Introduce 'exclusions' parameter to scanner

datman/README.md

@@ -8,5 +8,6 @@ Features:
 * (optional) Compression using Zstd and a specifiable dictionary
 * (optional) Encryption
 * Ability to back up to remote machines over SSH
+* Labelling of files in a backup source; different destinations can choose to backup either all or a subset of the labels.
 
 See the documentation for more information.

datman/src/commands/backup.rs

@@ -23,7 +23,7 @@ use anyhow::{anyhow, bail};
 use arc_interner::ArcIntern;
 use chrono::{DateTime, NaiveDateTime, TimeZone, Utc};
 use log::{info, warn};
-use std::collections::{HashMap, HashSet};
+use std::collections::{BTreeSet, HashMap, HashSet};
 use std::fmt::Debug;
 use std::io::Write;
 use std::path::Path;
@@ -151,7 +151,7 @@ pub fn backup_source_to_destination<PT: ProgressTracker>(
         } => {
             info!("Looking to backup {} to {}", source_name, dest_name);
             info!("Scanning.");
-            let tree = scan(directory, !*cross_filesystems)?
+            let tree = scan(directory, !*cross_filesystems, &BTreeSet::new())?
                 .ok_or_else(|| anyhow!("Source does not exist."))?;
 
             let absolute_source_path = desc_path.join(directory);

datman/src/commands/ibrowse.rs

@@ -15,6 +15,7 @@ You should have received a copy of the GNU General Public License
 along with Yama.  If not, see <https://www.gnu.org/licenses/>.
 */
 
+use std::collections::BTreeSet;
 use std::path::Path;
 
 use anyhow::{anyhow, bail};
@@ -80,7 +81,7 @@ pub fn session(path: &Path, source_name: String) -> anyhow::Result<()> {
     };
 
     println!("Scanning source; this might take a little while...");
-    let mut dir_scan: FileTree1<Option<State>> = scan(directory, one_filesystem)?
+    let mut dir_scan: FileTree1<Option<State>> = scan(directory, one_filesystem, &BTreeSet::new())?
         .ok_or_else(|| anyhow!("Empty source."))?
         .replace_meta(&None);
 

datman/src/commands/ilabel.rs

@@ -15,6 +15,7 @@ You should have received a copy of the GNU General Public License
 along with Yama.  If not, see <https://www.gnu.org/licenses/>.
 */
 
+use std::collections::BTreeSet;
 use std::io;
 use std::io::{StdinLock, Stdout, Write};
 use std::path::Path;
@@ -192,7 +193,7 @@ pub fn interactive_labelling_session(path: &Path, source_name: String) -> anyhow
         let my_hostname = get_hostname();
         let mut dir_scan = if &my_hostname == hostname {
             info!("Scanning source; this might take a little while...");
-            scan(directory, !*cross_filesystems)?
+            scan(directory, !*cross_filesystems, &BTreeSet::new())?
                 .ok_or_else(|| anyhow!("Empty source."))?
                 .replace_meta(&None)
         } else {

datman/src/remote/backup_source_responder.rs

@@ -1,6 +1,7 @@
 // This file implements the responder side of the backup source protocol -- the protocol used
 // to connect to remote backup sources.
 
+use std::collections::BTreeSet;
 use std::io::{stdin, stdout, Read, Write};
 use std::path::PathBuf;
 use std::sync::Arc;
@@ -43,7 +44,7 @@ pub fn introduction<R: Read, W: Write>(read: &mut R, write: &mut W) -> anyhow::R
 pub fn scanning<R: Read, W: Write>(read: &mut R, write: &mut W) -> anyhow::Result<()> {
     let path: PathBuf = read_message(read)?;
     let one_filesystem: bool = read_message(read)?;
-    let scan_result = scan(&path, one_filesystem)?;
+    let scan_result = scan(&path, one_filesystem, &BTreeSet::new())?;
     write_message(write, &scan_result)?;
     write.flush()?;
     Ok(())

datman/src/tree.rs

@@ -15,12 +15,12 @@ You should have received a copy of the GNU General Public License
 along with Yama.  If not, see <https://www.gnu.org/licenses/>.
 */
 
-use std::collections::BTreeMap;
+use std::collections::{BTreeMap, BTreeSet};
 use std::fmt::Debug;
 use std::fs::{read_link, symlink_metadata, DirEntry, Metadata};
 use std::io::ErrorKind;
 use std::os::unix::fs::MetadataExt;
-use std::path::Path;
+use std::path::{Path, PathBuf};
 
 use anyhow::anyhow;
 use indicatif::{ProgressBar, ProgressDrawTarget, ProgressStyle};
@@ -216,14 +216,18 @@ pub fn mtime_msec(metadata: &Metadata) -> u64 {
 }
 
 /// Scan the filesystem to produce a Tree, using a default progress bar.
-pub fn scan(path: &Path, one_filesystem: bool) -> anyhow::Result<Option<FileTree<(), (), (), ()>>> {
+pub fn scan(
+    path: &Path,
+    one_filesystem: bool,
+    exclusions: &BTreeSet<PathBuf>,
+) -> anyhow::Result<Option<FileTree<(), (), (), ()>>> {
     let pbar = ProgressBar::with_draw_target(0, ProgressDrawTarget::stdout_with_hz(2));
     pbar.set_style(ProgressStyle::default_spinner().template("{spinner} {pos:7} {msg}"));
     pbar.set_message("dir scan");
 
     let one_filesystem = if one_filesystem { Some(None) } else { None };
 
-    let result = scan_with_progress_bar(path, &pbar, one_filesystem);
+    let result = scan_with_progress_bar(path, &pbar, one_filesystem, exclusions);
     pbar.finish_at_current_pos();
     result
 }
@@ -233,7 +237,13 @@ pub fn scan_with_progress_bar(
     path: &Path,
     progress_bar: &ProgressBar,
     mut one_filesystem: Option<Option<u64>>,
+    exclusions: &BTreeSet<PathBuf>,
 ) -> anyhow::Result<Option<FileTree<(), (), (), ()>>> {
+    if exclusions.contains(path) {
+        // Don't enter excluded paths.
+        return Ok(None);
+    }
+
     let metadata_res = symlink_metadata(path);
     progress_bar.inc(1);
     if let Err(e) = &metadata_res {
@@ -305,7 +315,8 @@ pub fn scan_with_progress_bar(
 
         for entry in dir_read? {
             let entry: DirEntry = entry?;
-            let scanned = scan_with_progress_bar(&entry.path(), progress_bar, one_filesystem)?;
+            let scanned =
+                scan_with_progress_bar(&entry.path(), progress_bar, one_filesystem, exclusions)?;
             if let Some(scanned) = scanned {
                 if let Ok(filename) = entry.file_name().into_string() {
                     children.insert(filename, scanned);