff --git a/doc/mkdwarfs.md b/doc/mkdwarfs.md
index eb20d546..b8bedc35 100644
--- a/doc/mkdwarfs.md
+++ b/doc/mkdwarfs.md
@@ -27,9 +27,17 @@ There two mandatory options for specifying the input and output:
 
 - `-i`, `--input=`*path*|*file*:
   Path to the root directory containing the files from which you want to
-  build a filesystem. If the `--recompress` option is given, this argument
+  build a file system. If the `--recompress` option is given, this argument
   is the source filesystem.
 
+- `--input-list=`*file*|`-`:
+  Read list of paths to add to the file system from this file or from stdin.
+  The pathames will be interpreted relative to the path given with `--input`.
+  If `--input` is omitted, the pathames will be interpreted relative to the
+  current directory. If you want files to be stored in the exact same order
+  as read from this list (because, for example, you have already sorted them
+  by similarity or access frequency), you must also pass `--order=none`.
+
 - `-o`, `--output=`*file*:
   File name of the output filesystem.
 
diff --git a/include/dwarfs/entry.h b/include/dwarfs/entry.h
index bd160ba8..eed7357a 100644
--- a/include/dwarfs/entry.h
+++ b/include/dwarfs/entry.h
@@ -174,11 +174,17 @@ class dir : public entry {
     return inode_num_;
   }
 
+  std::shared_ptr<entry> find(std::string_view name);
+
  private:
   using entry_ptr = std::shared_ptr<entry>;
+  using lookup_table = std::unordered_map<std::string_view, entry_ptr>;
+
+  void populate_lookup_table();
 
   std::vector<std::shared_ptr<entry>> entries_;
   std::optional<uint32_t> inode_num_;
+  std::unique_ptr<lookup_table> lookup_;
 };
 
 class link : public entry {
diff --git a/include/dwarfs/scanner.h b/include/dwarfs/scanner.h
index 6bc37fa9..4e38dbf2 100644
--- a/include/dwarfs/scanner.h
+++ b/include/dwarfs/scanner.h
@@ -21,7 +21,10 @@
 
 #pragma once
 
+#include <filesystem>
 #include <memory>
+#include <optional>
+#include <span>
 #include <string>
 
 #include "dwarfs/block_manager.h"
@@ -44,8 +47,10 @@ class scanner {
           std::shared_ptr<entry_factory> ef, std::shared_ptr<os_access> os,
           std::shared_ptr<script> scr, const scanner_options& options);
 
-  void scan(filesystem_writer& fsw, const std::string& path, progress& prog) {
-    impl_->scan(fsw, path, prog);
+  void scan(filesystem_writer& fsw, const std::string& path, progress& prog,
+            std::optional<std::span<std::filesystem::path const>> list =
+                std::nullopt) {
+    impl_->scan(fsw, path, prog, list);
   }
 
   class impl {
@@ -53,7 +58,8 @@ class scanner {
     virtual ~impl() = default;
 
     virtual void
-    scan(filesystem_writer& fsw, const std::string& path, progress& prog) = 0;
+    scan(filesystem_writer& fsw, const std::string& path, progress& prog,
+         std::optional<std::span<std::filesystem::path const>> list) = 0;
   };
 
  private:
diff --git a/src/dwarfs/entry.cpp b/src/dwarfs/entry.cpp
index 65aa99cf..844eac73 100644
--- a/src/dwarfs/entry.cpp
+++ b/src/dwarfs/entry.cpp
@@ -235,7 +235,13 @@ void file::hardlink(file* other, progress& prog) {
 
 entry::type_t dir::type() const { return E_DIR; }
 
-void dir::add(std::shared_ptr<entry> e) { entries_.emplace_back(std::move(e)); }
+void dir::add(std::shared_ptr<entry> e) {
+  if (lookup_) {
+    auto r [[maybe_unused]] = lookup_->emplace(e->name(), e);
+    assert(r.second);
+  }
+  entries_.emplace_back(std::move(e));
+}
 
 void dir::walk(std::function<void(entry*)> const& f) {
   f(this);
@@ -323,6 +329,40 @@ void dir::remove_empty_dirs(progress& prog) {
     prog.dirs_found -= num;
     entries_.erase(last, entries_.end());
   }
+
+  lookup_.reset();
+}
+
+std::shared_ptr<entry> dir::find(std::string_view name) {
+  if (!lookup_ && entries_.size() >= 16) {
+    populate_lookup_table();
+  }
+
+  if (lookup_) {
+    if (auto it = lookup_->find(name); it != lookup_->end()) {
+      return it->second;
+    }
+  } else {
+    auto it = std::find_if(entries_.begin(), entries_.end(),
+                           [name](auto& e) { return e->name() == name; });
+    if (it != entries_.end()) {
+      return *it;
+    }
+  }
+
+  return nullptr;
+}
+
+void dir::populate_lookup_table() {
+  assert(!lookup_);
+
+  lookup_ = std::make_unique<lookup_table>();
+  lookup_->reserve(entries_.size());
+
+  for (auto const& e : entries_) {
+    auto r [[maybe_unused]] = lookup_->emplace(e->name(), e);
+    assert(r.second);
+  }
 }
 
 entry::type_t link::type() const { return E_LINK; }
diff --git a/src/dwarfs/scanner.cpp b/src/dwarfs/scanner.cpp
index 86aeef73..a442d598 100644
--- a/src/dwarfs/scanner.cpp
+++ b/src/dwarfs/scanner.cpp
@@ -284,13 +284,23 @@ class scanner_ final : public scanner::impl {
            std::shared_ptr<entry_factory> ef, std::shared_ptr<os_access> os,
            std::shared_ptr<script> scr, const scanner_options& options);
 
-  void scan(filesystem_writer& fsw, const std::string& path,
-            progress& prog) override;
+  void
+  scan(filesystem_writer& fsw, const std::string& path, progress& prog,
+       std::optional<std::span<std::filesystem::path const>> list) override;
 
  private:
   std::shared_ptr<entry>
   scan_tree(const std::string& path, progress& prog, detail::file_scanner& fs);
 
+  std::shared_ptr<entry> scan_list(const std::string& path,
+                                   std::span<std::filesystem::path const> list,
+                                   progress& prog, detail::file_scanner& fs);
+
+  std::shared_ptr<entry>
+  add_entry(std::string const& name, std::shared_ptr<dir> parent,
+            progress& prog, detail::file_scanner& fs,
+            bool debug_filter = false);
+
   const block_manager::config& cfg_;
   const scanner_options& options_;
   std::shared_ptr<entry_factory> entry_;
@@ -317,6 +327,111 @@ scanner_<LoggerPolicy>::scanner_(logger& lgr, worker_group& wg,
     , lgr_(lgr)
     , LOG_PROXY_INIT(lgr_) {}
 
+template <typename LoggerPolicy>
+std::shared_ptr<entry>
+scanner_<LoggerPolicy>::add_entry(std::string const& name,
+                                  std::shared_ptr<dir> parent, progress& prog,
+                                  detail::file_scanner& fs, bool debug_filter) {
+  try {
+    auto pe = entry_->create(*os_, name, parent);
+    bool exclude = false;
+
+    if (script_) {
+      if (script_->has_filter() && !script_->filter(*pe)) {
+        exclude = true;
+      } else if (script_->has_transform()) {
+        script_->transform(*pe);
+      }
+    }
+
+    if (debug_filter) {
+      (*options_.debug_filter_function)(exclude, pe.get());
+    }
+
+    if (exclude) {
+      if (!debug_filter) {
+        LOG_DEBUG << "excluding " << pe->dpath();
+      }
+
+      return nullptr;
+    }
+
+    if (pe) {
+      switch (pe->type()) {
+      case entry::E_FILE:
+        if (os_->access(pe->path(), R_OK)) {
+          LOG_ERROR << "cannot access: " << pe->path();
+          prog.errors++;
+          return nullptr;
+        }
+        break;
+
+      case entry::E_DEVICE:
+        if (!options_.with_devices) {
+          return nullptr;
+        }
+        break;
+
+      case entry::E_OTHER:
+        if (!options_.with_specials) {
+          return nullptr;
+        }
+        break;
+
+      default:
+        break;
+      }
+
+      parent->add(pe);
+
+      switch (pe->type()) {
+      case entry::E_DIR:
+        // prog.current.store(pe.get());
+        prog.dirs_found++;
+        if (!debug_filter) {
+          pe->scan(*os_, prog);
+        }
+        break;
+
+      case entry::E_FILE:
+        prog.files_found++;
+        if (!debug_filter) {
+          fs.scan(dynamic_cast<file*>(pe.get()));
+        }
+        break;
+
+      case entry::E_LINK:
+        prog.symlinks_found++;
+        if (!debug_filter) {
+          pe->scan(*os_, prog);
+        }
+        prog.symlinks_scanned++;
+        break;
+
+      case entry::E_DEVICE:
+      case entry::E_OTHER:
+        prog.specials_found++;
+        if (!debug_filter) {
+          pe->scan(*os_, prog);
+        }
+        break;
+
+      default:
+        LOG_ERROR << "unsupported entry type: " << int(pe->type());
+        prog.errors++;
+        break;
+      }
+    }
+
+    return pe;
+  } catch (const std::system_error& e) {
+    LOG_ERROR << "error reading entry: " << e.what();
+    prog.errors++;
+  }
+
+  return nullptr;
+}
+
 template <typename LoggerPolicy>
 std::shared_ptr<entry>
 scanner_<LoggerPolicy>::scan_tree(const std::string& path, progress& prog,
@@ -349,100 +464,10 @@ scanner_<LoggerPolicy>::scan_tree(const std::string& path, progress& prog,
           continue;
         }
 
-        try {
-          auto pe = entry_->create(*os_, name, parent);
-          bool exclude = false;
-
-          if (script_) {
-            if (script_->has_filter() && !script_->filter(*pe)) {
-              exclude = true;
-            } else if (script_->has_transform()) {
-              script_->transform(*pe);
-            }
+        if (auto pe = add_entry(name, parent, prog, fs, debug_filter)) {
+          if (pe->type() == entry::E_DIR) {
+            subdirs.push_back(pe);
           }
-
-          if (debug_filter) {
-            (*options_.debug_filter_function)(exclude, pe.get());
-          }
-
-          if (exclude) {
-            if (!debug_filter) {
-              LOG_DEBUG << "excluding " << pe->dpath();
-            }
-
-            continue;
-          }
-
-          if (pe) {
-            switch (pe->type()) {
-            case entry::E_FILE:
-              if (os_->access(pe->path(), R_OK)) {
-                LOG_ERROR << "cannot access: " << pe->path();
-                prog.errors++;
-                continue;
-              }
-              break;
-
-            case entry::E_DEVICE:
-              if (!options_.with_devices) {
-                continue;
-              }
-              break;
-
-            case entry::E_OTHER:
-              if (!options_.with_specials) {
-                continue;
-              }
-              break;
-
-            default:
-              break;
-            }
-
-            parent->add(pe);
-
-            switch (pe->type()) {
-            case entry::E_DIR:
-              // prog.current.store(pe.get());
-              prog.dirs_found++;
-              if (!debug_filter) {
-                pe->scan(*os_, prog);
-              }
-              subdirs.push_back(pe);
-              break;
-
-            case entry::E_FILE:
-              prog.files_found++;
-              if (!debug_filter) {
-                fs.scan(dynamic_cast<file*>(pe.get()));
-              }
-              break;
-
-            case entry::E_LINK:
-              prog.symlinks_found++;
-              if (!debug_filter) {
-                pe->scan(*os_, prog);
-              }
-              prog.symlinks_scanned++;
-              break;
-
-            case entry::E_DEVICE:
-            case entry::E_OTHER:
-              prog.specials_found++;
-              if (!debug_filter) {
-                pe->scan(*os_, prog);
-              }
-              break;
-
-            default:
-              LOG_ERROR << "unsupported entry type: " << int(pe->type());
-              prog.errors++;
-              break;
-            }
-          }
-        } catch (const std::system_error& e) {
-          LOG_ERROR << "error reading entry: " << e.what();
-          prog.errors++;
         }
       }
 
@@ -459,8 +484,88 @@ scanner_<LoggerPolicy>::scan_tree(const std::string& path, progress& prog,
 }
 
 template <typename LoggerPolicy>
-void scanner_<LoggerPolicy>::scan(filesystem_writer& fsw,
-                                  const std::string& path, progress& prog) {
+std::shared_ptr<entry>
+scanner_<LoggerPolicy>::scan_list(const std::string& path,
+                                  std::span<std::filesystem::path const> list,
+                                  progress& prog, detail::file_scanner& fs) {
+  if (script_ && script_->has_filter()) {
+    DWARFS_THROW(runtime_error, "cannot use filters with file lists");
+  }
+
+  auto ti = LOG_TIMED_INFO;
+
+  auto root = entry_->create(*os_, path);
+  auto root_path = std::filesystem::path(path);
+
+  if (root->type() != entry::E_DIR) {
+    DWARFS_THROW(runtime_error, fmt::format("'{}' must be a directory", path));
+  }
+
+  auto ensure_path = [this, &prog, &fs](std::filesystem::path const& path,
+                                        std::shared_ptr<entry> root) {
+    for (auto const& p : path) {
+      if (auto d = std::dynamic_pointer_cast<dir>(root)) {
+        if (auto e = d->find(p.string())) {
+          root = e;
+        } else {
+          root = add_entry(p.string(), d, prog, fs);
+          if (root && root->type() == entry::E_DIR) {
+            prog.dirs_scanned++;
+          } else {
+            DWARFS_THROW(runtime_error,
+                         fmt::format("invalid path '{}'", path.string()));
+          }
+        }
+      } else {
+        DWARFS_THROW(runtime_error,
+                     fmt::format("invalid path '{}'", path.string()));
+      }
+    }
+
+    return root;
+  };
+
+  std::unordered_map<std::string, std::shared_ptr<dir>> dir_cache;
+
+  for (auto const& p : list) {
+    auto pp = p.parent_path();
+    std::shared_ptr<dir> pd;
+
+    if (auto it = dir_cache.find(pp.string()); it != dir_cache.end()) {
+      pd = it->second;
+    } else {
+      pd = std::dynamic_pointer_cast<dir>(ensure_path(pp, root));
+
+      if (pd) {
+        dir_cache.emplace(pp.string(), pd);
+      } else {
+        DWARFS_THROW(runtime_error,
+                     fmt::format("invalid path '{}'", p.string()));
+      }
+    }
+
+    auto const& fname = p.filename().string();
+
+    if (auto pe = pd->find(fname)) {
+      continue;
+    }
+
+    if (auto pe = add_entry(fname, pd, prog, fs)) {
+      if (pe->type() == entry::E_DIR) {
+        prog.dirs_scanned++;
+      }
+    }
+  }
+
+  ti << "scanned input list";
+
+  return root;
+}
+
+template <typename LoggerPolicy>
+void scanner_<LoggerPolicy>::scan(
+    filesystem_writer& fsw, const std::string& path, progress& prog,
+    std::optional<std::span<std::filesystem::path const>> list) {
   if (!options_.debug_filter_function) {
     LOG_INFO << "scanning " << path;
   }
@@ -471,7 +576,8 @@ void scanner_<LoggerPolicy>::scan(filesystem_writer& fsw,
   detail::file_scanner fs(wg_, *os_, im, options_.inode,
                           options_.file_hash_algorithm, prog);
 
-  auto root = scan_tree(path, prog, fs);
+  auto root =
+      list ? scan_list(path, *list, prog, fs) : scan_tree(path, prog, fs);
 
   if (options_.debug_filter_function) {
     return;
diff --git a/src/mkdwarfs.cpp b/src/mkdwarfs.cpp
index ebe36258..40022d99 100644
--- a/src/mkdwarfs.cpp
+++ b/src/mkdwarfs.cpp
@@ -372,7 +372,7 @@ int mkdwarfs(int argc, char** argv) {
   std::string path, output, memory_limit, script_arg, compression, header,
       schema_compression, metadata_compression, log_level_str, timestamp,
       time_resolution, order, progress_mode, recompress_opts, pack_metadata,
-      file_hash_algo, debug_filter, max_similarity_size;
+      file_hash_algo, debug_filter, max_similarity_size, input_list_str;
   std::vector<std::string> filter;
   size_t num_workers, num_scanner_workers;
   bool no_progress = false, remove_header = false, no_section_index = false,
@@ -413,6 +413,9 @@ int mkdwarfs(int argc, char** argv) {
     ("force,f",
         po::value<bool>(&force_overwrite)->zero_tokens(),
         "force overwrite of existing output image")
+    ("input-list",
+        po::value<std::string>(&input_list_str),
+        "file containing list of paths relative to root directory")
     ("compress-level,l",
         po::value<unsigned>(&level)->default_value(default_level),
         "compression level (0=fast, 9=best, please see man page for details)")
@@ -548,7 +551,7 @@ int mkdwarfs(int argc, char** argv) {
     return 1;
   }
 
-  if (vm.count("help") or !vm.count("input") or
+  if (vm.count("help") or !(vm.count("input") or vm.count("input-list")) or
       (!vm.count("output") and !vm.count("debug-filter"))) {
     size_t l_dc = 0, l_sc = 0, l_mc = 0, l_or = 0;
     for (auto const& l : levels) {
@@ -643,6 +646,42 @@ int mkdwarfs(int argc, char** argv) {
     return 1;
   }
 
+  std::optional<std::vector<std::filesystem::path>> input_list;
+
+  if (vm.count("input-list")) {
+    if (vm.count("filter")) {
+      std::cerr << "error: cannot use --input-list and --filter\n";
+      return 1;
+    }
+
+    if (!vm.count("input")) {
+      path = std::filesystem::current_path().string();
+    }
+
+    std::unique_ptr<std::ifstream> ifs;
+    std::istream* is;
+
+    if (input_list_str == "-") {
+      is = &std::cin;
+    } else {
+      ifs = std::make_unique<std::ifstream>(input_list_str);
+
+      if (!ifs->is_open()) {
+        throw std::runtime_error(
+            fmt::format("error opening file: {}", input_list_str));
+      }
+
+      is = ifs.get();
+    }
+
+    std::string line;
+    input_list.emplace();
+
+    while (std::getline(*is, line)) {
+      input_list->emplace_back(line);
+    }
+  }
+
   bool recompress = vm.count("recompress");
   rewrite_options rw_opts;
   if (recompress) {
@@ -1011,7 +1050,11 @@ int mkdwarfs(int argc, char** argv) {
                 std::make_shared<os_access_posix>(), std::move(script),
                 options);
 
-      s.scan(fsw, path, prog);
+      if (input_list) {
+        s.scan(fsw, path, prog, *input_list);
+      } else {
+        s.scan(fsw, path, prog);
+      }
     }
   } catch (runtime_error const& e) {
     LOG_ERROR << e.what();