Use walk_inode_order in dwarfsextract

src/dwarfsextract.cpp

@@ -144,22 +144,6 @@ int dwarfsextract(int argc, char** argv) {
                  : nullptr));
     }
 
-    // TODO: This can surely be implemented more efficiently inside libdwarfs
-    //       using entry indices / inode numbers, which should reduce the
-    //       amount of memory needed by roughly a factor of 8.
-
-    std::vector<std::pair<entry_view, directory_view>> entries;
-
-    fs.walk([&](auto entry, auto parent) {
-      if (!S_ISDIR(entry.mode())) {
-        entries.emplace_back(entry, parent);
-      }
-    });
-
-    std::sort(entries.begin(), entries.end(), [](auto& a, auto& b) {
-      return a.first.inode() < b.first.inode();
-    });
-
     auto lr = ::archive_entry_linkresolver_new();
 
     ::archive_entry_linkresolver_set_strategy(lr, ::archive_format(a));
@@ -178,7 +162,11 @@ int dwarfsextract(int argc, char** argv) {
       }
     };
 
-    for (auto& [entry, parent] : entries) {
+    fs.walk_inode_order([&](auto entry, auto parent) {
+      if (entry.inode() == 0) {
+        return;
+      }
+
       auto ae = ::archive_entry_new();
       struct ::stat stbuf;
 
@@ -221,7 +209,7 @@ int dwarfsextract(int argc, char** argv) {
         do_archive(spare, *ev);
         ::archive_entry_free(spare);
       }
-    }
+    });
 
     // As we're visiting *all* hardlinks, we should never see any deferred
     // entries.