Add version for 0.10.0

init: Add `--copy-chunker-parameters` option

.github/ISSUE_TEMPLATE/Feature.md

@@ -39,8 +39,8 @@ Please describe the feature you'd like us to add here.
 -->
 
 
-What are you trying to do?
---------------------------
+What are you trying to do? What problem would this solve?
+---------------------------------------------------------
 
 <!--
 This section should contain a brief description what you're trying to do, which

.github/PULL_REQUEST_TEMPLATE.md

@@ -10,11 +10,11 @@ your time and add more commits. If you're done and ready for review, please
 check the last box.
 -->
 
-What is the purpose of this change? What does it change?
---------------------------------------------------------
+What does this PR change? What problem does it solve?
+-----------------------------------------------------
 
 <!--
-Describe the changes here, as detailed as needed.
+Describe the changes and their purpose here, as detailed as needed.
 -->
 
 Was the change discussed in an issue or in the forum before?
@@ -23,8 +23,8 @@ Was the change discussed in an issue or in the forum before?
 <!--
 Link issues and relevant forum posts here.
 
-If this PR resolves an issue on GitHub, use "closes #1234" so that the issue is
-closed automatically when this PR is merged.
+If this PR resolves an issue on GitHub, use "Closes #1234" so that the issue
+is closed automatically when this PR is merged.
 -->
 
 Checklist

.travis.yml

@@ -3,22 +3,6 @@ sudo: false
 
 matrix:
   include:
-    - os: linux
-      go: "1.11.x"
-      env: RESTIC_TEST_FUSE=0 RESTIC_TEST_CLOUD_BACKENDS=0
-      cache:
-        directories:
-          - $HOME/.cache/go-build
-          - $HOME/gopath/pkg/mod
-
-    - os: linux
-      go: "1.12.x"
-      env: RESTIC_TEST_FUSE=0 RESTIC_TEST_CLOUD_BACKENDS=0
-      cache:
-        directories:
-          - $HOME/.cache/go-build
-          - $HOME/gopath/pkg/mod
-
     - os: linux
       go: "1.13.x"
       env: RESTIC_TEST_FUSE=0 RESTIC_TEST_CLOUD_BACKENDS=0
@@ -27,9 +11,17 @@ matrix:
           - $HOME/.cache/go-build
           - $HOME/gopath/pkg/mod
 
-    # only run fuse and cloud backends tests on Travis for the latest Go on Linux
     - os: linux
       go: "1.14.x"
+      env: RESTIC_TEST_FUSE=0 RESTIC_TEST_CLOUD_BACKENDS=0
+      cache:
+        directories:
+          - $HOME/.cache/go-build
+          - $HOME/gopath/pkg/mod
+
+    # only run fuse and cloud backends tests on Travis for the latest Go on Linux
+    - os: linux
+      go: "1.15.x"
       sudo: true
       cache:
         directories:
@@ -37,7 +29,7 @@ matrix:
           - $HOME/gopath/pkg/mod
 
     - os: osx
-      go: "1.14.x"
+      go: "1.15.x"
       env: RESTIC_TEST_FUSE=0 RESTIC_TEST_CLOUD_BACKENDS=0
       cache:
         directories:

CHANGELOG.md

@@ -1,3 +1,467 @@
+Changelog for restic 0.10.0 (2020-09-19)
+=======================================
+
+The following sections list the changes in restic 0.10.0 relevant to
+restic users. The changes are ordered by importance.
+
+Summary
+-------
+
+ * Fix #1863: Report correct number of directories processed by backup
+ * Fix #2254: Fix tar issues when dumping `/`
+ * Fix #2281: Handle format verbs like '%' properly in `find` output
+ * Fix #2298: Do not hang when run as a background job
+ * Fix #2389: Fix mangled json output of backup command
+ * Fix #2390: Refresh lock timestamp
+ * Fix #2429: Backup --json reports total_bytes_processed as 0
+ * Fix #2469: Fix incorrect bytes stats in `diff` command
+ * Fix #2518: Do not crash with Synology NAS sftp server
+ * Fix #2531: Fix incorrect size calculation in `stats --mode restore-size`
+ * Fix #2537: Fix incorrect file counts in `stats --mode restore-size`
+ * Fix #2592: SFTP backend supports IPv6 addresses
+ * Fix #2607: Honor RESTIC_CACHE_DIR environment variable on Mac and Windows
+ * Fix #2668: Don't abort the stats command when data blobs are missing
+ * Fix #2674: Add stricter prune error checks
+ * Fix #2899: Fix possible crash in the progress bar of check --read-data
+ * Chg #2482: Remove vendored dependencies
+ * Chg #2546: Return exit code 3 when failing to backup all source data
+ * Chg #2600: Update dependencies, require Go >= 1.13
+ * Chg #1597: Honor the --no-lock flag in the mount command
+ * Enh #1570: Support specifying multiple host flags for various commands
+ * Enh #1680: Optimize `restic mount`
+ * Enh #2072: Display snapshot date when using `restic find`
+ * Enh #2175: Allow specifying user and host when creating keys
+ * Enh #2277: Add support for ppc64le
+ * Enh #2395: Ignore sync errors when operation not supported by local filesystem
+ * Enh #2427: Add flag `--iexclude-file` to backup command
+ * Enh #2569: Support excluding files by their size
+ * Enh #2571: Self-heal missing file parts during backup of unchanged files
+ * Enh #2858: Support filtering snapshots by tag and path in the stats command
+ * Enh #323: Add command for copying snapshots between repositories
+ * Enh #551: Use optimized library for hash calculation of file chunks
+ * Enh #2195: Simplify and improve restore performance
+ * Enh #2328: Improve speed of check command
+ * Enh #2423: Support user@domain parsing as user
+ * Enh #2576: Improve the chunking algorithm
+ * Enh #2598: Improve speed of diff command
+ * Enh #2599: Slightly reduce memory usage of prune and stats commands
+ * Enh #2733: S3 backend: Add support for WebIdentityTokenFile
+ * Enh #2773: Optimize handling of new index entries
+ * Enh #2781: Reduce memory consumption of in-memory index
+ * Enh #2786: Optimize `list blobs` command
+ * Enh #2790: Optimized file access in restic mount
+ * Enh #2840: Speed-up file deletion in forget, prune and rebuild-index
+
+Details
+-------
+
+ * Bugfix #1863: Report correct number of directories processed by backup
+
+   The directory statistics calculation was fixed to report the actual number of processed
+   directories instead of always zero.
+
+   https://github.com/restic/restic/issues/1863
+
+ * Bugfix #2254: Fix tar issues when dumping `/`
+
+   We've fixed an issue with dumping either `/` or files on the first sublevel e.g. `/foo` to tar.
+   This also fixes tar dumping issues on Windows where this issue could also happen.
+
+   https://github.com/restic/restic/issues/2254
+   https://github.com/restic/restic/issues/2357
+   https://github.com/restic/restic/pull/2255
+
+ * Bugfix #2281: Handle format verbs like '%' properly in `find` output
+
+   The JSON or "normal" output of the `find` command can now deal with file names that contain
+   substrings which the Golang `fmt` package considers "format verbs" like `%s`.
+
+   https://github.com/restic/restic/issues/2281
+
+ * Bugfix #2298: Do not hang when run as a background job
+
+   Restic did hang on exit while restoring the terminal configuration when it was started as a
+   background job, for example using `restic ... &`. This has been fixed by only restoring the
+   terminal configuration when restic is interrupted while reading a password from the
+   terminal.
+
+   https://github.com/restic/restic/issues/2298
+
+ * Bugfix #2389: Fix mangled json output of backup command
+
+   We've fixed a race condition in the json output of the backup command that could cause multiple
+   lines to get mixed up. We've also ensured that the backup summary is printed last.
+
+   https://github.com/restic/restic/issues/2389
+   https://github.com/restic/restic/pull/2545
+
+ * Bugfix #2390: Refresh lock timestamp
+
+   Long-running operations did not refresh lock timestamp, resulting in locks becoming stale.
+   This is now fixed.
+
+   https://github.com/restic/restic/issues/2390
+
+ * Bugfix #2429: Backup --json reports total_bytes_processed as 0
+
+   We've fixed the json output of total_bytes_processed. The non-json output was already fixed
+   with pull request #2138 but left the json output untouched.
+
+   https://github.com/restic/restic/issues/2429
+
+ * Bugfix #2469: Fix incorrect bytes stats in `diff` command
+
+   In some cases, the wrong number of bytes (e.g. 16777215.998 TiB) were reported by the `diff`
+   command. This is now fixed.
+
+   https://github.com/restic/restic/issues/2469
+
+ * Bugfix #2518: Do not crash with Synology NAS sftp server
+
+   It was found that when restic is used to store data on an sftp server on a Synology NAS with a
+   relative path (one which does not start with a slash), it may go into an endless loop trying to
+   create directories on the server. We've fixed this bug by using a function in the sftp library
+   instead of our own implementation.
+
+   The bug was discovered because the Synology sftp server behaves erratic with non-absolute
+   path (e.g. `home/restic-repo`). This can be resolved by just using an absolute path instead
+   (`/home/restic-repo`). We've also added a paragraph in the FAQ.
+
+   https://github.com/restic/restic/issues/2518
+   https://github.com/restic/restic/issues/2363
+   https://github.com/restic/restic/pull/2530
+
+ * Bugfix #2531: Fix incorrect size calculation in `stats --mode restore-size`
+
+   The restore-size mode of stats was counting hard-linked files as if they were independent.
+
+   https://github.com/restic/restic/issues/2531
+
+ * Bugfix #2537: Fix incorrect file counts in `stats --mode restore-size`
+
+   The restore-size mode of stats was failing to count empty directories and some files with hard
+   links.
+
+   https://github.com/restic/restic/issues/2537
+
+ * Bugfix #2592: SFTP backend supports IPv6 addresses
+
+   The SFTP backend now supports IPv6 addresses natively, without relying on aliases in the
+   external SSH configuration.
+
+   https://github.com/restic/restic/pull/2592
+
+ * Bugfix #2607: Honor RESTIC_CACHE_DIR environment variable on Mac and Windows
+
+   On Mac and Windows, the RESTIC_CACHE_DIR environment variable was ignored. This variable can
+   now be used on all platforms to set the directory where restic stores caches.
+
+   https://github.com/restic/restic/pull/2607
+
+ * Bugfix #2668: Don't abort the stats command when data blobs are missing
+
+   Runing the stats command in the blobs-per-file mode on a repository with missing data blobs
+   previously resulted in a crash.
+
+   https://github.com/restic/restic/pull/2668
+
+ * Bugfix #2674: Add stricter prune error checks
+
+   Additional checks were added to the prune command in order to improve resiliency to backend,
+   hardware and/or networking issues. The checks now detect a few more cases where such outside
+   factors could potentially cause data loss.
+
+   https://github.com/restic/restic/pull/2674
+
+ * Bugfix #2899: Fix possible crash in the progress bar of check --read-data
+
+   We've fixed a possible crash while displaying the progress bar for the check --read-data
+   command. The crash occurred when the length of the progress bar status exceeded the terminal
+   width, which only happened for very narrow terminal windows.
+
+   https://github.com/restic/restic/pull/2899
+   https://forum.restic.net/t/restic-rclone-pcloud-connection-issues/2963/15
+
+ * Change #2482: Remove vendored dependencies
+
+   We've removed the vendored dependencies (in the subdir `vendor/`). When building restic, the
+   Go compiler automatically fetches the dependencies. It will also cryptographically verify
+   that the correct code has been fetched by using the hashes in `go.sum` (see the link to the
+   documentation below).
+
+   https://github.com/restic/restic/issues/2482
+   https://golang.org/cmd/go/#hdr-Module_downloading_and_verification
+
+ * Change #2546: Return exit code 3 when failing to backup all source data
+
+   The backup command used to return a zero exit code as long as a snapshot could be created
+   successfully, even if some of the source files could not be read (in which case the snapshot
+   would contain the rest of the files).
+
+   This made it hard for automation/scripts to detect failures/incomplete backups by looking at
+   the exit code. Restic now returns the following exit codes for the backup command:
+
+   - 0 when the command was successful - 1 when there was a fatal error (no snapshot created) - 3 when
+   some source data could not be read (incomplete snapshot created)
+
+   https://github.com/restic/restic/issues/956
+   https://github.com/restic/restic/issues/2064
+   https://github.com/restic/restic/issues/2526
+   https://github.com/restic/restic/issues/2364
+   https://github.com/restic/restic/pull/2546
+
+ * Change #2600: Update dependencies, require Go >= 1.13
+
+   Restic now requires Go to be at least 1.13. This allows simplifications in the build process and
+   removing workarounds.
+
+   This is also probably the last version of restic still supporting mounting repositories via
+   fuse on macOS. The library we're using for fuse does not support macOS any more and osxfuse is not
+   open source any more.
+
+   https://github.com/bazil/fuse/issues/224
+   https://github.com/osxfuse/osxfuse/issues/590
+   https://github.com/restic/restic/pull/2600
+   https://github.com/restic/restic/pull/2852
+   https://github.com/restic/restic/pull/2927
+
+ * Change #1597: Honor the --no-lock flag in the mount command
+
+   The mount command now does not lock the repository if given the --no-lock flag. This allows to
+   mount repositories which are archived on a read only backend/filesystem.
+
+   https://github.com/restic/restic/issues/1597
+   https://github.com/restic/restic/pull/2821
+
+ * Enhancement #1570: Support specifying multiple host flags for various commands
+
+   Previously commands didn't take more than one `--host` or `-H` argument into account, which
+   could be limiting with e.g. the `forget` command.
+
+   The `dump`, `find`, `forget`, `ls`, `mount`, `restore`, `snapshots`, `stats` and `tag`
+   commands will now take into account multiple `--host` and `-H` flags.
+
+   https://github.com/restic/restic/issues/1570
+
+ * Enhancement #1680: Optimize `restic mount`
+
+   We've optimized the FUSE implementation used within restic. `restic mount` is now more
+   responsive and uses less memory.
+
+   https://github.com/restic/restic/issues/1680
+   https://github.com/restic/restic/pull/2587
+   https://github.com/restic/restic/pull/2787
+
+ * Enhancement #2072: Display snapshot date when using `restic find`
+
+   Added the respective snapshot date to the output of `restic find`.
+
+   https://github.com/restic/restic/issues/2072
+
+ * Enhancement #2175: Allow specifying user and host when creating keys
+
+   When adding a new key to the repository, the username and hostname for the new key can be
+   specified on the command line. This allows overriding the defaults, for example if you would
+   prefer to use the FQDN to identify the host or if you want to add keys for several different hosts
+   without having to run the key add command on those hosts.
+
+   https://github.com/restic/restic/issues/2175
+
+ * Enhancement #2277: Add support for ppc64le
+
+   Adds support for ppc64le, the processor architecture from IBM.
+
+   https://github.com/restic/restic/issues/2277
+
+ * Enhancement #2395: Ignore sync errors when operation not supported by local filesystem
+
+   The local backend has been modified to work with filesystems which doesn't support the `sync`
+   operation. This operation is normally used by restic to ensure that data files are fully
+   written to disk before continuing.
+
+   For these limited filesystems, saving a file in the backend would previously fail with an
+   "operation not supported" error. This error is now ignored, which means that e.g. an SMB mount
+   on macOS can now be used as storage location for a repository.
+
+   https://github.com/restic/restic/issues/2395
+   https://forum.restic.net/t/sync-errors-on-mac-over-smb/1859
+
+ * Enhancement #2427: Add flag `--iexclude-file` to backup command
+
+   The backup command now supports the flag `--iexclude-file` which is a case-insensitive
+   version of `--exclude-file`.
+
+   https://github.com/restic/restic/issues/2427
+   https://github.com/restic/restic/pull/2898
+
+ * Enhancement #2569: Support excluding files by their size
+
+   The `backup` command now supports the `--exclude-larger-than` option to exclude files which
+   are larger than the specified maximum size. This can for example be useful to exclude
+   unimportant files with a large file size.
+
+   https://github.com/restic/restic/issues/2569
+   https://github.com/restic/restic/pull/2914
+
+ * Enhancement #2571: Self-heal missing file parts during backup of unchanged files
+
+   We've improved the resilience of restic to certain types of repository corruption.
+
+   For files that are unchanged since the parent snapshot, the backup command now verifies that
+   all parts of the files still exist in the repository. Parts that are missing, e.g. from a damaged
+   repository, are backed up again. This verification was already run for files that were
+   modified since the parent snapshot, but is now also done for unchanged files.
+
+   Note that restic will not backup file parts that are referenced in the index but where the actual
+   data is not present on disk, as this situation can only be detected by restic check. Please
+   ensure that you run `restic check` regularly.
+
+   https://github.com/restic/restic/issues/2571
+   https://github.com/restic/restic/pull/2827
+
+ * Enhancement #2858: Support filtering snapshots by tag and path in the stats command
+
+   We've added filtering snapshots by `--tag tagList` and by `--path path` to the `stats`
+   command. This includes filtering of only 'latest' snapshots or all snapshots in a repository.
+
+   https://github.com/restic/restic/issues/2858
+   https://github.com/restic/restic/pull/2859
+   https://forum.restic.net/t/stats-for-a-host-and-filtered-snapshots/3020
+
+ * Enhancement #323: Add command for copying snapshots between repositories
+
+   We've added a copy command, allowing you to copy snapshots from one repository to another.
+
+   Note that this process will have to read (download) and write (upload) the entire snapshot(s)
+   due to the different encryption keys used on the source and destination repository. Also, the
+   transferred files are not re-chunked, which may break deduplication between files already
+   stored in the destination repo and files copied there using this command.
+
+   To fully support deduplication between repositories when the copy command is used, the init
+   command now supports the `--copy-chunker-params` option, which initializes the new
+   repository with identical parameters for splitting files into chunks as an already existing
+   repository. This allows copied snapshots to be equally deduplicated in both repositories.
+
+   https://github.com/restic/restic/issues/323
+   https://github.com/restic/restic/pull/2606
+   https://github.com/restic/restic/pull/2928
+
+ * Enhancement #551: Use optimized library for hash calculation of file chunks
+
+   We've switched the library used to calculate the hashes of file chunks, which are used for
+   deduplication, to the optimized Minio SHA-256 implementation.
+
+   Depending on the CPU it improves the hashing throughput by 10-30%. Modern x86 CPUs with the SHA
+   Extension should be about two to three times faster.
+
+   https://github.com/restic/restic/issues/551
+   https://github.com/restic/restic/pull/2709
+
+ * Enhancement #2195: Simplify and improve restore performance
+
+   Significantly improves restore performance of large files (i.e. 50M+):
+   https://github.com/restic/restic/issues/2074
+   https://forum.restic.net/t/restore-using-rclone-gdrive-backend-is-slow/1112/8
+   https://forum.restic.net/t/degraded-restore-performance-s3-backend/1400
+
+   Fixes "not enough cache capacity" error during restore:
+   https://github.com/restic/restic/issues/2244
+
+   NOTE: This new implementation does not guarantee order in which blobs are written to the target
+   files and, for example, the last blob of a file can be written to the file before any of the
+   preceeding file blobs. It is therefore possible to have gaps in the data written to the target
+   files if restore fails or interrupted by the user.
+
+   The implementation will try to preallocate space for the restored files on the filesystem to
+   prevent file fragmentation. This ensures good read performance for large files, like for
+   example VM images. If preallocating space is not supported by the filesystem, then this step is
+   silently skipped.
+
+   https://github.com/restic/restic/pull/2195
+   https://github.com/restic/restic/pull/2893
+
+ * Enhancement #2328: Improve speed of check command
+
+   We've improved the check command to traverse trees only once independent of whether they are
+   contained in multiple snapshots. The check command is now much faster for repositories with a
+   large number of snapshots.
+
+   https://github.com/restic/restic/issues/2284
+   https://github.com/restic/restic/pull/2328
+
+ * Enhancement #2423: Support user@domain parsing as user
+
+   Added the ability for user@domain-like users to be authenticated over SFTP servers.
+
+   https://github.com/restic/restic/pull/2423
+
+ * Enhancement #2576: Improve the chunking algorithm
+
+   We've updated the chunker library responsible for splitting files into smaller blocks. It
+   should improve the chunking throughput by 5-15% depending on the CPU.
+
+   https://github.com/restic/restic/issues/2820
+   https://github.com/restic/restic/pull/2576
+   https://github.com/restic/restic/pull/2845
+
+ * Enhancement #2598: Improve speed of diff command
+
+   We've improved the performance of the diff command when comparing snapshots with similar
+   content. It should run up to twice as fast as before.
+
+   https://github.com/restic/restic/pull/2598
+
+ * Enhancement #2599: Slightly reduce memory usage of prune and stats commands
+
+   The prune and the stats command kept directory identifiers in memory twice while searching for
+   used blobs.
+
+   https://github.com/restic/restic/pull/2599
+
+ * Enhancement #2733: S3 backend: Add support for WebIdentityTokenFile
+
+   We've added support for EKS IAM roles for service accounts feature to the S3 backend.
+
+   https://github.com/restic/restic/issues/2703
+   https://github.com/restic/restic/pull/2733
+
+ * Enhancement #2773: Optimize handling of new index entries
+
+   Restic now uses less memory for backups which add a lot of data, e.g. large initial backups. In
+   addition, we've improved the stability in some edge cases.
+
+   https://github.com/restic/restic/pull/2773
+
+ * Enhancement #2781: Reduce memory consumption of in-memory index
+
+   We've improved how the index is stored in memory. This change can reduce memory usage for large
+   repositories by up to 50% (depending on the operation).
+
+   https://github.com/restic/restic/pull/2781
+   https://github.com/restic/restic/pull/2812
+
+ * Enhancement #2786: Optimize `list blobs` command
+
+   We've changed the implementation of `list blobs` which should be now a bit faster and consume
+   almost no memory even for large repositories.
+
+   https://github.com/restic/restic/pull/2786
+
+ * Enhancement #2790: Optimized file access in restic mount
+
+   Reading large (> 100GiB) files from restic mountpoints is now faster, and the speedup is
+   greater for larger files.
+
+   https://github.com/restic/restic/pull/2790
+
+ * Enhancement #2840: Speed-up file deletion in forget, prune and rebuild-index
+
+   We've sped up the file deletion for the commands forget, prune and rebuild-index, especially
+   for remote repositories. Deletion was sequential before and is now run in parallel.
+
+   https://github.com/restic/restic/pull/2840
+
+
 Changelog for restic 0.9.6 (2019-11-22)
 =======================================
 
@@ -1361,10 +1825,10 @@ Details
 
    Exploiting the vulnerability requires a Linux/Unix system which saves backups via restic and
    a Windows systems which restores files from the repo. In addition, the attackers need to be able
-   to create create files with arbitrary names which are then saved to the restic repo. For
-   example, by creating a file named "..\test.txt" (which is a perfectly legal filename on Linux)
-   and restoring a snapshot containing this file on Windows, it would be written to the parent of
-   the target directory.
+   to create files with arbitrary names which are then saved to the restic repo. For example, by
+   creating a file named "..\test.txt" (which is a perfectly legal filename on Linux) and
+   restoring a snapshot containing this file on Windows, it would be written to the parent of the
+   target directory.
 
    We'd like to thank Tyler Spivey for reporting this responsibly!
 

CONTRIBUTING.md

@@ -60,16 +60,11 @@ uploading it somewhere or post only the parts that are really relevant.
 Development Environment
 =======================
 
-The repository contains several sets of directories with code: `cmd/` and
-`internal/` contain the code written for restic, whereas `vendor/` contains
-copies of libraries restic depends on. The libraries are managed with the
-command `go mod vendor`.
+The repository contains the code written for restic in the directories
+`cmd/` and `internal/`.
 
-Go >= 1.11
-----------
-
-For Go version 1.11 or later, you should clone the repo (without having
-`$GOPATH` set) and `cd` into the directory:
+Restic requires Go version 1.12 or later for compiling. Clone the repo (without
+having `$GOPATH` set) and `cd` into the directory:
 
     $ unset GOPATH
     $ git clone https://github.com/restic/restic
@@ -79,40 +74,12 @@ Then use the `go` tool to build restic:
 
     $ go build ./cmd/restic
     $ ./restic version
-    restic 0.9.2-dev (compiled manually) compiled with go1.11 on linux/amd64
+    restic 0.9.6-dev (compiled manually) compiled with go1.14 on linux/amd64
 
 You can run all tests with the following command:
 
     $ go test ./...
 
-Go < 1.11
----------
-
-In order to compile restic with Go before 1.11, it needs to be checked out at
-the right path within a `GOPATH`. The concept of a `GOPATH` is explained in
-["How to write Go code"](https://golang.org/doc/code.html).
-
-If you do not have a directory with Go code yet, executing the following
-instructions in your shell will create one for you and check out the restic
-repo:
-
-    $ export GOPATH="$HOME/go"
-    $ mkdir -p "$GOPATH/src/github.com/restic"
-    $ cd "$GOPATH/src/github.com/restic"
-    $ git clone https://github.com/restic/restic
-    $ cd restic
-
-You can then build restic as follows:
-
-    $ go build ./cmd/restic
-    $ ./restic version
-    restic compiled manually
-    compiled with go1.8.3 on linux/amd64
-
-The following commands can be used to run all the tests:
-
-    $ go test ./...
-
 Providing Patches
 =================
 
@@ -125,15 +92,14 @@ down to the following steps:
     GitHub. For a new feature, please add an issue before starting to work on
     it, so that duplicate work is prevented.
 
- 1. First we would kindly ask you to fork our project on GitHub if you haven't
-    done so already.
+ 1. Next, fork our project on GitHub if you haven't done so already.
 
- 2. Clone the repository locally and create a new branch. If you are working on
-    the code itself, please set up the development environment as described in
-    the previous section.
+ 2. Clone your fork of the repository locally and **create a new branch** for
+    your changes. If you are working on the code itself, please set up the
+    development environment as described in the previous section.
 
- 3. Then commit your changes as fine grained as possible, as smaller patches,
-    that handle one and only one issue are easier to discuss and merge.
+ 3. Commit your changes to the new branch as fine grained as possible, as
+    smaller patches, for individual changes, are easier to discuss and merge.
 
  4. Push the new branch with your changes to your fork of the repository.
 
@@ -146,20 +112,19 @@ down to the following steps:
     existing commit, use common sense to decide which is better), they will be
     automatically added to the pull request.
 
- 7. If your pull request changes anything that users should be aware
-    of (a bugfix, a new feature, ...) please add an entry as a new
-    file in `changelog/unreleased` including the issue number in the
-    filename (e.g. `issue-8756`). Use the template in
-    `changelog/TEMPLATE` for the content. It will be used in the
-    announcement of the next stable release. While writing, ask
-    yourself: If I were the user, what would I need to be aware of
-    with this change.
+ 7. If your pull request changes anything that users should be aware of
+    (a bugfix, a new feature, ...) please add an entry as a new file in
+    `changelog/unreleased` including the issue number in the filename (e.g.
+    `issue-8756`). Use the template in `changelog/TEMPLATE` for the content.
+    It will be used in the announcement of the next stable release. While
+    writing, ask yourself: If I were the user, what would I need to be aware
+    of with this change?
 
  8. Once your code looks good and passes all the tests, we'll merge it. Thanks
     a lot for your contribution!
 
 Please provide the patches for each bug or feature in a separate branch and
-open up a pull request for each.
+open up a pull request for each, as this simplifies discussion and merging.
 
 The restic project uses the `gofmt` tool for Go source indentation, so please
 run

VERSION

@@ -1 +1 @@
-0.9.6
+0.10.0

appveyor.yml

@@ -20,8 +20,8 @@ init:
 
 install:
   - rmdir c:\go /s /q
-  - appveyor DownloadFile https://dl.google.com/go/go1.14.windows-amd64.msi
-  - msiexec /i go1.14.windows-amd64.msi /q
+  - appveyor DownloadFile https://dl.google.com/go/go1.15.2.windows-amd64.msi
+  - msiexec /i go1.15.2.windows-amd64.msi /q
   - go version
   - go env
   - appveyor DownloadFile http://sourceforge.netcologne.de/project/gnuwin32/tar/1.13-1/tar-1.13-1-bin.zip -FileName tar.zip

build.go

@@ -3,7 +3,7 @@
 // This program aims to make building Go programs for end users easier by just
 // calling it with `go run`, without having to setup a GOPATH.
 //
-// This program needs Go >= 1.11. It'll use Go modules for compilation. It
+// This program needs Go >= 1.12. It'll use Go modules for compilation. It
 // builds the package configured as Main in the Config struct.
 
 // BSD 2-Clause License
@@ -327,8 +327,8 @@ func (v GoVersion) String() string {
 }
 
 func main() {
-	if !goVersion.AtLeast(GoVersion{1, 11, 0}) {
-		die("Go version (%v) is too old, Go <= 1.11 does not support Go Modules\n", goVersion)
+	if !goVersion.AtLeast(GoVersion{1, 12, 0}) {
+		die("Go version (%v) is too old, restic requires Go >= 1.12\n", goVersion)
 	}
 
 	if !goVersion.AtLeast(config.MinVersion) {

changelog/0.10.0_2020-09-19/issue-1680

@@ -0,0 +1,8 @@
+Enhancement: Optimize `restic mount`
+
+We've optimized the FUSE implementation used within restic.
+`restic mount` is now more responsive and uses less memory.
+
+https://github.com/restic/restic/issues/1680
+https://github.com/restic/restic/pull/2587
+https://github.com/restic/restic/pull/2787

changelog/0.10.0_2020-09-19/issue-1863

@@ -0,0 +1,6 @@
+Bugfix: Report correct number of directories processed by backup
+
+The directory statistics calculation was fixed to report the actual number
+of processed directories instead of always zero.
+
+https://github.com/restic/restic/issues/1863

changelog/0.10.0_2020-09-19/issue-2175

@@ -0,0 +1,9 @@
+Enhancement: Allow specifying user and host when creating keys
+
+When adding a new key to the repository, the username and hostname for the new
+key can be specified on the command line. This allows overriding the defaults,
+for example if you would prefer to use the FQDN to identify the host or if you
+want to add keys for several different hosts without having to run the key add
+command on those hosts.
+
+https://github.com/restic/restic/issues/2175

changelog/0.10.0_2020-09-19/issue-2254

@@ -0,0 +1,9 @@
+Bugfix: Fix tar issues when dumping `/`
+
+We've fixed an issue with dumping either `/` or files on the first sublevel
+e.g. `/foo` to tar. This also fixes tar dumping issues on Windows where this
+issue could also happen.
+
+https://github.com/restic/restic/issues/2254
+https://github.com/restic/restic/issues/2357
+https://github.com/restic/restic/pull/2255

changelog/0.10.0_2020-09-19/issue-2395

@@ -0,0 +1,12 @@
+Enhancement: Ignore sync errors when operation not supported by local filesystem
+
+The local backend has been modified to work with filesystems which doesn't support
+the `sync` operation. This operation is normally used by restic to ensure that data
+files are fully written to disk before continuing.
+
+For these limited filesystems, saving a file in the backend would previously fail with
+an "operation not supported" error. This error is now ignored, which means that e.g.
+an SMB mount on macOS can now be used as storage location for a repository.
+
+https://github.com/restic/restic/issues/2395
+https://forum.restic.net/t/sync-errors-on-mac-over-smb/1859

changelog/0.10.0_2020-09-19/issue-2427

@@ -0,0 +1,7 @@
+Enhancement: Add flag `--iexclude-file` to backup command
+
+The backup command now supports the flag `--iexclude-file` which is a
+case-insensitive version of `--exclude-file`.
+
+https://github.com/restic/restic/issues/2427
+https://github.com/restic/restic/pull/2898

changelog/0.10.0_2020-09-19/issue-2569

@@ -0,0 +1,8 @@
+Enhancement: Support excluding files by their size
+
+The `backup` command now supports the `--exclude-larger-than` option to exclude files which are
+larger than the specified maximum size. This can for example be useful to exclude unimportant
+files with a large file size.
+
+https://github.com/restic/restic/issues/2569
+https://github.com/restic/restic/pull/2914

changelog/0.10.0_2020-09-19/issue-2571

@@ -0,0 +1,16 @@
+Enhancement: Self-heal missing file parts during backup of unchanged files
+
+We've improved the resilience of restic to certain types of repository corruption.
+
+For files that are unchanged since the parent snapshot, the backup command now
+verifies that all parts of the files still exist in the repository. Parts that are
+missing, e.g. from a damaged repository, are backed up again. This verification
+was already run for files that were modified since the parent snapshot, but is
+now also done for unchanged files.
+
+Note that restic will not backup file parts that are referenced in the index but
+where the actual data is not present on disk, as this situation can only be
+detected by restic check. Please ensure that you run `restic check` regularly.
+
+https://github.com/restic/restic/issues/2571
+https://github.com/restic/restic/pull/2827

changelog/0.10.0_2020-09-19/issue-2858

@@ -0,0 +1,9 @@
+Enhancement: Support filtering snapshots by tag and path in the stats command
+
+We've added filtering snapshots by `--tag tagList` and by `--path path` to
+the `stats` command. This includes filtering of only 'latest' snapshots or
+all snapshots in a repository.
+
+https://github.com/restic/restic/issues/2858
+https://github.com/restic/restic/pull/2859
+https://forum.restic.net/t/stats-for-a-host-and-filtered-snapshots/3020

changelog/0.10.0_2020-09-19/issue-323

@@ -0,0 +1,20 @@
+Enhancement: Add command for copying snapshots between repositories
+
+We've added a copy command, allowing you to copy snapshots from one
+repository to another.
+
+Note that this process will have to read (download) and write (upload) the
+entire snapshot(s) due to the different encryption keys used on the source
+and destination repository. Also, the transferred files are not re-chunked,
+which may break deduplication between files already stored in the
+destination repo and files copied there using this command.
+
+To fully support deduplication between repositories when the copy command is
+used, the init command now supports the `--copy-chunker-params` option,
+which initializes the new repository with identical parameters for splitting
+files into chunks as an already existing repository. This allows copied
+snapshots to be equally deduplicated in both repositories.
+
+https://github.com/restic/restic/issues/323
+https://github.com/restic/restic/pull/2606
+https://github.com/restic/restic/pull/2928

changelog/0.10.0_2020-09-19/issue-551

@@ -0,0 +1,10 @@
+Enhancement: Use optimized library for hash calculation of file chunks
+
+We've switched the library used to calculate the hashes of file chunks, which
+are used for deduplication, to the optimized Minio SHA-256 implementation.
+
+Depending on the CPU it improves the hashing throughput by 10-30%. Modern x86
+CPUs with the SHA Extension should be about two to three times faster.
+
+https://github.com/restic/restic/issues/551
+https://github.com/restic/restic/pull/2709

changelog/0.10.0_2020-09-19/pull-2195

@@ -14,4 +14,10 @@ file can be written to the file before any of the preceeding file blobs.
 It is therefore possible to have gaps in the data written to the target
 files if restore fails or interrupted by the user.
 
+The implementation will try to preallocate space for the restored files
+on the filesystem to prevent file fragmentation. This ensures good read
+performance for large files, like for example VM images. If preallocating
+space is not supported by the filesystem, then this step is silently skipped.
+
 https://github.com/restic/restic/pull/2195
+https://github.com/restic/restic/pull/2893

changelog/0.10.0_2020-09-19/pull-2328

@@ -0,0 +1,8 @@
+Enhancement: Improve speed of check command
+
+We've improved the check command to traverse trees only once independent of
+whether they are contained in multiple snapshots. The check command is now much
+faster for repositories with a large number of snapshots.
+
+https://github.com/restic/restic/pull/2328
+https://github.com/restic/restic/issues/2284

changelog/0.10.0_2020-09-19/pull-2546

@@ -0,0 +1,19 @@
+Change: Return exit code 3 when failing to backup all source data
+
+The backup command used to return a zero exit code as long as a snapshot
+could be created successfully, even if some of the source files could not
+be read (in which case the snapshot would contain the rest of the files).
+
+This made it hard for automation/scripts to detect failures/incomplete
+backups by looking at the exit code. Restic now returns the following exit
+codes for the backup command:
+
+ - 0 when the command was successful
+ - 1 when there was a fatal error (no snapshot created)
+ - 3 when some source data could not be read (incomplete snapshot created)
+
+https://github.com/restic/restic/pull/2546
+https://github.com/restic/restic/issues/956
+https://github.com/restic/restic/issues/2064
+https://github.com/restic/restic/issues/2526
+https://github.com/restic/restic/issues/2364

changelog/0.10.0_2020-09-19/pull-2576

@@ -1,7 +1,9 @@
 Enhancement: Improve the chunking algorithm
 
 We've updated the chunker library responsible for splitting files into smaller
-blocks. It should improve the chunking throughput by 5-10% depending on the
+blocks. It should improve the chunking throughput by 5-15% depending on the
 CPU.
 
 https://github.com/restic/restic/pull/2576
+https://github.com/restic/restic/pull/2845
+https://github.com/restic/restic/issues/2820

changelog/0.10.0_2020-09-19/pull-2598

@@ -0,0 +1,6 @@
+Enhancement: Improve speed of diff command
+
+We've improved the performance of the diff command when comparing snapshots
+with similar content. It should run up to twice as fast as before.
+
+https://github.com/restic/restic/pull/2598

changelog/0.10.0_2020-09-19/pull-2599

@@ -0,0 +1,6 @@
+Enhancement: Slightly reduce memory usage of prune and stats commands
+
+The prune and the stats command kept directory identifiers in memory twice
+while searching for used blobs.
+
+https://github.com/restic/restic/pull/2599

changelog/0.10.0_2020-09-19/pull-2600

@@ -0,0 +1,14 @@
+Change: Update dependencies, require Go >= 1.13
+
+Restic now requires Go to be at least 1.13. This allows simplifications in the
+build process and removing workarounds.
+
+This is also probably the last version of restic still supporting mounting
+repositories via fuse on macOS. The library we're using for fuse does not
+support macOS any more and osxfuse is not open source any more.
+
+https://github.com/restic/restic/pull/2600
+https://github.com/restic/restic/pull/2852
+https://github.com/restic/restic/pull/2927
+https://github.com/bazil/fuse/issues/224
+https://github.com/osxfuse/osxfuse/issues/590

changelog/0.10.0_2020-09-19/pull-2674

@@ -0,0 +1,8 @@
+Bugfix: Add stricter prune error checks
+
+Additional checks were added to the prune command in order to improve
+resiliency to backend, hardware and/or networking issues. The checks now
+detect a few more cases where such outside factors could potentially cause
+data loss.
+
+https://github.com/restic/restic/pull/2674

changelog/0.10.0_2020-09-19/pull-2733

@@ -0,0 +1,6 @@
+Enhancement: S3 backend: Add support for WebIdentityTokenFile
+
+We've added support for EKS IAM roles for service accounts feature to the S3 backend.
+
+https://github.com/restic/restic/pull/2733
+https://github.com/restic/restic/issues/2703

changelog/0.10.0_2020-09-19/pull-2773

@@ -0,0 +1,6 @@
+Enhancement: Optimize handling of new index entries
+
+Restic now uses less memory for backups which add a lot of data, e.g. large initial backups.
+In addition, we've improved the stability in some edge cases.
+
+https://github.com/restic/restic/pull/2773

changelog/0.10.0_2020-09-19/pull-2781

@@ -0,0 +1,8 @@
+Enhancement: Reduce memory consumption of in-memory index
+
+We've improved how the index is stored in memory.
+This change can reduce memory usage for large repositories by up to 50%
+(depending on the operation).
+
+https://github.com/restic/restic/pull/2781
+https://github.com/restic/restic/pull/2812

changelog/0.10.0_2020-09-19/pull-2786

@@ -0,0 +1,6 @@
+Enhancement: Optimize `list blobs` command
+
+We've changed the implementation of `list blobs` which should be now a bit faster
+and consume almost no memory even for large repositories.
+
+https://github.com/restic/restic/pull/2786

changelog/0.10.0_2020-09-19/pull-2790

@@ -0,0 +1,6 @@
+Enhancement: Optimized file access in restic mount
+
+Reading large (> 100GiB) files from restic mountpoints is now faster,
+and the speedup is greater for larger files.
+
+https://github.com/restic/restic/pull/2790

changelog/0.10.0_2020-09-19/pull-2821

@@ -0,0 +1,8 @@
+Change: Honor the --no-lock flag in the mount command
+
+The mount command now does not lock the repository if given the
+--no-lock flag. This allows to mount repositories which are archived
+on a read only backend/filesystem.
+
+https://github.com/restic/restic/issues/1597
+https://github.com/restic/restic/pull/2821

changelog/0.10.0_2020-09-19/pull-2840

@@ -0,0 +1,7 @@
+Enhancement: Speed-up file deletion in forget, prune and rebuild-index
+
+We've sped up the file deletion for the commands forget, prune and
+rebuild-index, especially for remote repositories.
+Deletion was sequential before and is now run in parallel.
+
+https://github.com/restic/restic/pull/2840

changelog/0.10.0_2020-09-19/pull-2899

@@ -0,0 +1,9 @@
+Bugfix: Fix possible crash in the progress bar of check --read-data
+
+We've fixed a possible crash while displaying the progress bar for the
+check --read-data command. The crash occurred when the length of the
+progress bar status exceeded the terminal width, which only happened for
+very narrow terminal windows.
+
+https://github.com/restic/restic/pull/2899
+https://forum.restic.net/t/restic-rclone-pcloud-connection-issues/2963/15

changelog/0.8.0_2017-11-26/issue-1445

@@ -7,7 +7,7 @@ vulnerability, but urge all users to upgrade to the latest version of restic.
 
 Exploiting the vulnerability requires a Linux/Unix system which saves backups
 via restic and a Windows systems which restores files from the repo. In
-addition, the attackers need to be able to create create files with arbitrary
+addition, the attackers need to be able to create files with arbitrary
 names which are then saved to the restic repo. For example, by creating a file
 named "..\test.txt" (which is a perfectly legal filename on Linux) and
 restoring a snapshot containing this file on Windows, it would be written to

changelog/unreleased/pull-2600

@@ -1,6 +0,0 @@
-Change: Require Go >= 1.11
-
-Restic now requires Go to be at least 1.11. This allows simplifications in the
-build process and removing workarounds.
-
-https://github.com/restic/restic/pull/2600

cmd/restic/cleanup.go

@@ -1,7 +1,6 @@
 package main
 
 import (
-	"fmt"
 	"os"
 	"os/signal"
 	"sync"
@@ -17,8 +16,6 @@ var cleanupHandlers struct {
 	ch   chan os.Signal
 }
 
-var stderr = os.Stderr
-
 func init() {
 	cleanupHandlers.ch = make(chan os.Signal, 1)
 	go CleanupHandler(cleanupHandlers.ch)
@@ -51,7 +48,7 @@ func RunCleanupHandlers() {
 	for _, f := range cleanupHandlers.list {
 		err := f()
 		if err != nil {
-			fmt.Fprintf(stderr, "error in cleanup handler: %v\n", err)
+			Warnf("error in cleanup handler: %v\n", err)
 		}
 	}
 	cleanupHandlers.list = nil
@@ -61,7 +58,7 @@ func RunCleanupHandlers() {
 func CleanupHandler(c <-chan os.Signal) {
 	for s := range c {
 		debug.Log("signal %v received, cleaning up", s)
-		fmt.Fprintf(stderr, "%ssignal %v received, cleaning up\n", ClearLine(), s)
+		Warnf("%ssignal %v received, cleaning up\n", ClearLine(), s)
 
 		code := 0
 

cmd/restic/cmd_backup.go

@@ -39,10 +39,9 @@ given as the arguments.
 EXIT STATUS
 ===========
 
-Exit status is 0 if the command was successful, and non-zero if there was any error.
-
-Note that some issues such as unreadable or deleted files during backup
-currently doesn't result in a non-zero error exit status.
+Exit status is 0 if the command was successful.
+Exit status is 1 if there was a fatal error (no snapshot created).
+Exit status is 3 if some source data could not be read (incomplete snapshot created).
 `,
 	PreRun: func(cmd *cobra.Command, args []string) {
 		if backupOptions.Host == "" {
@@ -79,26 +78,31 @@ currently doesn't result in a non-zero error exit status.
 
 // BackupOptions bundles all options for the backup command.
 type BackupOptions struct {
-	Parent              string
-	Force               bool
-	Excludes            []string
-	InsensitiveExcludes []string
-	ExcludeFiles        []string
-	ExcludeOtherFS      bool
-	ExcludeIfPresent    []string
-	ExcludeCaches       bool
-	Stdin               bool
-	StdinFilename       string
-	Tags                []string
-	Host                string
-	FilesFrom           []string
-	TimeStamp           string
-	WithAtime           bool
-	IgnoreInode         bool
+	Parent                  string
+	Force                   bool
+	Excludes                []string
+	InsensitiveExcludes     []string
+	ExcludeFiles            []string
+	InsensitiveExcludeFiles []string
+	ExcludeOtherFS          bool
+	ExcludeIfPresent        []string
+	ExcludeCaches           bool
+	ExcludeLargerThan       string
+	Stdin                   bool
+	StdinFilename           string
+	Tags                    []string
+	Host                    string
+	FilesFrom               []string
+	TimeStamp               string
+	WithAtime               bool
+	IgnoreInode             bool
 }
 
 var backupOptions BackupOptions
 
+// ErrInvalidSourceData is used to report an incomplete backup
+var ErrInvalidSourceData = errors.New("failed to read all source data during backup")
+
 func init() {
 	cmdRoot.AddCommand(cmdBackup)
 
@@ -108,9 +112,11 @@ func init() {
 	f.StringArrayVarP(&backupOptions.Excludes, "exclude", "e", nil, "exclude a `pattern` (can be specified multiple times)")
 	f.StringArrayVar(&backupOptions.InsensitiveExcludes, "iexclude", nil, "same as --exclude `pattern` but ignores the casing of filenames")
 	f.StringArrayVar(&backupOptions.ExcludeFiles, "exclude-file", nil, "read exclude patterns from a `file` (can be specified multiple times)")
+	f.StringArrayVar(&backupOptions.InsensitiveExcludeFiles, "iexclude-file", nil, "same as --exclude-file but ignores casing of `file`names in patterns")
 	f.BoolVarP(&backupOptions.ExcludeOtherFS, "one-file-system", "x", false, "exclude other file systems")
 	f.StringArrayVar(&backupOptions.ExcludeIfPresent, "exclude-if-present", nil, "takes `filename[:header]`, exclude contents of directories containing filename (except filename itself) if header of that file is as provided (can be specified multiple times)")
-	f.BoolVar(&backupOptions.ExcludeCaches, "exclude-caches", false, `excludes cache directories that are marked with a CACHEDIR.TAG file. See http://bford.info/cachedir/spec.html for the Cache Directory Tagging Standard`)
+	f.BoolVar(&backupOptions.ExcludeCaches, "exclude-caches", false, `excludes cache directories that are marked with a CACHEDIR.TAG file. See https://bford.info/cachedir/ for the Cache Directory Tagging Standard`)
+	f.StringVar(&backupOptions.ExcludeLargerThan, "exclude-larger-than", "", "max `size` of the files to be backed up (allowed suffixes: k/K, m/M, g/G, t/T)")
 	f.BoolVar(&backupOptions.Stdin, "stdin", false, "read backup from stdin")
 	f.StringVar(&backupOptions.StdinFilename, "stdin-filename", "stdin", "`filename` to use when reading from stdin")
 	f.StringArrayVar(&backupOptions.Tags, "tag", nil, "add a `tag` for the new snapshot (can be specified multiple times)")
@@ -237,6 +243,14 @@ func collectRejectByNameFuncs(opts BackupOptions, repo *repository.Repository, t
 		opts.Excludes = append(opts.Excludes, excludes...)
 	}
 
+	if len(opts.InsensitiveExcludeFiles) > 0 {
+		excludes, err := readExcludePatternsFromFiles(opts.InsensitiveExcludeFiles)
+		if err != nil {
+			return nil, err
+		}
+		opts.InsensitiveExcludes = append(opts.InsensitiveExcludes, excludes...)
+	}
+
 	if len(opts.InsensitiveExcludes) > 0 {
 		fs = append(fs, rejectByInsensitivePattern(opts.InsensitiveExcludes))
 	}
@@ -273,6 +287,14 @@ func collectRejectFuncs(opts BackupOptions, repo *repository.Repository, targets
 		fs = append(fs, f)
 	}
 
+	if len(opts.ExcludeLargerThan) != 0 && !opts.Stdin {
+		f, err := rejectBySize(opts.ExcludeLargerThan)
+		if err != nil {
+			return nil, err
+		}
+		fs = append(fs, f)
+	}
+
 	return fs, nil
 }
 
@@ -415,7 +437,7 @@ func runBackup(opts BackupOptions, gopts GlobalOptions, term *termstatus.Termina
 	var t tomb.Tomb
 
 	if gopts.verbosity >= 2 && !gopts.JSON {
-		term.Print("open repository\n")
+		Verbosef("open repository\n")
 	}
 
 	repo, err := OpenRepository(gopts)
@@ -557,7 +579,11 @@ func runBackup(opts BackupOptions, gopts GlobalOptions, term *termstatus.Termina
 	arch.SelectByName = selectByNameFilter
 	arch.Select = selectFilter
 	arch.WithAtime = opts.WithAtime
-	arch.Error = p.Error
+	success := true
+	arch.Error = func(item string, fi os.FileInfo, err error) error {
+		success = false
+		return p.Error(item, fi, err)
+	}
 	arch.CompleteItem = p.CompleteItem
 	arch.StartFile = p.StartFile
 	arch.CompleteBlob = p.CompleteBlob
@@ -575,24 +601,6 @@ func runBackup(opts BackupOptions, gopts GlobalOptions, term *termstatus.Termina
 		ParentSnapshot: *parentSnapshotID,
 	}
 
-	uploader := archiver.IndexUploader{
-		Repository: repo,
-		Start: func() {
-			if !gopts.JSON {
-				p.VV("uploading intermediate index")
-			}
-		},
-		Complete: func(id restic.ID) {
-			if !gopts.JSON {
-				p.V("uploaded intermediate index %v", id.Str())
-			}
-		},
-	}
-
-	t.Go(func() error {
-		return uploader.Upload(gopts.ctx, t.Context(gopts.ctx), 30*time.Second)
-	})
-
 	if !gopts.JSON {
 		p.V("start backup on %v", targets)
 	}
@@ -612,6 +620,9 @@ func runBackup(opts BackupOptions, gopts GlobalOptions, term *termstatus.Termina
 	if !gopts.JSON {
 		p.P("snapshot %s saved\n", id.Str())
 	}
+	if !success {
+		return ErrInvalidSourceData
+	}
 
 	// Return error if any
 	return err

cmd/restic/cmd_cat.go

@@ -2,8 +2,6 @@ package main
 
 import (
 	"encoding/json"
-	"fmt"
-	"os"
 
 	"github.com/spf13/cobra"
 
@@ -76,7 +74,7 @@ func runCat(gopts GlobalOptions, args []string) error {
 			return err
 		}
 
-		fmt.Println(string(buf))
+		Println(string(buf))
 		return nil
 	case "index":
 		buf, err := repo.LoadAndDecrypt(gopts.ctx, nil, restic.IndexFile, id)
@@ -84,9 +82,8 @@ func runCat(gopts GlobalOptions, args []string) error {
 			return err
 		}
 
-		_, err = os.Stdout.Write(append(buf, '\n'))
-		return err
-
+		Println(string(buf))
+		return nil
 	case "snapshot":
 		sn := &restic.Snapshot{}
 		err = repo.LoadJSONUnpacked(gopts.ctx, restic.SnapshotFile, id, sn)
@@ -99,8 +96,7 @@ func runCat(gopts GlobalOptions, args []string) error {
 			return err
 		}
 
-		fmt.Println(string(buf))
-
+		Println(string(buf))
 		return nil
 	case "key":
 		h := restic.Handle{Type: restic.KeyFile, Name: id.String()}
@@ -120,7 +116,7 @@ func runCat(gopts GlobalOptions, args []string) error {
 			return err
 		}
 
-		fmt.Println(string(buf))
+		Println(string(buf))
 		return nil
 	case "masterkey":
 		buf, err := json.MarshalIndent(repo.Key(), "", "  ")
@@ -128,7 +124,7 @@ func runCat(gopts GlobalOptions, args []string) error {
 			return err
 		}
 
-		fmt.Println(string(buf))
+		Println(string(buf))
 		return nil
 	case "lock":
 		lock, err := restic.LoadLock(gopts.ctx, repo, id)
@@ -141,8 +137,7 @@ func runCat(gopts GlobalOptions, args []string) error {
 			return err
 		}
 
-		fmt.Println(string(buf))
-
+		Println(string(buf))
 		return nil
 	}
 
@@ -154,7 +149,7 @@ func runCat(gopts GlobalOptions, args []string) error {
 
 	switch tpe {
 	case "pack":
-		h := restic.Handle{Type: restic.DataFile, Name: id.String()}
+		h := restic.Handle{Type: restic.PackFile, Name: id.String()}
 		buf, err := backend.LoadAll(gopts.ctx, nil, repo.Backend(), h)
 		if err != nil {
 			return err
@@ -162,16 +157,15 @@ func runCat(gopts GlobalOptions, args []string) error {
 
 		hash := restic.Hash(buf)
 		if !hash.Equal(id) {
-			fmt.Fprintf(stderr, "Warning: hash of data does not match ID, want\n  %v\ngot:\n  %v\n", id.String(), hash.String())
+			Warnf("Warning: hash of data does not match ID, want\n  %v\ngot:\n  %v\n", id.String(), hash.String())
 		}
 
-		_, err = os.Stdout.Write(buf)
+		_, err = globalOptions.stdout.Write(buf)
 		return err
 
 	case "blob":
 		for _, t := range []restic.BlobType{restic.DataBlob, restic.TreeBlob} {
-			_, found := repo.Index().Lookup(id, t)
-			if !found {
+			if !repo.Index().Has(id, t) {
 				continue
 			}
 
@@ -180,7 +174,7 @@ func runCat(gopts GlobalOptions, args []string) error {
 				return err
 			}
 
-			_, err = os.Stdout.Write(buf)
+			_, err = globalOptions.stdout.Write(buf)
 			return err
 		}
 

cmd/restic/cmd_check.go

@@ -3,10 +3,8 @@ package main
 import (
 	"fmt"
 	"io/ioutil"
-	"os"
 	"strconv"
 	"strings"
-	"time"
 
 	"github.com/spf13/cobra"
 
@@ -100,36 +98,6 @@ func stringToIntSlice(param string) (split []uint, err error) {
 	return result, nil
 }
 
-func newReadProgress(gopts GlobalOptions, todo restic.Stat) *restic.Progress {
-	if gopts.Quiet {
-		return nil
-	}
-
-	readProgress := restic.NewProgress()
-
-	readProgress.OnUpdate = func(s restic.Stat, d time.Duration, ticker bool) {
-		status := fmt.Sprintf("[%s] %s  %d / %d items",
-			formatDuration(d),
-			formatPercent(s.Blobs, todo.Blobs),
-			s.Blobs, todo.Blobs)
-
-		if w := stdoutTerminalWidth(); w > 0 {
-			if len(status) > w {
-				max := w - len(status) - 4
-				status = status[:max] + "... "
-			}
-		}
-
-		PrintProgress("%s", status)
-	}
-
-	readProgress.OnDone = func(s restic.Stat, d time.Duration, ticker bool) {
-		fmt.Printf("\nduration: %s\n", formatDuration(d))
-	}
-
-	return readProgress
-}
-
 // prepareCheckCache configures a special cache directory for check.
 //
 //  * if --with-cache is specified, the default cache is used
@@ -235,7 +203,7 @@ func runCheck(opts CheckOptions, gopts GlobalOptions, args []string) error {
 			continue
 		}
 		errorsFound = true
-		fmt.Fprintf(os.Stderr, "%v\n", err)
+		Warnf("%v\n", err)
 	}
 
 	if orphanedPacks > 0 {
@@ -249,18 +217,18 @@ func runCheck(opts CheckOptions, gopts GlobalOptions, args []string) error {
 	for err := range errChan {
 		errorsFound = true
 		if e, ok := err.(checker.TreeError); ok {
-			fmt.Fprintf(os.Stderr, "error for tree %v:\n", e.ID.Str())
+			Warnf("error for tree %v:\n", e.ID.Str())
 			for _, treeErr := range e.Errors {
-				fmt.Fprintf(os.Stderr, "  %v\n", treeErr)
+				Warnf("  %v\n", treeErr)
 			}
 		} else {
-			fmt.Fprintf(os.Stderr, "error: %v\n", err)
+			Warnf("error: %v\n", err)
 		}
 	}
 
 	if opts.CheckUnused {
 		for _, id := range chkr.UnusedBlobs() {
-			Verbosef("unused blob %v\n", id.Str())
+			Verbosef("unused blob %v\n", id)
 			errorsFound = true
 		}
 	}
@@ -282,14 +250,14 @@ func runCheck(opts CheckOptions, gopts GlobalOptions, args []string) error {
 			Verbosef("read all data\n")
 		}
 
-		p := newReadProgress(gopts, restic.Stat{Blobs: packCount})
+		p := newProgressMax(!gopts.Quiet, packCount, "packs")
 		errChan := make(chan error)
 
 		go chkr.ReadPacks(gopts.ctx, packs, p, errChan)
 
 		for err := range errChan {
 			errorsFound = true
-			fmt.Fprintf(os.Stderr, "%v\n", err)
+			Warnf("%v\n", err)
 		}
 	}
 

cmd/restic/cmd_copy.go

@@ -0,0 +1,233 @@
+package main
+
+import (
+	"context"
+	"fmt"
+
+	"github.com/restic/restic/internal/debug"
+	"github.com/restic/restic/internal/restic"
+
+	"github.com/spf13/cobra"
+)
+
+var cmdCopy = &cobra.Command{
+	Use:   "copy [flags] [snapshotID ...]",
+	Short: "Copy snapshots from one repository to another",
+	Long: `
+The "copy" command copies one or more snapshots from one repository to another
+repository. Note that this will have to read (download) and write (upload) the
+entire snapshot(s) due to the different encryption keys on the source and
+destination, and that transferred files are not re-chunked, which may break
+their deduplication. This can be mitigated by the "--copy-chunker-params"
+option when initializing a new destination repository using the "init" command.
+`,
+	RunE: func(cmd *cobra.Command, args []string) error {
+		return runCopy(copyOptions, globalOptions, args)
+	},
+}
+
+// CopyOptions bundles all options for the copy command.
+type CopyOptions struct {
+	secondaryRepoOptions
+	Hosts []string
+	Tags  restic.TagLists
+	Paths []string
+}
+
+var copyOptions CopyOptions
+
+func init() {
+	cmdRoot.AddCommand(cmdCopy)
+
+	f := cmdCopy.Flags()
+	initSecondaryRepoOptions(f, &copyOptions.secondaryRepoOptions, "destination", "to copy snapshots to")
+	f.StringArrayVarP(&copyOptions.Hosts, "host", "H", nil, "only consider snapshots for this `host`, when no snapshot ID is given (can be specified multiple times)")
+	f.Var(&copyOptions.Tags, "tag", "only consider snapshots which include this `taglist`, when no snapshot ID is given")
+	f.StringArrayVar(&copyOptions.Paths, "path", nil, "only consider snapshots which include this (absolute) `path`, when no snapshot ID is given")
+}
+
+func runCopy(opts CopyOptions, gopts GlobalOptions, args []string) error {
+	dstGopts, err := fillSecondaryGlobalOpts(opts.secondaryRepoOptions, gopts, "destination")
+	if err != nil {
+		return err
+	}
+
+	ctx, cancel := context.WithCancel(gopts.ctx)
+	defer cancel()
+
+	srcRepo, err := OpenRepository(gopts)
+	if err != nil {
+		return err
+	}
+
+	dstRepo, err := OpenRepository(dstGopts)
+	if err != nil {
+		return err
+	}
+
+	srcLock, err := lockRepo(srcRepo)
+	defer unlockRepo(srcLock)
+	if err != nil {
+		return err
+	}
+
+	dstLock, err := lockRepo(dstRepo)
+	defer unlockRepo(dstLock)
+	if err != nil {
+		return err
+	}
+
+	debug.Log("Loading source index")
+	if err := srcRepo.LoadIndex(ctx); err != nil {
+		return err
+	}
+
+	debug.Log("Loading destination index")
+	if err := dstRepo.LoadIndex(ctx); err != nil {
+		return err
+	}
+
+	dstSnapshotByOriginal := make(map[restic.ID][]*restic.Snapshot)
+	for sn := range FindFilteredSnapshots(ctx, dstRepo, opts.Hosts, opts.Tags, opts.Paths, nil) {
+		if sn.Original != nil && !sn.Original.IsNull() {
+			dstSnapshotByOriginal[*sn.Original] = append(dstSnapshotByOriginal[*sn.Original], sn)
+		}
+		// also consider identical snapshot copies
+		dstSnapshotByOriginal[*sn.ID()] = append(dstSnapshotByOriginal[*sn.ID()], sn)
+	}
+
+	cloner := &treeCloner{
+		srcRepo:      srcRepo,
+		dstRepo:      dstRepo,
+		visitedTrees: restic.NewIDSet(),
+		buf:          nil,
+	}
+
+	for sn := range FindFilteredSnapshots(ctx, srcRepo, opts.Hosts, opts.Tags, opts.Paths, args) {
+		Verbosef("\nsnapshot %s of %v at %s)\n", sn.ID().Str(), sn.Paths, sn.Time)
+
+		// check whether the destination has a snapshot with the same persistent ID which has similar snapshot fields
+		srcOriginal := *sn.ID()
+		if sn.Original != nil {
+			srcOriginal = *sn.Original
+		}
+		if originalSns, ok := dstSnapshotByOriginal[srcOriginal]; ok {
+			isCopy := false
+			for _, originalSn := range originalSns {
+				if similarSnapshots(originalSn, sn) {
+					Verbosef("skipping source snapshot %s, was already copied to snapshot %s\n", sn.ID().Str(), originalSn.ID().Str())
+					isCopy = true
+					break
+				}
+			}
+			if isCopy {
+				continue
+			}
+		}
+		Verbosef("  copy started, this may take a while...\n")
+
+		if err := cloner.copyTree(ctx, *sn.Tree); err != nil {
+			return err
+		}
+		debug.Log("tree copied")
+
+		if err = dstRepo.Flush(ctx); err != nil {
+			return err
+		}
+		debug.Log("flushed packs and saved index")
+
+		// save snapshot
+		sn.Parent = nil // Parent does not have relevance in the new repo.
+		// Use Original as a persistent snapshot ID
+		if sn.Original == nil {
+			sn.Original = sn.ID()
+		}
+		newID, err := dstRepo.SaveJSONUnpacked(ctx, restic.SnapshotFile, sn)
+		if err != nil {
+			return err
+		}
+		Verbosef("snapshot %s saved\n", newID.Str())
+	}
+	return nil
+}
+
+func similarSnapshots(sna *restic.Snapshot, snb *restic.Snapshot) bool {
+	// everything except Parent and Original must match
+	if !sna.Time.Equal(snb.Time) || !sna.Tree.Equal(*snb.Tree) || sna.Hostname != snb.Hostname ||
+		sna.Username != snb.Username || sna.UID != snb.UID || sna.GID != snb.GID ||
+		len(sna.Paths) != len(snb.Paths) || len(sna.Excludes) != len(snb.Excludes) ||
+		len(sna.Tags) != len(snb.Tags) {
+		return false
+	}
+	if !sna.HasPaths(snb.Paths) || !sna.HasTags(snb.Tags) {
+		return false
+	}
+	for i, a := range sna.Excludes {
+		if a != snb.Excludes[i] {
+			return false
+		}
+	}
+	return true
+}
+
+type treeCloner struct {
+	srcRepo      restic.Repository
+	dstRepo      restic.Repository
+	visitedTrees restic.IDSet
+	buf          []byte
+}
+
+func (t *treeCloner) copyTree(ctx context.Context, treeID restic.ID) error {
+	// We have already processed this tree
+	if t.visitedTrees.Has(treeID) {
+		return nil
+	}
+
+	tree, err := t.srcRepo.LoadTree(ctx, treeID)
+	if err != nil {
+		return fmt.Errorf("LoadTree(%v) returned error %v", treeID.Str(), err)
+	}
+	t.visitedTrees.Insert(treeID)
+
+	// Do we already have this tree blob?
+	if !t.dstRepo.Index().Has(treeID, restic.TreeBlob) {
+		newTreeID, err := t.dstRepo.SaveTree(ctx, tree)
+		if err != nil {
+			return fmt.Errorf("SaveTree(%v) returned error %v", treeID.Str(), err)
+		}
+		// Assurance only.
+		if newTreeID != treeID {
+			return fmt.Errorf("SaveTree(%v) returned unexpected id %s", treeID.Str(), newTreeID.Str())
+		}
+	}
+
+	// TODO: parellize this stuff, likely only needed inside a tree.
+
+	for _, entry := range tree.Nodes {
+		// If it is a directory, recurse
+		if entry.Type == "dir" && entry.Subtree != nil {
+			if err := t.copyTree(ctx, *entry.Subtree); err != nil {
+				return err
+			}
+		}
+		// Copy the blobs for this file.
+		for _, blobID := range entry.Content {
+			// Do we already have this data blob?
+			if t.dstRepo.Index().Has(blobID, restic.DataBlob) {
+				continue
+			}
+			debug.Log("Copying blob %s\n", blobID.Str())
+			t.buf, err = t.srcRepo.LoadBlob(ctx, restic.DataBlob, blobID, t.buf)
+			if err != nil {
+				return fmt.Errorf("LoadBlob(%v) returned error %v", blobID, err)
+			}
+
+			_, _, err = t.dstRepo.SaveBlob(ctx, restic.DataBlob, t.buf, blobID, false)
+			if err != nil {
+				return fmt.Errorf("SaveBlob(%v) returned error %v", blobID, err)
+			}
+		}
+	}
+
+	return nil
+}

cmd/restic/cmd_debug.go

@@ -84,12 +84,12 @@ type Blob struct {
 
 func printPacks(repo *repository.Repository, wr io.Writer) error {
 
-	return repo.List(context.TODO(), restic.DataFile, func(id restic.ID, size int64) error {
-		h := restic.Handle{Type: restic.DataFile, Name: id.String()}
+	return repo.List(context.TODO(), restic.PackFile, func(id restic.ID, size int64) error {
+		h := restic.Handle{Type: restic.PackFile, Name: id.String()}
 
 		blobs, err := pack.List(repo.Key(), restic.ReaderAt(repo.Backend(), h), size)
 		if err != nil {
-			fmt.Fprintf(globalOptions.stderr, "error for pack %v: %v\n", id.Str(), err)
+			Warnf("error for pack %v: %v\n", id.Str(), err)
 			return nil
 		}
 
@@ -112,7 +112,7 @@ func printPacks(repo *repository.Repository, wr io.Writer) error {
 
 func dumpIndexes(repo restic.Repository, wr io.Writer) error {
 	return repo.List(context.TODO(), restic.IndexFile, func(id restic.ID, size int64) error {
-		fmt.Printf("index_id: %v\n", id)
+		Printf("index_id: %v\n", id)
 
 		idx, err := repository.LoadIndex(context.TODO(), repo, id)
 		if err != nil {
@@ -151,13 +151,13 @@ func runDebugDump(gopts GlobalOptions, args []string) error {
 	case "packs":
 		return printPacks(repo, gopts.stdout)
 	case "all":
-		fmt.Printf("snapshots:\n")
+		Printf("snapshots:\n")
 		err := debugPrintSnapshots(repo, gopts.stdout)
 		if err != nil {
 			return err
 		}
 
-		fmt.Printf("\nindexes:\n")
+		Printf("\nindexes:\n")
 		err = dumpIndexes(repo, gopts.stdout)
 		if err != nil {
 			return err

cmd/restic/cmd_diff.go

@@ -14,7 +14,7 @@ import (
 )
 
 var cmdDiff = &cobra.Command{
-	Use:   "diff snapshot-ID snapshot-ID",
+	Use:   "diff [flags] snapshot-ID snapshot-ID",
 	Short: "Show differences between two snapshots",
 	Long: `
 The "diff" command shows differences from the first to the second snapshot. The
@@ -116,10 +116,10 @@ func addBlobs(bs restic.BlobSet, node *restic.Node) {
 
 // DiffStats collects the differences between two snapshots.
 type DiffStats struct {
-	ChangedFiles            int
-	Added                   DiffStat
-	Removed                 DiffStat
-	BlobsBefore, BlobsAfter restic.BlobSet
+	ChangedFiles                         int
+	Added                                DiffStat
+	Removed                              DiffStat
+	BlobsBefore, BlobsAfter, BlobsCommon restic.BlobSet
 }
 
 // NewDiffStats creates new stats for a diff run.
@@ -127,6 +127,7 @@ func NewDiffStats() *DiffStats {
 	return &DiffStats{
 		BlobsBefore: restic.NewBlobSet(),
 		BlobsAfter:  restic.NewBlobSet(),
+		BlobsCommon: restic.NewBlobSet(),
 	}
 }
 
@@ -177,6 +178,27 @@ func (c *Comparer) printDir(ctx context.Context, mode string, stats *DiffStat, b
 	return nil
 }
 
+func (c *Comparer) collectDir(ctx context.Context, blobs restic.BlobSet, id restic.ID) error {
+	debug.Log("print tree %v", id)
+	tree, err := c.repo.LoadTree(ctx, id)
+	if err != nil {
+		return err
+	}
+
+	for _, node := range tree.Nodes {
+		addBlobs(blobs, node)
+
+		if node.Type == "dir" {
+			err := c.collectDir(ctx, blobs, *node.Subtree)
+			if err != nil {
+				Warnf("error: %v\n", err)
+			}
+		}
+	}
+
+	return nil
+}
+
 func uniqueNodeNames(tree1, tree2 *restic.Tree) (tree1Nodes, tree2Nodes map[string]*restic.Node, uniqueNames []string) {
 	names := make(map[string]struct{})
 	tree1Nodes = make(map[string]*restic.Node)
@@ -196,7 +218,7 @@ func uniqueNodeNames(tree1, tree2 *restic.Tree) (tree1Nodes, tree2Nodes map[stri
 		uniqueNames = append(uniqueNames, name)
 	}
 
-	sort.Sort(sort.StringSlice(uniqueNames))
+	sort.Strings(uniqueNames)
 	return tree1Nodes, tree2Nodes, uniqueNames
 }
 
@@ -248,7 +270,12 @@ func (c *Comparer) diffTree(ctx context.Context, stats *DiffStats, prefix string
 			}
 
 			if node1.Type == "dir" && node2.Type == "dir" {
-				err := c.diffTree(ctx, stats, name, *node1.Subtree, *node2.Subtree)
+				var err error
+				if (*node1.Subtree).Equal(*node2.Subtree) {
+					err = c.collectDir(ctx, stats.BlobsCommon, *node1.Subtree)
+				} else {
+					err = c.diffTree(ctx, stats, name, *node1.Subtree, *node2.Subtree)
+				}
 				if err != nil {
 					Warnf("error: %v\n", err)
 				}
@@ -345,8 +372,8 @@ func runDiff(opts DiffOptions, gopts GlobalOptions, args []string) error {
 	}
 
 	both := stats.BlobsBefore.Intersect(stats.BlobsAfter)
-	updateBlobs(repo, stats.BlobsBefore.Sub(both), &stats.Removed)
-	updateBlobs(repo, stats.BlobsAfter.Sub(both), &stats.Added)
+	updateBlobs(repo, stats.BlobsBefore.Sub(both).Sub(stats.BlobsCommon), &stats.Removed)
+	updateBlobs(repo, stats.BlobsAfter.Sub(both).Sub(stats.BlobsCommon), &stats.Added)
 
 	Printf("\n")
 	Printf("Files:       %5d new, %5d removed, %5d changed\n", stats.Added.Files, stats.Removed.Files, stats.ChangedFiles)

cmd/restic/cmd_dump.go

@@ -1,19 +1,16 @@
 package main
 
 import (
-	"archive/tar"
 	"context"
 	"fmt"
-	"io"
 	"os"
 	"path"
 	"path/filepath"
-	"strings"
 
 	"github.com/restic/restic/internal/debug"
+	"github.com/restic/restic/internal/dump"
 	"github.com/restic/restic/internal/errors"
 	"github.com/restic/restic/internal/restic"
-	"github.com/restic/restic/internal/walker"
 
 	"github.com/spf13/cobra"
 )
@@ -22,8 +19,10 @@ var cmdDump = &cobra.Command{
 	Use:   "dump [flags] snapshotID file",
 	Short: "Print a backed-up file to stdout",
 	Long: `
-The "dump" command extracts a single file from a snapshot from the repository and
-prints its contents to stdout.
+The "dump" command extracts files from a snapshot from the repository. If a
+single file is selected, it prints its contents to stdout. Folders are output
+as a tar file containing the contents of the specified folder.  Pass "/" as
+file name to dump the whole snapshot as a tar file.
 
 The special snapshot "latest" can be used to use the latest snapshot in the
 repository.
@@ -59,17 +58,14 @@ func init() {
 
 func splitPath(p string) []string {
 	d, f := path.Split(p)
-	if d == "" {
+	if d == "" || d == "/" {
 		return []string{f}
 	}
-	if d == "/" {
-		return []string{d}
-	}
-	s := splitPath(path.Clean(d))
+	s := splitPath(path.Join("/", d))
 	return append(s, f)
 }
 
-func printFromTree(ctx context.Context, tree *restic.Tree, repo restic.Repository, prefix string, pathComponents []string, pathToPrint string) error {
+func printFromTree(ctx context.Context, tree *restic.Tree, repo restic.Repository, prefix string, pathComponents []string) error {
 
 	if tree == nil {
 		return fmt.Errorf("called with a nil tree")
@@ -81,24 +77,42 @@ func printFromTree(ctx context.Context, tree *restic.Tree, repo restic.Repositor
 	if l == 0 {
 		return fmt.Errorf("empty path components")
 	}
+
+	// If we print / we need to assume that there are multiple nodes at that
+	// level in the tree.
+	if pathComponents[0] == "" {
+		if err := checkStdoutTar(); err != nil {
+			return err
+		}
+		return dump.WriteTar(ctx, repo, tree, "/", os.Stdout)
+	}
+
 	item := filepath.Join(prefix, pathComponents[0])
 	for _, node := range tree.Nodes {
-		if node.Name == pathComponents[0] || pathComponents[0] == "/" {
+		// If dumping something in the highest level it will just take the
+		// first item it finds and dump that according to the switch case below.
+		if node.Name == pathComponents[0] {
 			switch {
-			case l == 1 && node.Type == "file":
-				return getNodeData(ctx, os.Stdout, repo, node)
-			case l > 1 && node.Type == "dir":
+			case l == 1 && dump.IsFile(node):
+				return dump.GetNodeData(ctx, os.Stdout, repo, node)
+			case l > 1 && dump.IsDir(node):
 				subtree, err := repo.LoadTree(ctx, *node.Subtree)
 				if err != nil {
 					return errors.Wrapf(err, "cannot load subtree for %q", item)
 				}
-				return printFromTree(ctx, subtree, repo, item, pathComponents[1:], pathToPrint)
-			case node.Type == "dir":
-				node.Path = pathToPrint
-				return tarTree(ctx, repo, node, pathToPrint)
+				return printFromTree(ctx, subtree, repo, item, pathComponents[1:])
+			case dump.IsDir(node):
+				if err := checkStdoutTar(); err != nil {
+					return err
+				}
+				subtree, err := repo.LoadTree(ctx, *node.Subtree)
+				if err != nil {
+					return err
+				}
+				return dump.WriteTar(ctx, repo, subtree, item, os.Stdout)
 			case l > 1:
 				return fmt.Errorf("%q should be a dir, but is a %q", item, node.Type)
-			case node.Type != "file":
+			case !dump.IsFile(node):
 				return fmt.Errorf("%q should be a file, but is a %q", item, node.Type)
 			}
 		}
@@ -162,7 +176,7 @@ func runDump(opts DumpOptions, gopts GlobalOptions, args []string) error {
 		Exitf(2, "loading tree for snapshot %q failed: %v", snapshotIDString, err)
 	}
 
-	err = printFromTree(ctx, tree, repo, "", splittedPath, pathToPrint)
+	err = printFromTree(ctx, tree, repo, "/", splittedPath)
 	if err != nil {
 		Exitf(2, "cannot dump file: %v", err)
 	}
@@ -170,126 +184,9 @@ func runDump(opts DumpOptions, gopts GlobalOptions, args []string) error {
 	return nil
 }
 
-func getNodeData(ctx context.Context, output io.Writer, repo restic.Repository, node *restic.Node) error {
-	var (
-		buf []byte
-		err error
-	)
-	for _, id := range node.Content {
-		buf, err = repo.LoadBlob(ctx, restic.DataBlob, id, buf)
-		if err != nil {
-			return err
-		}
-
-		_, err = output.Write(buf)
-		if err != nil {
-			return errors.Wrap(err, "Write")
-		}
-
-	}
-	return nil
-}
-
-func tarTree(ctx context.Context, repo restic.Repository, rootNode *restic.Node, rootPath string) error {
-
+func checkStdoutTar() error {
 	if stdoutIsTerminal() {
 		return fmt.Errorf("stdout is the terminal, please redirect output")
 	}
-
-	tw := tar.NewWriter(os.Stdout)
-	defer tw.Close()
-
-	// If we want to dump "/" we'll need to add the name of the first node, too
-	// as it would get lost otherwise.
-	if rootNode.Path == "/" {
-		rootNode.Path = path.Join(rootNode.Path, rootNode.Name)
-		rootPath = rootNode.Path
-	}
-
-	// we know that rootNode is a folder and walker.Walk will already process
-	// the next node, so we have to tar this one first, too
-	if err := tarNode(ctx, tw, rootNode, repo); err != nil {
-		return err
-	}
-
-	err := walker.Walk(ctx, repo, *rootNode.Subtree, nil, func(_ restic.ID, nodepath string, node *restic.Node, err error) (bool, error) {
-		if err != nil {
-			return false, err
-		}
-		if node == nil {
-			return false, nil
-		}
-
-		node.Path = path.Join(rootPath, nodepath)
-
-		if node.Type == "file" || node.Type == "symlink" || node.Type == "dir" {
-			err := tarNode(ctx, tw, node, repo)
-			if err != err {
-				return false, err
-			}
-		}
-
-		return false, nil
-	})
-
-	return err
-}
-
-func tarNode(ctx context.Context, tw *tar.Writer, node *restic.Node, repo restic.Repository) error {
-
-	header := &tar.Header{
-		Name:       node.Path,
-		Size:       int64(node.Size),
-		Mode:       int64(node.Mode),
-		Uid:        int(node.UID),
-		Gid:        int(node.GID),
-		ModTime:    node.ModTime,
-		AccessTime: node.AccessTime,
-		ChangeTime: node.ChangeTime,
-		PAXRecords: parseXattrs(node.ExtendedAttributes),
-	}
-
-	if node.Type == "symlink" {
-		header.Typeflag = tar.TypeSymlink
-		header.Linkname = node.LinkTarget
-	}
-
-	if node.Type == "dir" {
-		header.Typeflag = tar.TypeDir
-	}
-
-	err := tw.WriteHeader(header)
-
-	if err != nil {
-		return errors.Wrap(err, "TarHeader ")
-	}
-
-	return getNodeData(ctx, tw, repo, node)
-
-}
-
-func parseXattrs(xattrs []restic.ExtendedAttribute) map[string]string {
-	tmpMap := make(map[string]string)
-
-	for _, attr := range xattrs {
-		attrString := string(attr.Value)
-
-		if strings.HasPrefix(attr.Name, "system.posix_acl_") {
-			na := acl{}
-			na.decode(attr.Value)
-
-			if na.String() != "" {
-				if strings.Contains(attr.Name, "system.posix_acl_access") {
-					tmpMap["SCHILY.acl.access"] = na.String()
-				} else if strings.Contains(attr.Name, "system.posix_acl_default") {
-					tmpMap["SCHILY.acl.default"] = na.String()
-				}
-			}
-
-		} else {
-			tmpMap["SCHILY.xattr."+attr.Name] = attrString
-		}
-	}
-
-	return tmpMap
+	return nil
 }

cmd/restic/cmd_dump_test.go

@@ -0,0 +1,27 @@
+package main
+
+import (
+	"testing"
+
+	rtest "github.com/restic/restic/internal/test"
+)
+
+func TestDumpSplitPath(t *testing.T) {
+	testPaths := []struct {
+		path   string
+		result []string
+	}{
+		{"", []string{""}},
+		{"test", []string{"test"}},
+		{"test/dir", []string{"test", "dir"}},
+		{"test/dir/sub", []string{"test", "dir", "sub"}},
+		{"/", []string{""}},
+		{"/test", []string{"test"}},
+		{"/test/dir", []string{"test", "dir"}},
+		{"/test/dir/sub", []string{"test", "dir", "sub"}},
+	}
+	for _, path := range testPaths {
+		parts := splitPath(path.path)
+		rtest.Equals(t, path.result, parts)
+	}
+}

cmd/restic/cmd_find.go

@@ -270,7 +270,7 @@ func (f *Finder) findInSnapshot(ctx context.Context, sn *restic.Snapshot) error
 
 			Printf("Unable to load tree %s\n ... which belongs to snapshot %s.\n", parentTreeID, sn.ID())
 
-			return false, walker.SkipNode
+			return false, walker.ErrSkipNode
 		}
 
 		if node == nil {
@@ -314,7 +314,7 @@ func (f *Finder) findInSnapshot(ctx context.Context, sn *restic.Snapshot) error
 
 			if !childMayMatch {
 				ignoreIfNoMatch = true
-				errIfNoMatch = walker.SkipNode
+				errIfNoMatch = walker.ErrSkipNode
 			} else {
 				ignoreIfNoMatch = false
 			}
@@ -354,7 +354,7 @@ func (f *Finder) findIDs(ctx context.Context, sn *restic.Snapshot) error {
 
 			Printf("Unable to load tree %s\n ... which belongs to snapshot %s.\n", parentTreeID, sn.ID())
 
-			return false, walker.SkipNode
+			return false, walker.ErrSkipNode
 		}
 
 		if node == nil {
@@ -417,7 +417,7 @@ func (f *Finder) packsToBlobs(ctx context.Context, packs []string) error {
 	packsFound := 0
 
 	debug.Log("Looking for packs...")
-	err := f.repo.List(ctx, restic.DataFile, func(id restic.ID, size int64) error {
+	err := f.repo.List(ctx, restic.PackFile, func(id restic.ID, size int64) error {
 		if allPacksFound {
 			return nil
 		}
@@ -465,8 +465,8 @@ func (f *Finder) findObjectPack(ctx context.Context, id string, t restic.BlobTyp
 		return
 	}
 
-	blobs, found := idx.Lookup(rid, t)
-	if !found {
+	blobs := idx.Lookup(rid, t)
+	if len(blobs) == 0 {
 		Printf("Object %s not found in the index\n", rid.Str())
 		return
 	}

cmd/restic/cmd_forget.go

@@ -94,34 +94,22 @@ func runForget(opts ForgetOptions, gopts GlobalOptions, args []string) error {
 		return err
 	}
 
-	removeSnapshots := 0
-
 	ctx, cancel := context.WithCancel(gopts.ctx)
 	defer cancel()
 
 	var snapshots restic.Snapshots
+	removeSnIDs := restic.NewIDSet()
 
 	for sn := range FindFilteredSnapshots(ctx, repo, opts.Hosts, opts.Tags, opts.Paths, args) {
 		snapshots = append(snapshots, sn)
 	}
 
+	var jsonGroups []*ForgetGroup
+
 	if len(args) > 0 {
 		// When explicit snapshots args are given, remove them immediately.
 		for _, sn := range snapshots {
-			if !opts.DryRun {
-				h := restic.Handle{Type: restic.SnapshotFile, Name: sn.ID().String()}
-				if err = repo.Backend().Remove(gopts.ctx, h); err != nil {
-					return err
-				}
-				if !gopts.JSON {
-					Verbosef("removed snapshot %v\n", sn.ID().Str())
-				}
-				removeSnapshots++
-			} else {
-				if !gopts.JSON {
-					Verbosef("would have removed snapshot %v\n", sn.ID().Str())
-				}
-			}
+			removeSnIDs.Insert(*sn.ID())
 		}
 	} else {
 		snapshotGroups, _, err := restic.GroupSnapshots(snapshots, opts.GroupBy)
@@ -151,8 +139,6 @@ func runForget(opts ForgetOptions, gopts GlobalOptions, args []string) error {
 				Verbosef("Applying Policy: %v\n", policy)
 			}
 
-			var jsonGroups []*ForgetGroup
-
 			for k, snapshotGroup := range snapshotGroups {
 				if gopts.Verbose >= 1 && !gopts.JSON {
 					err = PrintSnapshotGroupHeader(gopts.stdout, k)
@@ -191,37 +177,37 @@ func runForget(opts ForgetOptions, gopts GlobalOptions, args []string) error {
 
 				jsonGroups = append(jsonGroups, &fg)
 
-				removeSnapshots += len(remove)
-
-				if !opts.DryRun {
-					for _, sn := range remove {
-						h := restic.Handle{Type: restic.SnapshotFile, Name: sn.ID().String()}
-						err = repo.Backend().Remove(gopts.ctx, h)
-						if err != nil {
-							return err
-						}
-					}
-				}
-			}
-
-			if gopts.JSON {
-				err = printJSONForget(gopts.stdout, jsonGroups)
-				if err != nil {
-					return err
+				for _, sn := range remove {
+					removeSnIDs.Insert(*sn.ID())
 				}
 			}
 		}
 	}
 
-	if removeSnapshots > 0 && opts.Prune {
-		if !gopts.JSON {
-			Verbosef("%d snapshots have been removed, running prune\n", removeSnapshots)
-		}
+	if len(removeSnIDs) > 0 {
 		if !opts.DryRun {
-			return pruneRepository(gopts, repo)
+			err := DeleteFilesChecked(gopts, repo, removeSnIDs, restic.SnapshotFile)
+			if err != nil {
+				return err
+			}
+		} else {
+			if !gopts.JSON {
+				Printf("Would have removed the following snapshots:\n%v\n\n", removeSnIDs)
+			}
 		}
 	}
 
+	if gopts.JSON && len(jsonGroups) > 0 {
+		err = printJSONForget(gopts.stdout, jsonGroups)
+		if err != nil {
+			return err
+		}
+	}
+
+	if len(removeSnIDs) > 0 && opts.Prune && !opts.DryRun {
+		return pruneRepository(gopts, repo)
+	}
+
 	return nil
 }
 

cmd/restic/cmd_generate.go

@@ -9,7 +9,7 @@ import (
 )
 
 var cmdGenerate = &cobra.Command{
-	Use:   "generate [command]",
+	Use:   "generate [flags]",
 	Short: "Generate manual pages and auto-completion files (bash, zsh)",
 	Long: `
 The "generate" command writes automatically generated files (like the man pages

cmd/restic/cmd_init.go

@@ -1,6 +1,7 @@
 package main
 
 import (
+	"github.com/restic/chunker"
 	"github.com/restic/restic/internal/errors"
 	"github.com/restic/restic/internal/repository"
 
@@ -20,19 +21,36 @@ Exit status is 0 if the command was successful, and non-zero if there was any er
 `,
 	DisableAutoGenTag: true,
 	RunE: func(cmd *cobra.Command, args []string) error {
-		return runInit(globalOptions, args)
+		return runInit(initOptions, globalOptions, args)
 	},
 }
 
-func init() {
-	cmdRoot.AddCommand(cmdInit)
+// InitOptions bundles all options for the init command.
+type InitOptions struct {
+	secondaryRepoOptions
+	CopyChunkerParameters bool
 }
 
-func runInit(gopts GlobalOptions, args []string) error {
+var initOptions InitOptions
+
+func init() {
+	cmdRoot.AddCommand(cmdInit)
+
+	f := cmdInit.Flags()
+	initSecondaryRepoOptions(f, &initOptions.secondaryRepoOptions, "secondary", "to copy chunker parameters from")
+	f.BoolVar(&initOptions.CopyChunkerParameters, "copy-chunker-params", false, "copy chunker parameters from the secondary repository (useful with the copy command)")
+}
+
+func runInit(opts InitOptions, gopts GlobalOptions, args []string) error {
 	if gopts.Repo == "" {
 		return errors.Fatal("Please specify repository location (-r)")
 	}
 
+	chunkerPolynomial, err := maybeReadChunkerPolynomial(opts, gopts)
+	if err != nil {
+		return err
+	}
+
 	be, err := create(gopts.Repo, gopts.extended)
 	if err != nil {
 		return errors.Fatalf("create repository at %s failed: %v\n", gopts.Repo, err)
@@ -47,7 +65,7 @@ func runInit(gopts GlobalOptions, args []string) error {
 
 	s := repository.New(be)
 
-	err = s.Init(gopts.ctx, gopts.password)
+	err = s.Init(gopts.ctx, gopts.password, chunkerPolynomial)
 	if err != nil {
 		return errors.Fatalf("create key in repository at %s failed: %v\n", gopts.Repo, err)
 	}
@@ -60,3 +78,25 @@ func runInit(gopts GlobalOptions, args []string) error {
 
 	return nil
 }
+
+func maybeReadChunkerPolynomial(opts InitOptions, gopts GlobalOptions) (*chunker.Pol, error) {
+	if opts.CopyChunkerParameters {
+		otherGopts, err := fillSecondaryGlobalOpts(opts.secondaryRepoOptions, gopts, "secondary")
+		if err != nil {
+			return nil, err
+		}
+
+		otherRepo, err := OpenRepository(otherGopts)
+		if err != nil {
+			return nil, err
+		}
+
+		pol := otherRepo.Config().ChunkerPolynomial
+		return &pol, nil
+	}
+
+	if opts.Repo != "" {
+		return nil, errors.Fatal("Secondary repository must only be specified when copying the chunker parameters")
+	}
+	return nil, nil
+}

cmd/restic/cmd_key.go

@@ -16,7 +16,7 @@ import (
 )
 
 var cmdKey = &cobra.Command{
-	Use:   "key [list|add|remove|passwd] [ID]",
+	Use:   "key [flags] [list|add|remove|passwd] [ID]",
 	Short: "Manage keys (passwords)",
 	Long: `
 The "key" command manages keys (passwords) for accessing the repository.
@@ -32,13 +32,19 @@ Exit status is 0 if the command was successful, and non-zero if there was any er
 	},
 }
 
-var newPasswordFile string
+var (
+	newPasswordFile string
+	keyUsername     string
+	keyHostname     string
+)
 
 func init() {
 	cmdRoot.AddCommand(cmdKey)
 
 	flags := cmdKey.Flags()
-	flags.StringVarP(&newPasswordFile, "new-password-file", "", "", "the file from which to load a new password")
+	flags.StringVarP(&newPasswordFile, "new-password-file", "", "", "`file` from which to read the new password")
+	flags.StringVarP(&keyUsername, "user", "", "", "the username for new keys")
+	flags.StringVarP(&keyHostname, "host", "", "", "the hostname for new keys")
 }
 
 func listKeys(ctx context.Context, s *repository.Repository, gopts GlobalOptions) error {
@@ -110,7 +116,7 @@ func getNewPassword(gopts GlobalOptions) (string, error) {
 	newopts.password = ""
 
 	return ReadPasswordTwice(newopts,
-		"enter password for new key: ",
+		"enter new password: ",
 		"enter password again: ")
 }
 
@@ -120,7 +126,7 @@ func addKey(gopts GlobalOptions, repo *repository.Repository) error {
 		return err
 	}
 
-	id, err := repository.AddKey(gopts.ctx, repo, pw, repo.Key())
+	id, err := repository.AddKey(gopts.ctx, repo, pw, keyUsername, keyHostname, repo.Key())
 	if err != nil {
 		return errors.Fatalf("creating new key failed: %v\n", err)
 	}
@@ -151,7 +157,7 @@ func changePassword(gopts GlobalOptions, repo *repository.Repository) error {
 		return err
 	}
 
-	id, err := repository.AddKey(gopts.ctx, repo, pw, repo.Key())
+	id, err := repository.AddKey(gopts.ctx, repo, pw, "", "", repo.Key())
 	if err != nil {
 		return errors.Fatalf("creating new key failed: %v\n", err)
 	}

cmd/restic/cmd_list.go

@@ -1,17 +1,15 @@
 package main
 
 import (
-	"fmt"
-
 	"github.com/restic/restic/internal/errors"
-	"github.com/restic/restic/internal/index"
+	"github.com/restic/restic/internal/repository"
 	"github.com/restic/restic/internal/restic"
 
 	"github.com/spf13/cobra"
 )
 
 var cmdList = &cobra.Command{
-	Use:   "list [blobs|packs|index|snapshots|keys|locks]",
+	Use:   "list [flags] [blobs|packs|index|snapshots|keys|locks]",
 	Short: "List objects in the repository",
 	Long: `
 The "list" command allows listing objects in the repository based on type.
@@ -52,7 +50,7 @@ func runList(cmd *cobra.Command, opts GlobalOptions, args []string) error {
 	var t restic.FileType
 	switch args[0] {
 	case "packs":
-		t = restic.DataFile
+		t = restic.PackFile
 	case "index":
 		t = restic.IndexFile
 	case "snapshots":
@@ -62,18 +60,17 @@ func runList(cmd *cobra.Command, opts GlobalOptions, args []string) error {
 	case "locks":
 		t = restic.LockFile
 	case "blobs":
-		idx, err := index.Load(opts.ctx, repo, nil)
-		if err != nil {
-			return err
-		}
-
-		for _, pack := range idx.Packs {
-			for _, entry := range pack.Entries {
-				fmt.Printf("%v %v\n", entry.Type, entry.ID)
+		return repo.List(opts.ctx, restic.IndexFile, func(id restic.ID, size int64) error {
+			idx, err := repository.LoadIndex(opts.ctx, repo, id)
+			if err != nil {
+				return err
 			}
-		}
+			for blobs := range idx.Each(opts.ctx) {
+				Printf("%v %v\n", blobs.Type, blobs.ID)
+			}
+			return nil
+		})
 
-		return nil
 	default:
 		return errors.Fatal("invalid type")
 	}

cmd/restic/cmd_ls.go

@@ -16,7 +16,7 @@ import (
 )
 
 var cmdLs = &cobra.Command{
-	Use:   "ls [flags] [snapshotID] [dir...]",
+	Use:   "ls [flags] snapshotID [dir...]",
 	Short: "List files in a snapshot",
 	Long: `
 The "ls" command lists files and directories in a snapshot.
@@ -89,8 +89,8 @@ type lsNode struct {
 }
 
 func runLs(opts LsOptions, gopts GlobalOptions, args []string) error {
-	if len(args) == 0 && len(opts.Hosts) == 0 && len(opts.Tags) == 0 && len(opts.Paths) == 0 {
-		return errors.Fatal("Invalid arguments, either give one or more snapshot IDs or set filters.")
+	if len(args) == 0 {
+		return errors.Fatal("no snapshot ID specified")
 	}
 
 	// extract any specific directories to walk
@@ -222,7 +222,7 @@ func runLs(opts LsOptions, gopts GlobalOptions, args []string) error {
 			// otherwise, signal the walker to not walk recursively into any
 			// subdirs
 			if node.Type == "dir" {
-				return false, walker.SkipNode
+				return false, walker.ErrSkipNode
 			}
 			return false, nil
 		})

cmd/restic/cmd_migrate.go

@@ -8,7 +8,7 @@ import (
 )
 
 var cmdMigrate = &cobra.Command{
-	Use:   "migrate [name]",
+	Use:   "migrate [flags] [name]",
 	Short: "Apply migrations",
 	Long: `
 The "migrate" command applies migrations to a repository. When no migration

cmd/restic/cmd_mount.go

@@ -1,7 +1,4 @@
-// +build !netbsd
-// +build !openbsd
-// +build !solaris
-// +build !windows
+// +build darwin freebsd linux
 
 package main
 
@@ -93,10 +90,12 @@ func mount(opts MountOptions, gopts GlobalOptions, mountpoint string) error {
 		return err
 	}
 
-	lock, err := lockRepo(repo)
-	defer unlockRepo(lock)
-	if err != nil {
-		return err
+	if !gopts.NoLock {
+		lock, err := lockRepo(repo)
+		defer unlockRepo(lock)
+		if err != nil {
+			return err
+		}
 	}
 
 	err = repo.LoadIndex(gopts.ctx)
@@ -142,10 +141,7 @@ func mount(opts MountOptions, gopts GlobalOptions, mountpoint string) error {
 		Paths:            opts.Paths,
 		SnapshotTemplate: opts.SnapshotTemplate,
 	}
-	root, err := fuse.NewRoot(gopts.ctx, repo, cfg)
-	if err != nil {
-		return err
-	}
+	root := fuse.NewRoot(repo, cfg)
 
 	Printf("Now serving the repository at %s\n", mountpoint)
 	Printf("When finished, quit with Ctrl-c or umount the mountpoint.\n")
@@ -183,7 +179,7 @@ func runMount(opts MountOptions, gopts GlobalOptions, args []string) error {
 		debug.Log("running umount cleanup handler for mount at %v", mountpoint)
 		err := umount(mountpoint)
 		if err != nil {
-			Warnf("unable to umount (maybe already umounted?): %v\n", err)
+			Warnf("unable to umount (maybe already umounted or still in use?): %v\n", err)
 		}
 		return nil
 	})

cmd/restic/cmd_prune.go

@@ -1,9 +1,6 @@
 package main
 
 import (
-	"fmt"
-	"time"
-
 	"github.com/restic/restic/internal/debug"
 	"github.com/restic/restic/internal/errors"
 	"github.com/restic/restic/internal/index"
@@ -47,34 +44,6 @@ func shortenStatus(maxLength int, s string) string {
 	return s[:maxLength-3] + "..."
 }
 
-// newProgressMax returns a progress that counts blobs.
-func newProgressMax(show bool, max uint64, description string) *restic.Progress {
-	if !show {
-		return nil
-	}
-
-	p := restic.NewProgress()
-
-	p.OnUpdate = func(s restic.Stat, d time.Duration, ticker bool) {
-		status := fmt.Sprintf("[%s] %s  %d / %d %s",
-			formatDuration(d),
-			formatPercent(s.Blobs, max),
-			s.Blobs, max, description)
-
-		if w := stdoutTerminalWidth(); w > 0 {
-			status = shortenStatus(w, status)
-		}
-
-		PrintProgress("%s", status)
-	}
-
-	p.OnDone = func(s restic.Stat, d time.Duration, ticker bool) {
-		fmt.Printf("\n")
-	}
-
-	return p
-}
-
 func runPrune(gopts GlobalOptions) error {
 	repo, err := OpenRepository(gopts)
 	if err != nil {
@@ -87,6 +56,9 @@ func runPrune(gopts GlobalOptions) error {
 		return err
 	}
 
+	// we do not need index updates while pruning!
+	repo.DisableAutoIndexUpdate()
+
 	return pruneRepository(gopts, repo)
 }
 
@@ -125,7 +97,7 @@ func pruneRepository(gopts GlobalOptions, repo restic.Repository) error {
 	}
 
 	Verbosef("counting files in repo\n")
-	err = repo.List(ctx, restic.DataFile, func(restic.ID, int64) error {
+	err = repo.List(ctx, restic.PackFile, func(restic.ID, int64) error {
 		stats.packs++
 		return nil
 	})
@@ -183,34 +155,22 @@ func pruneRepository(gopts GlobalOptions, repo restic.Repository) error {
 
 	stats.snapshots = len(snapshots)
 
-	Verbosef("find data that is still in use for %d snapshots\n", stats.snapshots)
-
-	usedBlobs := restic.NewBlobSet()
-	seenBlobs := restic.NewBlobSet()
-
-	bar = newProgressMax(!gopts.Quiet, uint64(len(snapshots)), "snapshots")
-	bar.Start()
-	for _, sn := range snapshots {
-		debug.Log("process snapshot %v", sn.ID())
-
-		err = restic.FindUsedBlobs(ctx, repo, *sn.Tree, usedBlobs, seenBlobs)
-		if err != nil {
-			if repo.Backend().IsNotExist(err) {
-				return errors.Fatal("unable to load a tree from the repo: " + err.Error())
-			}
-
-			return err
-		}
-
-		debug.Log("processed snapshot %v", sn.ID())
-		bar.Report(restic.Stat{Blobs: 1})
+	usedBlobs, err := getUsedBlobs(gopts, repo, snapshots)
+	if err != nil {
+		return err
 	}
-	bar.Done()
 
-	if len(usedBlobs) > stats.blobs {
-		return errors.Fatalf("number of used blobs is larger than number of available blobs!\n" +
-			"Please report this error (along with the output of the 'prune' run) at\n" +
-			"https://github.com/restic/restic/issues/new")
+	var missingBlobs []restic.BlobHandle
+	for h := range usedBlobs {
+		if _, ok := blobCount[h]; !ok {
+			missingBlobs = append(missingBlobs, h)
+		}
+	}
+	if len(missingBlobs) > 0 {
+		return errors.Fatalf("%v not found in the new index\n"+
+			"Data blobs seem to be missing, aborting prune to prevent further data loss!\n"+
+			"Please report this error (along with the output of the 'prune' run) at\n"+
+			"https://github.com/restic/restic/issues/new/choose", missingBlobs)
 	}
 
 	Verbosef("found %d of %d data blobs still in use, removing %d blobs\n",
@@ -278,13 +238,11 @@ func pruneRepository(gopts GlobalOptions, repo restic.Repository) error {
 
 	var obsoletePacks restic.IDSet
 	if len(rewritePacks) != 0 {
-		bar = newProgressMax(!gopts.Quiet, uint64(len(rewritePacks)), "packs rewritten")
-		bar.Start()
+		bar := newProgressMax(!gopts.Quiet, uint64(len(rewritePacks)), "packs rewritten")
 		obsoletePacks, err = repository.Repack(ctx, repo, rewritePacks, usedBlobs, bar)
 		if err != nil {
 			return err
 		}
-		bar.Done()
 	}
 
 	removePacks.Merge(obsoletePacks)
@@ -294,19 +252,38 @@ func pruneRepository(gopts GlobalOptions, repo restic.Repository) error {
 	}
 
 	if len(removePacks) != 0 {
-		bar = newProgressMax(!gopts.Quiet, uint64(len(removePacks)), "packs deleted")
-		bar.Start()
-		for packID := range removePacks {
-			h := restic.Handle{Type: restic.DataFile, Name: packID.String()}
-			err = repo.Backend().Remove(ctx, h)
-			if err != nil {
-				Warnf("unable to remove file %v from the repository\n", packID.Str())
-			}
-			bar.Report(restic.Stat{Blobs: 1})
-		}
-		bar.Done()
+		Verbosef("remove %d old packs\n", len(removePacks))
+		DeleteFiles(gopts, repo, removePacks, restic.PackFile)
 	}
 
 	Verbosef("done\n")
 	return nil
 }
+
+func getUsedBlobs(gopts GlobalOptions, repo restic.Repository, snapshots []*restic.Snapshot) (usedBlobs restic.BlobSet, err error) {
+	ctx := gopts.ctx
+
+	Verbosef("find data that is still in use for %d snapshots\n", len(snapshots))
+
+	usedBlobs = restic.NewBlobSet()
+
+	bar := newProgressMax(!gopts.Quiet, uint64(len(snapshots)), "snapshots")
+	bar.Start()
+	defer bar.Done()
+	for _, sn := range snapshots {
+		debug.Log("process snapshot %v", sn.ID())
+
+		err = restic.FindUsedBlobs(ctx, repo, *sn.Tree, usedBlobs)
+		if err != nil {
+			if repo.Backend().IsNotExist(err) {
+				return nil, errors.Fatal("unable to load a tree from the repo: " + err.Error())
+			}
+
+			return nil, err
+		}
+
+		debug.Log("processed snapshot %v", sn.ID())
+		bar.Report(restic.Stat{Blobs: 1})
+	}
+	return usedBlobs, nil
+}

cmd/restic/cmd_rebuild_index.go

@@ -53,7 +53,7 @@ func rebuildIndex(ctx context.Context, repo restic.Repository, ignorePacks resti
 	Verbosef("counting files in repo\n")
 
 	var packs uint64
-	err := repo.List(ctx, restic.DataFile, func(restic.ID, int64) error {
+	err := repo.List(ctx, restic.PackFile, func(restic.ID, int64) error {
 		packs++
 		return nil
 	})
@@ -92,14 +92,9 @@ func rebuildIndex(ctx context.Context, repo restic.Repository, ignorePacks resti
 	Verbosef("saved new indexes as %v\n", ids)
 
 	Verbosef("remove %d old index files\n", len(supersedes))
-
-	for _, id := range supersedes {
-		if err := repo.Backend().Remove(ctx, restic.Handle{
-			Type: restic.IndexFile,
-			Name: id.String(),
-		}); err != nil {
-			Warnf("error removing old index %v: %v\n", id.Str(), err)
-		}
+	err = DeleteFilesChecked(globalOptions, repo, restic.NewIDSet(supersedes...), restic.IndexFile)
+	if err != nil {
+		return errors.Fatalf("unable to remove an old index: %v\n", err)
 	}
 
 	return nil

cmd/restic/cmd_recover.go

@@ -130,11 +130,6 @@ func runRecover(gopts GlobalOptions) error {
 		return errors.Fatalf("unable to save blobs to the repo: %v", err)
 	}
 
-	err = repo.SaveIndex(gopts.ctx)
-	if err != nil {
-		return errors.Fatalf("unable to save new index to the repo: %v", err)
-	}
-
 	sn, err := restic.NewSnapshot([]string{"/recover"}, []string{}, hostname, time.Now())
 	if err != nil {
 		return errors.Fatalf("unable to save snapshot: %v", err)

cmd/restic/cmd_snapshots.go

@@ -14,7 +14,7 @@ import (
 )
 
 var cmdSnapshots = &cobra.Command{
-	Use:   "snapshots [snapshotID ...]",
+	Use:   "snapshots [flags] [snapshotID ...]",
 	Short: "List all snapshots",
 	Long: `
 The "snapshots" command lists all snapshots stored in the repository.
@@ -251,9 +251,8 @@ func PrintSnapshots(stdout io.Writer, list restic.Snapshots, reasons []restic.Ke
 // Prints nothing, if we did not group at all.
 func PrintSnapshotGroupHeader(stdout io.Writer, groupKeyJSON string) error {
 	var key restic.SnapshotGroupKey
-	var err error
 
-	err = json.Unmarshal([]byte(groupKeyJSON), &key)
+	err := json.Unmarshal([]byte(groupKeyJSON), &key)
 	if err != nil {
 		return err
 	}

cmd/restic/cmd_stats.go

@@ -2,31 +2,31 @@ package main
 
 import (
 	"context"
-	"crypto/sha256"
 	"encoding/json"
 	"fmt"
-	"os"
 	"path/filepath"
 
-	"github.com/restic/restic/internal/errors"
 	"github.com/restic/restic/internal/restic"
 	"github.com/restic/restic/internal/walker"
+
+	"github.com/minio/sha256-simd"
 	"github.com/spf13/cobra"
 )
 
 var cmdStats = &cobra.Command{
-	Use:   "stats [flags] [snapshot-ID]",
+	Use:   "stats [flags] [snapshot ID] [...]",
 	Short: "Scan the repository and show basic statistics",
 	Long: `
-The "stats" command walks one or all snapshots in a repository and
-accumulates statistics about the data stored therein. It reports on
-the number of unique files and their sizes, according to one of
+The "stats" command walks one or multiple snapshots in a repository
+and accumulates statistics about the data stored therein. It reports 
+on the number of unique files and their sizes, according to one of
 the counting modes as given by the --mode flag.
 
-If no snapshot is specified, all snapshots will be considered. Some
-modes make more sense over just a single snapshot, while others
-are useful across all snapshots, depending on what you are trying
-to calculate.
+It operates on all snapshots matching the selection criteria or all
+snapshots if nothing is specified. The special snapshot ID "latest"
+is also supported. Some modes make more sense over 
+just a single snapshot, while others are useful across all snapshots,
+depending on what you are trying to calculate.
 
 The modes are:
 
@@ -50,11 +50,26 @@ Exit status is 0 if the command was successful, and non-zero if there was any er
 	},
 }
 
+// StatsOptions collects all options for the stats command.
+type StatsOptions struct {
+	// the mode of counting to perform (see consts for available modes)
+	countMode string
+
+	// filter snapshots by, if given by user
+	Hosts []string
+	Tags  restic.TagLists
+	Paths []string
+}
+
+var statsOptions StatsOptions
+
 func init() {
 	cmdRoot.AddCommand(cmdStats)
 	f := cmdStats.Flags()
-	f.StringVar(&countMode, "mode", countModeRestoreSize, "counting mode: restore-size (default), files-by-contents, blobs-per-file, or raw-data")
-	f.StringArrayVarP(&snapshotByHosts, "host", "H", nil, "filter latest snapshot by this hostname (can be specified multiple times)")
+	f.StringVar(&statsOptions.countMode, "mode", countModeRestoreSize, "counting mode: restore-size (default), files-by-contents, blobs-per-file or raw-data")
+	f.StringArrayVarP(&statsOptions.Hosts, "host", "H", nil, "only consider snapshots with the given `host` (can be specified multiple times)")
+	f.Var(&statsOptions.Tags, "tag", "only consider snapshots which include this `taglist` in the format `tag[,tag,...]` (can be specified multiple times)")
+	f.StringArrayVar(&statsOptions.Paths, "path", nil, "only consider snapshots which include this (absolute) `path` (can be specified multiple times)")
 }
 
 func runStats(gopts GlobalOptions, args []string) error {
@@ -89,53 +104,25 @@ func runStats(gopts GlobalOptions, args []string) error {
 
 	// create a container for the stats (and other needed state)
 	stats := &statsContainer{
-		uniqueFiles:  make(map[fileID]struct{}),
-		uniqueInodes: make(map[uint64]struct{}),
-		fileBlobs:    make(map[string]restic.IDSet),
-		blobs:        restic.NewBlobSet(),
-		blobsSeen:    restic.NewBlobSet(),
+		uniqueFiles:    make(map[fileID]struct{}),
+		uniqueInodes:   make(map[uint64]struct{}),
+		fileBlobs:      make(map[string]restic.IDSet),
+		blobs:          restic.NewBlobSet(),
+		snapshotsCount: 0,
 	}
 
-	if snapshotIDString != "" {
-		// scan just a single snapshot
-
-		var sID restic.ID
-		if snapshotIDString == "latest" {
-			sID, err = restic.FindLatestSnapshot(ctx, repo, []string{}, []restic.TagList{}, snapshotByHosts)
-			if err != nil {
-				return errors.Fatalf("latest snapshot for criteria not found: %v", err)
-			}
-		} else {
-			sID, err = restic.FindSnapshot(repo, snapshotIDString)
-			if err != nil {
-				return errors.Fatalf("error loading snapshot: %v", err)
-			}
-		}
-
-		snapshot, err := restic.LoadSnapshot(ctx, repo, sID)
-		if err != nil {
-			return errors.Fatalf("error loading snapshot from repo: %v", err)
-		}
-
-		err = statsWalkSnapshot(ctx, snapshot, repo, stats)
+	for sn := range FindFilteredSnapshots(ctx, repo, statsOptions.Hosts, statsOptions.Tags, statsOptions.Paths, args) {
+		err = statsWalkSnapshot(ctx, sn, repo, stats)
 		if err != nil {
 			return fmt.Errorf("error walking snapshot: %v", err)
 		}
-	} else {
-		// iterate every snapshot in the repo
-		err = repo.List(ctx, restic.SnapshotFile, func(snapshotID restic.ID, size int64) error {
-			snapshot, err := restic.LoadSnapshot(ctx, repo, snapshotID)
-			if err != nil {
-				return fmt.Errorf("Error loading snapshot %s: %v", snapshotID.Str(), err)
-			}
-			return statsWalkSnapshot(ctx, snapshot, repo, stats)
-		})
 	}
+
 	if err != nil {
 		return err
 	}
 
-	if countMode == countModeRawData {
+	if statsOptions.countMode == countModeRawData {
 		// the blob handles have been collected, but not yet counted
 		for blobHandle := range stats.blobs {
 			blobSize, found := repo.LookupBlobSize(blobHandle.ID, blobHandle.Type)
@@ -148,29 +135,23 @@ func runStats(gopts GlobalOptions, args []string) error {
 	}
 
 	if gopts.JSON {
-		err = json.NewEncoder(os.Stdout).Encode(stats)
+		err = json.NewEncoder(globalOptions.stdout).Encode(stats)
 		if err != nil {
 			return fmt.Errorf("encoding output: %v", err)
 		}
 		return nil
 	}
 
-	// inform the user what was scanned and how it was scanned
-	snapshotsScanned := snapshotIDString
-	if snapshotsScanned == "latest" {
-		snapshotsScanned = "the latest snapshot"
-	} else if snapshotsScanned == "" {
-		snapshotsScanned = "all snapshots"
-	}
-	Printf("Stats for %s in %s mode:\n", snapshotsScanned, countMode)
+	Printf("Stats in %s mode:\n", statsOptions.countMode)
+	Printf("Snapshots processed:   %d\n", stats.snapshotsCount)
 
 	if stats.TotalBlobCount > 0 {
-		Printf("  Total Blob Count:   %d\n", stats.TotalBlobCount)
+		Printf("   Total Blob Count:   %d\n", stats.TotalBlobCount)
 	}
 	if stats.TotalFileCount > 0 {
-		Printf("  Total File Count:   %d\n", stats.TotalFileCount)
+		Printf("   Total File Count:   %d\n", stats.TotalFileCount)
 	}
-	Printf("        Total Size:   %-5s\n", formatBytes(stats.TotalSize))
+	Printf("         Total Size:   %-5s\n", formatBytes(stats.TotalSize))
 
 	return nil
 }
@@ -180,16 +161,19 @@ func statsWalkSnapshot(ctx context.Context, snapshot *restic.Snapshot, repo rest
 		return fmt.Errorf("snapshot %s has nil tree", snapshot.ID().Str())
 	}
 
-	if countMode == countModeRawData {
+	stats.snapshotsCount++
+
+	if statsOptions.countMode == countModeRawData {
 		// count just the sizes of unique blobs; we don't need to walk the tree
 		// ourselves in this case, since a nifty function does it for us
-		return restic.FindUsedBlobs(ctx, repo, *snapshot.Tree, stats.blobs, stats.blobsSeen)
+		return restic.FindUsedBlobs(ctx, repo, *snapshot.Tree, stats.blobs)
 	}
 
 	err := walker.Walk(ctx, repo, *snapshot.Tree, restic.NewIDSet(), statsWalkTree(repo, stats))
 	if err != nil {
 		return fmt.Errorf("walking tree %s: %v", *snapshot.Tree, err)
 	}
+
 	return nil
 }
 
@@ -202,19 +186,19 @@ func statsWalkTree(repo restic.Repository, stats *statsContainer) walker.WalkFun
 			return true, nil
 		}
 
-		if countMode == countModeUniqueFilesByContents || countMode == countModeBlobsPerFile {
+		if statsOptions.countMode == countModeUniqueFilesByContents || statsOptions.countMode == countModeBlobsPerFile {
 			// only count this file if we haven't visited it before
 			fid := makeFileIDByContents(node)
 			if _, ok := stats.uniqueFiles[fid]; !ok {
 				// mark the file as visited
 				stats.uniqueFiles[fid] = struct{}{}
 
-				if countMode == countModeUniqueFilesByContents {
+				if statsOptions.countMode == countModeUniqueFilesByContents {
 					// simply count the size of each unique file (unique by contents only)
 					stats.TotalSize += node.Size
 					stats.TotalFileCount++
 				}
-				if countMode == countModeBlobsPerFile {
+				if statsOptions.countMode == countModeBlobsPerFile {
 					// count the size of each unique blob reference, which is
 					// by unique file (unique by contents and file path)
 					for _, blobID := range node.Content {
@@ -244,7 +228,7 @@ func statsWalkTree(repo restic.Repository, stats *statsContainer) walker.WalkFun
 			}
 		}
 
-		if countMode == countModeRestoreSize {
+		if statsOptions.countMode == countModeRestoreSize {
 			// as this is a file in the snapshot, we can simply count its
 			// size without worrying about uniqueness, since duplicate files
 			// will still be restored
@@ -276,23 +260,13 @@ func makeFileIDByContents(node *restic.Node) fileID {
 
 func verifyStatsInput(gopts GlobalOptions, args []string) error {
 	// require a recognized counting mode
-	switch countMode {
+	switch statsOptions.countMode {
 	case countModeRestoreSize:
 	case countModeUniqueFilesByContents:
 	case countModeBlobsPerFile:
 	case countModeRawData:
 	default:
-		return fmt.Errorf("unknown counting mode: %s (use the -h flag to get a list of supported modes)", countMode)
-	}
-
-	// ensure at most one snapshot was specified
-	if len(args) > 1 {
-		return fmt.Errorf("only one snapshot may be specified")
-	}
-
-	// if a snapshot was specified, mark it as the one to scan
-	if len(args) == 1 {
-		snapshotIDString = args[0]
+		return fmt.Errorf("unknown counting mode: %s (use the -h flag to get a list of supported modes)", statsOptions.countMode)
 	}
 
 	return nil
@@ -318,26 +292,17 @@ type statsContainer struct {
 	// blobs that have been seen as a part of the file
 	fileBlobs map[string]restic.IDSet
 
-	// blobs and blobsSeen are used to count individual
-	// unique blobs, independent of references to files
-	blobs, blobsSeen restic.BlobSet
+	// blobs is used to count individual unique blobs,
+	// independent of references to files
+	blobs restic.BlobSet
+
+	// holds count of all considered snapshots
+	snapshotsCount int
 }
 
 // fileID is a 256-bit hash that distinguishes unique files.
 type fileID [32]byte
 
-var (
-	// the mode of counting to perform
-	countMode string
-
-	// the snapshot to scan, as given by the user
-	snapshotIDString string
-
-	// snapshotByHost is the host to filter latest
-	// snapshot by, if given by user
-	snapshotByHosts []string
-)
-
 const (
 	countModeRestoreSize           = "restore-size"
 	countModeUniqueFilesByContents = "files-by-contents"

cmd/restic/delete.go

@@ -0,0 +1,62 @@
+package main
+
+import (
+	"golang.org/x/sync/errgroup"
+
+	"github.com/restic/restic/internal/restic"
+)
+
+// DeleteFiles deletes the given fileList of fileType in parallel
+// it will print a warning if there is an error, but continue deleting the remaining files
+func DeleteFiles(gopts GlobalOptions, repo restic.Repository, fileList restic.IDSet, fileType restic.FileType) {
+	deleteFiles(gopts, true, repo, fileList, fileType)
+}
+
+// DeleteFilesChecked deletes the given fileList of fileType in parallel
+// if an error occurs, it will cancel and return this error
+func DeleteFilesChecked(gopts GlobalOptions, repo restic.Repository, fileList restic.IDSet, fileType restic.FileType) error {
+	return deleteFiles(gopts, false, repo, fileList, fileType)
+}
+
+const numDeleteWorkers = 8
+
+// deleteFiles deletes the given fileList of fileType in parallel
+// if ignoreError=true, it will print a warning if there was an error, else it will abort.
+func deleteFiles(gopts GlobalOptions, ignoreError bool, repo restic.Repository, fileList restic.IDSet, fileType restic.FileType) error {
+	totalCount := len(fileList)
+	fileChan := make(chan restic.ID)
+	go func() {
+		for id := range fileList {
+			fileChan <- id
+		}
+		close(fileChan)
+	}()
+
+	bar := newProgressMax(!gopts.JSON && !gopts.Quiet, uint64(totalCount), "files deleted")
+	wg, ctx := errgroup.WithContext(gopts.ctx)
+	bar.Start()
+	for i := 0; i < numDeleteWorkers; i++ {
+		wg.Go(func() error {
+			for id := range fileChan {
+				h := restic.Handle{Type: fileType, Name: id.String()}
+				err := repo.Backend().Remove(ctx, h)
+				if err != nil {
+					if !gopts.JSON {
+						Warnf("unable to remove %v from the repository\n", h)
+					}
+					if !ignoreError {
+						return err
+					}
+				}
+				if !gopts.JSON && gopts.verbosity >= 2 {
+					Verbosef("removed %v\n", h)
+				}
+				bar.Report(restic.Stat{Blobs: 1})
+			}
+			return nil
+		})
+	}
+	err := wg.Wait()
+	bar.Done()
+	return err
+}

cmd/restic/exclude.go

@@ -6,6 +6,7 @@ import (
 	"io"
 	"os"
 	"path/filepath"
+	"strconv"
 	"strings"
 	"sync"
 
@@ -131,7 +132,7 @@ func rejectIfPresent(excludeFileSpec string) (RejectByNameFunc, error) {
 }
 
 // isExcludedByFile interprets filename as a path and returns true if that file
-// is in a excluded directory. A directory is identified as excluded if it contains a
+// is in an excluded directory. A directory is identified as excluded if it contains a
 // tagfile which bears the name specified in tagFilename and starts with
 // header. If rc is non-nil, it is used to expedite the evaluation of a
 // directory based on previous visits.
@@ -190,7 +191,7 @@ func isDirExcludedByFile(dir, tagFilename, header string) bool {
 		Warnf("could not read signature from exclusion tagfile %q: %v\n", tf, err)
 		return false
 	}
-	if bytes.Compare(buf, []byte(header)) != 0 {
+	if !bytes.Equal(buf, []byte(header)) {
 		Warnf("invalid signature in exclusion tagfile %q\n", tf)
 		return false
 	}
@@ -292,3 +293,50 @@ func rejectResticCache(repo *repository.Repository) (RejectByNameFunc, error) {
 		return false
 	}, nil
 }
+
+func rejectBySize(maxSizeStr string) (RejectFunc, error) {
+	maxSize, err := parseSizeStr(maxSizeStr)
+	if err != nil {
+		return nil, err
+	}
+
+	return func(item string, fi os.FileInfo) bool {
+		// directory will be ignored
+		if fi.IsDir() {
+			return false
+		}
+
+		filesize := fi.Size()
+		if filesize > maxSize {
+			debug.Log("file %s is oversize: %d", item, filesize)
+			return true
+		}
+
+		return false
+	}, nil
+}
+
+func parseSizeStr(sizeStr string) (int64, error) {
+	numStr := sizeStr[:len(sizeStr)-1]
+	var unit int64 = 1
+
+	switch sizeStr[len(sizeStr)-1] {
+	case 'b', 'B':
+		// use initialized values, do nothing here
+	case 'k', 'K':
+		unit = 1024
+	case 'm', 'M':
+		unit = 1024 * 1024
+	case 'g', 'G':
+		unit = 1024 * 1024 * 1024
+	case 't', 'T':
+		unit = 1024 * 1024 * 1024 * 1024
+	default:
+		numStr = sizeStr
+	}
+	value, err := strconv.ParseInt(numStr, 10, 64)
+	if err != nil {
+		return 0, nil
+	}
+	return value * unit, nil
+}

cmd/restic/exclude_test.go

@@ -189,3 +189,113 @@ func TestMultipleIsExcludedByFile(t *testing.T) {
 		}
 	}
 }
+
+func TestParseSizeStr(t *testing.T) {
+	sizeStrTests := []struct {
+		in       string
+		expected int64
+	}{
+		{"1024", 1024},
+		{"1024b", 1024},
+		{"1024B", 1024},
+		{"1k", 1024},
+		{"100k", 102400},
+		{"100K", 102400},
+		{"10M", 10485760},
+		{"100m", 104857600},
+		{"20G", 21474836480},
+		{"10g", 10737418240},
+		{"2T", 2199023255552},
+		{"2t", 2199023255552},
+	}
+
+	for _, tt := range sizeStrTests {
+		actual, err := parseSizeStr(tt.in)
+		test.OK(t, err)
+
+		if actual != tt.expected {
+			t.Errorf("parseSizeStr(%s) = %d; expected %d", tt.in, actual, tt.expected)
+		}
+	}
+}
+
+// TestIsExcludedByFileSize is for testing the instance of
+// --exclude-larger-than parameters
+func TestIsExcludedByFileSize(t *testing.T) {
+	tempDir, cleanup := test.TempDir(t)
+	defer cleanup()
+
+	// Max size of file is set to be 1k
+	maxSizeStr := "1k"
+
+	// Create some files in a temporary directory.
+	// Files in UPPERCASE will be used as exclusion triggers later on.
+	// We will test the inclusion later, so we add the expected value as
+	// a bool.
+	files := []struct {
+		path string
+		size int64
+		incl bool
+	}{
+		{"42", 100, true},
+
+		// everything in foodir except the FOOLARGE tagfile
+		// should not be included.
+		{"foodir/FOOLARGE", 2048, false},
+		{"foodir/foo", 1002, true},
+		{"foodir/foosub/underfoo", 100, true},
+
+		// everything in bardir except the BARLARGE tagfile
+		// should not be included.
+		{"bardir/BARLARGE", 1030, false},
+		{"bardir/bar", 1000, true},
+		{"bardir/barsub/underbar", 500, true},
+
+		// everything in bazdir should be included.
+		{"bazdir/baz", 100, true},
+		{"bazdir/bazsub/underbaz", 200, true},
+	}
+	var errs []error
+	for _, f := range files {
+		// create directories first, then the file
+		p := filepath.Join(tempDir, filepath.FromSlash(f.path))
+		errs = append(errs, os.MkdirAll(filepath.Dir(p), 0700))
+		file, err := os.OpenFile(p, os.O_RDWR|os.O_CREATE|os.O_TRUNC, 0600)
+		errs = append(errs, err)
+		if err == nil {
+			// create a file with given size
+			errs = append(errs, file.Truncate(f.size))
+		}
+		errs = append(errs, file.Close())
+	}
+	test.OKs(t, errs) // see if anything went wrong during the creation
+
+	// create rejection function
+	sizeExclude, _ := rejectBySize(maxSizeStr)
+
+	// To mock the archiver scanning walk, we create filepath.WalkFn
+	// that tests against the two rejection functions and stores
+	// the result in a map against we can test later.
+	m := make(map[string]bool)
+	walk := func(p string, fi os.FileInfo, err error) error {
+		if err != nil {
+			return err
+		}
+
+		excluded := sizeExclude(p, fi)
+		// the log message helps debugging in case the test fails
+		t.Logf("%q: dir:%t; size:%d; excluded:%v", p, fi.IsDir(), fi.Size(), excluded)
+		m[p] = !excluded
+		return nil
+	}
+	// walk through the temporary file and check the error
+	test.OK(t, filepath.Walk(tempDir, walk))
+
+	// compare whether the walk gave the expected values for the test cases
+	for _, f := range files {
+		p := filepath.Join(tempDir, filepath.FromSlash(f.path))
+		if m[p] != f.incl {
+			t.Errorf("inclusion status of %s is wrong: want %v, got %v", f.path, f.incl, m[p])
+		}
+	}
+}

cmd/restic/find.go

@@ -22,10 +22,10 @@ func FindFilteredSnapshots(ctx context.Context, repo *repository.Repository, hos
 			// Process all snapshot IDs given as arguments.
 			for _, s := range snapshotIDs {
 				if s == "latest" {
+					usedFilter = true
 					id, err = restic.FindLatestSnapshot(ctx, repo, paths, tags, hosts)
 					if err != nil {
 						Warnf("Ignoring %q, no snapshot matched given filter (Paths:%v Tags:%v Hosts:%v)\n", s, paths, tags, hosts)
-						usedFilter = true
 						continue
 					}
 				} else {

cmd/restic/format.go

@@ -51,12 +51,6 @@ func formatPercent(numerator uint64, denominator uint64) string {
 	return fmt.Sprintf("%3.2f%%", percent)
 }
 
-func formatRate(bytes uint64, duration time.Duration) string {
-	sec := float64(duration) / float64(time.Second)
-	rate := float64(bytes) / sec / (1 << 20)
-	return fmt.Sprintf("%.2fMiB/s", rate)
-}
-
 func formatDuration(d time.Duration) string {
 	sec := uint64(d / time.Second)
 	return formatSeconds(sec)

cmd/restic/global.go

@@ -39,11 +39,13 @@ import (
 	"golang.org/x/crypto/ssh/terminal"
 )
 
-var version = "0.9.6-dev (compiled manually)"
+var version = "0.10.0"
 
 // TimeFormat is the format used for all timestamps printed by restic.
 const TimeFormat = "2006-01-02 15:04:05"
 
+type backendWrapper func(r restic.Backend) (restic.Backend, error)
+
 // GlobalOptions hold all global options for restic.
 type GlobalOptions struct {
 	Repo            string
@@ -68,11 +70,13 @@ type GlobalOptions struct {
 	stdout   io.Writer
 	stderr   io.Writer
 
+	backendTestHook backendWrapper
+
 	// verbosity is set as follows:
 	//  0 means: don't print any messages except errors, this is used when --quiet is specified
 	//  1 is the default: print essential messages
 	//  2 means: print more messages, report minor things, this is used when --verbose is specified
-	//  3 means: print very detailed debug messages, this is used when --verbose 2 is specified
+	//  3 means: print very detailed debug messages, this is used when --verbose=2 is specified
 	verbosity uint
 
 	Options []string
@@ -97,11 +101,11 @@ func init() {
 
 	f := cmdRoot.PersistentFlags()
 	f.StringVarP(&globalOptions.Repo, "repo", "r", os.Getenv("RESTIC_REPOSITORY"), "`repository` to backup to or restore from (default: $RESTIC_REPOSITORY)")
-	f.StringVarP(&globalOptions.PasswordFile, "password-file", "p", os.Getenv("RESTIC_PASSWORD_FILE"), "read the repository password from a `file` (default: $RESTIC_PASSWORD_FILE)")
+	f.StringVarP(&globalOptions.PasswordFile, "password-file", "p", os.Getenv("RESTIC_PASSWORD_FILE"), "`file` to read the repository password from (default: $RESTIC_PASSWORD_FILE)")
 	f.StringVarP(&globalOptions.KeyHint, "key-hint", "", os.Getenv("RESTIC_KEY_HINT"), "`key` ID of key to try decrypting first (default: $RESTIC_KEY_HINT)")
-	f.StringVarP(&globalOptions.PasswordCommand, "password-command", "", os.Getenv("RESTIC_PASSWORD_COMMAND"), "specify a shell `command` to obtain a password (default: $RESTIC_PASSWORD_COMMAND)")
+	f.StringVarP(&globalOptions.PasswordCommand, "password-command", "", os.Getenv("RESTIC_PASSWORD_COMMAND"), "shell `command` to obtain the repository password from (default: $RESTIC_PASSWORD_COMMAND)")
 	f.BoolVarP(&globalOptions.Quiet, "quiet", "q", false, "do not output comprehensive progress report")
-	f.CountVarP(&globalOptions.Verbose, "verbose", "v", "be verbose (specify --verbose multiple times or level `n`)")
+	f.CountVarP(&globalOptions.Verbose, "verbose", "v", "be verbose (specify --verbose multiple times or level --verbose=`n`)")
 	f.BoolVar(&globalOptions.NoLock, "no-lock", false, "do not lock the repo, this allows some operations on read-only repos")
 	f.BoolVarP(&globalOptions.JSON, "json", "", false, "set output mode to JSON for commands that support it")
 	f.StringVar(&globalOptions.CacheDir, "cache-dir", "", "set the cache `directory`. (default: use system default cache directory)")
@@ -270,7 +274,7 @@ func Exitf(exitcode int, format string, args ...interface{}) {
 }
 
 // resolvePassword determines the password to be used for opening the repository.
-func resolvePassword(opts GlobalOptions) (string, error) {
+func resolvePassword(opts GlobalOptions, envStr string) (string, error) {
 	if opts.PasswordFile != "" && opts.PasswordCommand != "" {
 		return "", errors.Fatalf("Password file and command are mutually exclusive options")
 	}
@@ -295,7 +299,7 @@ func resolvePassword(opts GlobalOptions) (string, error) {
 		return strings.TrimSpace(string(s)), errors.Wrap(err, "Readfile")
 	}
 
-	if pwd := os.Getenv("RESTIC_PASSWORD"); pwd != "" {
+	if pwd := os.Getenv(envStr); pwd != "" {
 		return pwd, nil
 	}
 
@@ -395,6 +399,14 @@ func OpenRepository(opts GlobalOptions) (*repository.Repository, error) {
 		Warnf("%v returned error, retrying after %v: %v\n", msg, d, err)
 	})
 
+	// wrap backend if a test specified a hook
+	if opts.backendTestHook != nil {
+		be, err = opts.backendTestHook(be)
+		if err != nil {
+			return nil, err
+		}
+	}
+
 	s := repository.New(be)
 
 	passwordTriesLeft := 1

cmd/restic/integration_helpers_test.go

@@ -1,6 +1,7 @@
 package main
 
 import (
+	"bytes"
 	"context"
 	"fmt"
 	"io/ioutil"
@@ -54,7 +55,7 @@ func walkDir(dir string) <-chan *dirEntry {
 	}()
 
 	// first element is root
-	_ = <-ch
+	<-ch
 
 	return ch
 }
@@ -72,27 +73,16 @@ func sameModTime(fi1, fi2 os.FileInfo) bool {
 		}
 	}
 
-	same := fi1.ModTime().Equal(fi2.ModTime())
-	if !same && (runtime.GOOS == "darwin" || runtime.GOOS == "openbsd") {
-		// Allow up to 1μs difference, because macOS <10.13 cannot restore
-		// with nanosecond precision and the current version of Go (1.9.2)
-		// does not yet support the new syscall. (#1087)
-		mt1 := fi1.ModTime()
-		mt2 := fi2.ModTime()
-		usecDiff := (mt1.Nanosecond()-mt2.Nanosecond())/1000 + (mt1.Second()-mt2.Second())*1000000
-		same = usecDiff <= 1 && usecDiff >= -1
-	}
-	return same
+	return fi1.ModTime().Equal(fi2.ModTime())
 }
 
-// directoriesEqualContents checks if both directories contain exactly the same
-// contents.
-func directoriesEqualContents(dir1, dir2 string) bool {
+// directoriesContentsDiff returns a diff between both directories. If these
+// contain exactly the same contents, then the diff is an empty string.
+func directoriesContentsDiff(dir1, dir2 string) string {
+	var out bytes.Buffer
 	ch1 := walkDir(dir1)
 	ch2 := walkDir(dir2)
 
-	changes := false
-
 	var a, b *dirEntry
 	for {
 		var ok bool
@@ -116,36 +106,27 @@ func directoriesEqualContents(dir1, dir2 string) bool {
 		}
 
 		if ch1 == nil {
-			fmt.Printf("+%v\n", b.path)
-			changes = true
+			fmt.Fprintf(&out, "+%v\n", b.path)
 		} else if ch2 == nil {
-			fmt.Printf("-%v\n", a.path)
-			changes = true
-		} else if !a.equals(b) {
+			fmt.Fprintf(&out, "-%v\n", a.path)
+		} else if !a.equals(&out, b) {
 			if a.path < b.path {
-				fmt.Printf("-%v\n", a.path)
-				changes = true
+				fmt.Fprintf(&out, "-%v\n", a.path)
 				a = nil
 				continue
 			} else if a.path > b.path {
-				fmt.Printf("+%v\n", b.path)
-				changes = true
+				fmt.Fprintf(&out, "+%v\n", b.path)
 				b = nil
 				continue
 			} else {
-				fmt.Printf("%%%v\n", a.path)
-				changes = true
+				fmt.Fprintf(&out, "%%%v\n", a.path)
 			}
 		}
 
 		a, b = nil, nil
 	}
 
-	if changes {
-		return false
-	}
-
-	return true
+	return out.String()
 }
 
 type dirStat struct {

cmd/restic/integration_helpers_unix_test.go

@@ -4,25 +4,26 @@ package main
 
 import (
 	"fmt"
+	"io"
 	"io/ioutil"
 	"os"
 	"path/filepath"
 	"syscall"
 )
 
-func (e *dirEntry) equals(other *dirEntry) bool {
+func (e *dirEntry) equals(out io.Writer, other *dirEntry) bool {
 	if e.path != other.path {
-		fmt.Fprintf(os.Stderr, "%v: path does not match (%v != %v)\n", e.path, e.path, other.path)
+		fmt.Fprintf(out, "%v: path does not match (%v != %v)\n", e.path, e.path, other.path)
 		return false
 	}
 
 	if e.fi.Mode() != other.fi.Mode() {
-		fmt.Fprintf(os.Stderr, "%v: mode does not match (%v != %v)\n", e.path, e.fi.Mode(), other.fi.Mode())
+		fmt.Fprintf(out, "%v: mode does not match (%v != %v)\n", e.path, e.fi.Mode(), other.fi.Mode())
 		return false
 	}
 
 	if !sameModTime(e.fi, other.fi) {
-		fmt.Fprintf(os.Stderr, "%v: ModTime does not match (%v != %v)\n", e.path, e.fi.ModTime(), other.fi.ModTime())
+		fmt.Fprintf(out, "%v: ModTime does not match (%v != %v)\n", e.path, e.fi.ModTime(), other.fi.ModTime())
 		return false
 	}
 
@@ -30,17 +31,17 @@ func (e *dirEntry) equals(other *dirEntry) bool {
 	stat2, _ := other.fi.Sys().(*syscall.Stat_t)
 
 	if stat.Uid != stat2.Uid {
-		fmt.Fprintf(os.Stderr, "%v: UID does not match (%v != %v)\n", e.path, stat.Uid, stat2.Uid)
+		fmt.Fprintf(out, "%v: UID does not match (%v != %v)\n", e.path, stat.Uid, stat2.Uid)
 		return false
 	}
 
 	if stat.Gid != stat2.Gid {
-		fmt.Fprintf(os.Stderr, "%v: GID does not match (%v != %v)\n", e.path, stat.Gid, stat2.Gid)
+		fmt.Fprintf(out, "%v: GID does not match (%v != %v)\n", e.path, stat.Gid, stat2.Gid)
 		return false
 	}
 
 	if stat.Nlink != stat2.Nlink {
-		fmt.Fprintf(os.Stderr, "%v: Number of links do not match (%v != %v)\n", e.path, stat.Nlink, stat2.Nlink)
+		fmt.Fprintf(out, "%v: Number of links do not match (%v != %v)\n", e.path, stat.Nlink, stat2.Nlink)
 		return false
 	}
 

cmd/restic/integration_helpers_windows_test.go

@@ -4,23 +4,24 @@ package main
 
 import (
 	"fmt"
+	"io"
 	"io/ioutil"
 	"os"
 )
 
-func (e *dirEntry) equals(other *dirEntry) bool {
+func (e *dirEntry) equals(out io.Writer, other *dirEntry) bool {
 	if e.path != other.path {
-		fmt.Fprintf(os.Stderr, "%v: path does not match (%v != %v)\n", e.path, e.path, other.path)
+		fmt.Fprintf(out, "%v: path does not match (%v != %v)\n", e.path, e.path, other.path)
 		return false
 	}
 
 	if e.fi.Mode() != other.fi.Mode() {
-		fmt.Fprintf(os.Stderr, "%v: mode does not match (%v != %v)\n", e.path, e.fi.Mode(), other.fi.Mode())
+		fmt.Fprintf(out, "%v: mode does not match (%v != %v)\n", e.path, e.fi.Mode(), other.fi.Mode())
 		return false
 	}
 
 	if !sameModTime(e.fi, other.fi) {
-		fmt.Fprintf(os.Stderr, "%v: ModTime does not match (%v != %v)\n", e.path, e.fi.ModTime(), other.fi.ModTime())
+		fmt.Fprintf(out, "%v: ModTime does not match (%v != %v)\n", e.path, e.fi.ModTime(), other.fi.ModTime())
 		return false
 	}
 

cmd/restic/integration_test.go

@@ -13,6 +13,7 @@ import (
 	"os"
 	"path/filepath"
 	"regexp"
+	"runtime"
 	"strings"
 	"syscall"
 	"testing"
@@ -50,11 +51,11 @@ func testRunInit(t testing.TB, opts GlobalOptions) {
 	restic.TestDisableCheckPolynomial(t)
 	restic.TestSetLockTimeout(t, 0)
 
-	rtest.OK(t, runInit(opts, nil))
+	rtest.OK(t, runInit(InitOptions{}, opts, nil))
 	t.Logf("repository initialized at %v", opts.Repo)
 }
 
-func testRunBackup(t testing.TB, dir string, target []string, opts BackupOptions, gopts GlobalOptions) {
+func testRunBackupAssumeFailure(t testing.TB, dir string, target []string, opts BackupOptions, gopts GlobalOptions) error {
 	ctx, cancel := context.WithCancel(gopts.ctx)
 	defer cancel()
 
@@ -69,7 +70,7 @@ func testRunBackup(t testing.TB, dir string, target []string, opts BackupOptions
 		defer cleanup()
 	}
 
-	rtest.OK(t, runBackup(opts, gopts, term, target))
+	backupErr := runBackup(opts, gopts, term, target)
 
 	cancel()
 
@@ -77,6 +78,13 @@ func testRunBackup(t testing.TB, dir string, target []string, opts BackupOptions
 	if err != nil {
 		t.Fatal(err)
 	}
+
+	return backupErr
+}
+
+func testRunBackup(t testing.TB, dir string, target []string, opts BackupOptions, gopts GlobalOptions) {
+	err := testRunBackupAssumeFailure(t, dir, target, opts, gopts)
+	rtest.Assert(t, err == nil, "Error while backing up")
 }
 
 func testRunList(t testing.TB, tpe string, opts GlobalOptions) restic.IDs {
@@ -143,6 +151,21 @@ func testRunCheckOutput(gopts GlobalOptions) (string, error) {
 	}
 
 	err := runCheck(opts, gopts, nil)
+	return buf.String(), err
+}
+
+func testRunDiffOutput(gopts GlobalOptions, firstSnapshotID string, secondSnapshotID string) (string, error) {
+	buf := bytes.NewBuffer(nil)
+
+	globalOptions.stdout = buf
+	defer func() {
+		globalOptions.stdout = os.Stdout
+	}()
+
+	opts := DiffOptions{
+		ShowMetadata: false,
+	}
+	err := runDiff(opts, gopts, []string{firstSnapshotID, secondSnapshotID})
 	return string(buf.Bytes()), err
 }
 
@@ -169,7 +192,7 @@ func testRunLs(t testing.TB, gopts GlobalOptions, snapshotID string) []string {
 
 	rtest.OK(t, runLs(opts, gopts, []string{snapshotID}))
 
-	return strings.Split(string(buf.Bytes()), "\n")
+	return strings.Split(buf.String(), "\n")
 }
 
 func testRunFind(t testing.TB, wantJSON bool, gopts GlobalOptions, pattern string) []byte {
@@ -245,29 +268,24 @@ func testRunForgetJSON(t testing.TB, gopts GlobalOptions, args ...string) {
 		"Expected 1 snapshot to be kept, got %v", len(forgets[0].Keep))
 	rtest.Assert(t, len(forgets[0].Remove) == 2,
 		"Expected 2 snapshots to be removed, got %v", len(forgets[0].Remove))
-	return
 }
 
 func testRunPrune(t testing.TB, gopts GlobalOptions) {
 	rtest.OK(t, runPrune(gopts))
 }
 
+func testSetupBackupData(t testing.TB, env *testEnvironment) string {
+	datafile := filepath.Join("testdata", "backup-data.tar.gz")
+	testRunInit(t, env.gopts)
+	rtest.SetupTarTestFixture(t, env.testdata, datafile)
+	return datafile
+}
+
 func TestBackup(t *testing.T) {
 	env, cleanup := withTestEnvironment(t)
 	defer cleanup()
 
-	datafile := filepath.Join("testdata", "backup-data.tar.gz")
-	fd, err := os.Open(datafile)
-	if os.IsNotExist(errors.Cause(err)) {
-		t.Skipf("unable to find data file %q, skipping", datafile)
-		return
-	}
-	rtest.OK(t, err)
-	rtest.OK(t, fd.Close())
-
-	testRunInit(t, env.gopts)
-
-	rtest.SetupTarTestFixture(t, env.testdata, datafile)
+	testSetupBackupData(t, env)
 	opts := BackupOptions{}
 
 	// first backup
@@ -309,9 +327,9 @@ func TestBackup(t *testing.T) {
 	for i, snapshotID := range snapshotIDs {
 		restoredir := filepath.Join(env.base, fmt.Sprintf("restore%d", i))
 		t.Logf("restoring snapshot %v to %v", snapshotID.Str(), restoredir)
-		testRunRestore(t, env.gopts, restoredir, snapshotIDs[0])
-		rtest.Assert(t, directoriesEqualContents(env.testdata, filepath.Join(restoredir, "testdata")),
-			"directories are not equal")
+		testRunRestore(t, env.gopts, restoredir, snapshotID)
+		diff := directoriesContentsDiff(env.testdata, filepath.Join(restoredir, "testdata"))
+		rtest.Assert(t, diff == "", "directories are not equal: %v", diff)
 	}
 
 	testRunCheck(t, env.gopts)
@@ -321,18 +339,7 @@ func TestBackupNonExistingFile(t *testing.T) {
 	env, cleanup := withTestEnvironment(t)
 	defer cleanup()
 
-	datafile := filepath.Join("testdata", "backup-data.tar.gz")
-	fd, err := os.Open(datafile)
-	if os.IsNotExist(errors.Cause(err)) {
-		t.Skipf("unable to find data file %q, skipping", datafile)
-		return
-	}
-	rtest.OK(t, err)
-	rtest.OK(t, fd.Close())
-
-	rtest.SetupTarTestFixture(t, env.testdata, datafile)
-
-	testRunInit(t, env.gopts)
+	testSetupBackupData(t, env)
 	globalOptions.stderr = ioutil.Discard
 	defer func() {
 		globalOptions.stderr = os.Stderr
@@ -351,6 +358,58 @@ func TestBackupNonExistingFile(t *testing.T) {
 	testRunBackup(t, "", dirs, opts, env.gopts)
 }
 
+func removeDataPacksExcept(gopts GlobalOptions, t *testing.T, keep restic.IDSet) {
+	r, err := OpenRepository(gopts)
+	rtest.OK(t, err)
+
+	// Get all tree packs
+	rtest.OK(t, r.LoadIndex(gopts.ctx))
+	treePacks := restic.NewIDSet()
+	for _, idx := range r.Index().(*repository.MasterIndex).All() {
+		for _, id := range idx.TreePacks() {
+			treePacks.Insert(id)
+		}
+	}
+
+	// remove all packs containing data blobs
+	rtest.OK(t, r.List(gopts.ctx, restic.PackFile, func(id restic.ID, size int64) error {
+		if treePacks.Has(id) || keep.Has(id) {
+			return nil
+		}
+		return r.Backend().Remove(gopts.ctx, restic.Handle{Type: restic.PackFile, Name: id.String()})
+	}))
+}
+
+func TestBackupSelfHealing(t *testing.T) {
+	env, cleanup := withTestEnvironment(t)
+	defer cleanup()
+
+	testRunInit(t, env.gopts)
+
+	p := filepath.Join(env.testdata, "test/test")
+	rtest.OK(t, os.MkdirAll(filepath.Dir(p), 0755))
+	rtest.OK(t, appendRandomData(p, 5))
+
+	opts := BackupOptions{}
+
+	testRunBackup(t, filepath.Dir(env.testdata), []string{filepath.Base(env.testdata)}, opts, env.gopts)
+	testRunCheck(t, env.gopts)
+
+	// remove all data packs
+	removeDataPacksExcept(env.gopts, t, restic.NewIDSet())
+
+	testRunRebuildIndex(t, env.gopts)
+	// now the repo is also missing the data blob in the index; check should report this
+	rtest.Assert(t, runCheck(CheckOptions{}, env.gopts, nil) != nil,
+		"check should have reported an error")
+
+	// second backup should report an error but "heal" this situation
+	err := testRunBackupAssumeFailure(t, filepath.Dir(env.testdata), []string{filepath.Base(env.testdata)}, opts, env.gopts)
+	rtest.Assert(t, err != nil,
+		"backup should have reported an error")
+	testRunCheck(t, env.gopts)
+}
+
 func includes(haystack []string, needle string) bool {
 	for _, s := range haystack {
 		if s == needle {
@@ -405,7 +464,7 @@ func TestBackupExclude(t *testing.T) {
 		f, err := os.Create(fp)
 		rtest.OK(t, err)
 
-		fmt.Fprintf(f, filename)
+		fmt.Fprint(f, filename)
 		rtest.OK(t, f.Close())
 	}
 
@@ -436,6 +495,32 @@ func TestBackupExclude(t *testing.T) {
 		"expected file %q not in first snapshot, but it's included", "passwords.txt")
 }
 
+func TestBackupErrors(t *testing.T) {
+	if runtime.GOOS == "windows" {
+		return
+	}
+	env, cleanup := withTestEnvironment(t)
+	defer cleanup()
+
+	testSetupBackupData(t, env)
+
+	// Assume failure
+	inaccessibleFile := filepath.Join(env.testdata, "0", "0", "9", "0")
+	os.Chmod(inaccessibleFile, 0000)
+	defer func() {
+		os.Chmod(inaccessibleFile, 0644)
+	}()
+	opts := BackupOptions{}
+	gopts := env.gopts
+	gopts.stderr = ioutil.Discard
+	err := testRunBackupAssumeFailure(t, filepath.Dir(env.testdata), []string{"testdata"}, opts, gopts)
+	rtest.Assert(t, err != nil, "Assumed failure, but no error occured.")
+	rtest.Assert(t, err == ErrInvalidSourceData, "Wrong error returned")
+	snapshotIDs := testRunList(t, "snapshots", env.gopts)
+	rtest.Assert(t, len(snapshotIDs) == 1,
+		"expected one snapshot, got %v", snapshotIDs)
+}
+
 const (
 	incrementalFirstWrite  = 10 * 1042 * 1024
 	incrementalSecondWrite = 1 * 1042 * 1024
@@ -506,10 +591,7 @@ func TestBackupTags(t *testing.T) {
 	env, cleanup := withTestEnvironment(t)
 	defer cleanup()
 
-	datafile := filepath.Join("testdata", "backup-data.tar.gz")
-	testRunInit(t, env.gopts)
-	rtest.SetupTarTestFixture(t, env.testdata, datafile)
-
+	testSetupBackupData(t, env)
 	opts := BackupOptions{}
 
 	testRunBackup(t, "", []string{env.testdata}, opts, env.gopts)
@@ -532,6 +614,153 @@ func TestBackupTags(t *testing.T) {
 		"expected parent to be %v, got %v", parent.ID, newest.Parent)
 }
 
+func testRunCopy(t testing.TB, srcGopts GlobalOptions, dstGopts GlobalOptions) {
+	copyOpts := CopyOptions{
+		secondaryRepoOptions: secondaryRepoOptions{
+			Repo:     dstGopts.Repo,
+			password: dstGopts.password,
+		},
+	}
+
+	rtest.OK(t, runCopy(copyOpts, srcGopts, nil))
+}
+
+func TestCopy(t *testing.T) {
+	env, cleanup := withTestEnvironment(t)
+	defer cleanup()
+	env2, cleanup2 := withTestEnvironment(t)
+	defer cleanup2()
+
+	testSetupBackupData(t, env)
+	opts := BackupOptions{}
+	testRunBackup(t, "", []string{filepath.Join(env.testdata, "0", "0", "9")}, opts, env.gopts)
+	testRunBackup(t, "", []string{filepath.Join(env.testdata, "0", "0", "9", "2")}, opts, env.gopts)
+	testRunBackup(t, "", []string{filepath.Join(env.testdata, "0", "0", "9", "3")}, opts, env.gopts)
+	testRunCheck(t, env.gopts)
+
+	testRunInit(t, env2.gopts)
+	testRunCopy(t, env.gopts, env2.gopts)
+
+	snapshotIDs := testRunList(t, "snapshots", env.gopts)
+	copiedSnapshotIDs := testRunList(t, "snapshots", env2.gopts)
+
+	// Check that the copies size seems reasonable
+	rtest.Assert(t, len(snapshotIDs) == len(copiedSnapshotIDs), "expected %v snapshots, found %v",
+		len(snapshotIDs), len(copiedSnapshotIDs))
+	stat := dirStats(env.repo)
+	stat2 := dirStats(env2.repo)
+	sizeDiff := int64(stat.size) - int64(stat2.size)
+	if sizeDiff < 0 {
+		sizeDiff = -sizeDiff
+	}
+	rtest.Assert(t, sizeDiff < int64(stat.size)/50, "expected less than 2%% size difference: %v vs. %v",
+		stat.size, stat2.size)
+
+	// Check integrity of the copy
+	testRunCheck(t, env2.gopts)
+
+	// Check that the copied snapshots have the same tree contents as the old ones (= identical tree hash)
+	origRestores := make(map[string]struct{})
+	for i, snapshotID := range snapshotIDs {
+		restoredir := filepath.Join(env.base, fmt.Sprintf("restore%d", i))
+		origRestores[restoredir] = struct{}{}
+		testRunRestore(t, env.gopts, restoredir, snapshotID)
+	}
+	for i, snapshotID := range copiedSnapshotIDs {
+		restoredir := filepath.Join(env2.base, fmt.Sprintf("restore%d", i))
+		testRunRestore(t, env2.gopts, restoredir, snapshotID)
+		foundMatch := false
+		for cmpdir := range origRestores {
+			diff := directoriesContentsDiff(restoredir, cmpdir)
+			if diff == "" {
+				delete(origRestores, cmpdir)
+				foundMatch = true
+			}
+		}
+
+		rtest.Assert(t, foundMatch, "found no counterpart for snapshot %v", snapshotID)
+	}
+
+	rtest.Assert(t, len(origRestores) == 0, "found not copied snapshots")
+}
+
+func TestCopyIncremental(t *testing.T) {
+	env, cleanup := withTestEnvironment(t)
+	defer cleanup()
+	env2, cleanup2 := withTestEnvironment(t)
+	defer cleanup2()
+
+	testSetupBackupData(t, env)
+	opts := BackupOptions{}
+	testRunBackup(t, "", []string{filepath.Join(env.testdata, "0", "0", "9")}, opts, env.gopts)
+	testRunBackup(t, "", []string{filepath.Join(env.testdata, "0", "0", "9", "2")}, opts, env.gopts)
+	testRunCheck(t, env.gopts)
+
+	testRunInit(t, env2.gopts)
+	testRunCopy(t, env.gopts, env2.gopts)
+
+	snapshotIDs := testRunList(t, "snapshots", env.gopts)
+	copiedSnapshotIDs := testRunList(t, "snapshots", env2.gopts)
+
+	// Check that the copies size seems reasonable
+	testRunCheck(t, env2.gopts)
+	rtest.Assert(t, len(snapshotIDs) == len(copiedSnapshotIDs), "expected %v snapshots, found %v",
+		len(snapshotIDs), len(copiedSnapshotIDs))
+
+	// check that no snapshots are copied, as there are no new ones
+	testRunCopy(t, env.gopts, env2.gopts)
+	testRunCheck(t, env2.gopts)
+	copiedSnapshotIDs = testRunList(t, "snapshots", env2.gopts)
+	rtest.Assert(t, len(snapshotIDs) == len(copiedSnapshotIDs), "still expected %v snapshots, found %v",
+		len(snapshotIDs), len(copiedSnapshotIDs))
+
+	// check that only new snapshots are copied
+	testRunBackup(t, "", []string{filepath.Join(env.testdata, "0", "0", "9", "3")}, opts, env.gopts)
+	testRunCopy(t, env.gopts, env2.gopts)
+	testRunCheck(t, env2.gopts)
+	snapshotIDs = testRunList(t, "snapshots", env.gopts)
+	copiedSnapshotIDs = testRunList(t, "snapshots", env2.gopts)
+	rtest.Assert(t, len(snapshotIDs) == len(copiedSnapshotIDs), "still expected %v snapshots, found %v",
+		len(snapshotIDs), len(copiedSnapshotIDs))
+
+	// also test the reverse direction
+	testRunCopy(t, env2.gopts, env.gopts)
+	testRunCheck(t, env.gopts)
+	snapshotIDs = testRunList(t, "snapshots", env.gopts)
+	rtest.Assert(t, len(snapshotIDs) == len(copiedSnapshotIDs), "still expected %v snapshots, found %v",
+		len(copiedSnapshotIDs), len(snapshotIDs))
+}
+
+func TestInitCopyChunkerParams(t *testing.T) {
+	env, cleanup := withTestEnvironment(t)
+	defer cleanup()
+	env2, cleanup2 := withTestEnvironment(t)
+	defer cleanup2()
+
+	testRunInit(t, env2.gopts)
+
+	initOpts := InitOptions{
+		secondaryRepoOptions: secondaryRepoOptions{
+			Repo:     env2.gopts.Repo,
+			password: env2.gopts.password,
+		},
+	}
+	rtest.Assert(t, runInit(initOpts, env.gopts, nil) != nil, "expected invalid init options to fail")
+
+	initOpts.CopyChunkerParameters = true
+	rtest.OK(t, runInit(initOpts, env.gopts, nil))
+
+	repo, err := OpenRepository(env.gopts)
+	rtest.OK(t, err)
+
+	otherRepo, err := OpenRepository(env2.gopts)
+	rtest.OK(t, err)
+
+	rtest.Assert(t, repo.Config().ChunkerPolynomial == otherRepo.Config().ChunkerPolynomial,
+		"expected equal chunker polynomials, got %v expected %v", repo.Config().ChunkerPolynomial,
+		otherRepo.Config().ChunkerPolynomial)
+}
+
 func testRunTag(t testing.TB, opts TagOptions, gopts GlobalOptions) {
 	rtest.OK(t, runTag(opts, gopts, []string{}))
 }
@@ -540,10 +769,7 @@ func TestTag(t *testing.T) {
 	env, cleanup := withTestEnvironment(t)
 	defer cleanup()
 
-	datafile := filepath.Join("testdata", "backup-data.tar.gz")
-	testRunInit(t, env.gopts)
-	rtest.SetupTarTestFixture(t, env.testdata, datafile)
-
+	testSetupBackupData(t, env)
 	testRunBackup(t, "", []string{env.testdata}, BackupOptions{}, env.gopts)
 	testRunCheck(t, env.gopts)
 	newest, _ := testRunSnapshots(t, env.gopts)
@@ -639,6 +865,28 @@ func testRunKeyAddNewKey(t testing.TB, newPassword string, gopts GlobalOptions)
 	rtest.OK(t, runKey(gopts, []string{"add"}))
 }
 
+func testRunKeyAddNewKeyUserHost(t testing.TB, gopts GlobalOptions) {
+	testKeyNewPassword = "john's geheimnis"
+	defer func() {
+		testKeyNewPassword = ""
+		keyUsername = ""
+		keyHostname = ""
+	}()
+
+	cmdKey.Flags().Parse([]string{"--user=john", "--host=example.com"})
+
+	t.Log("adding key for john@example.com")
+	rtest.OK(t, runKey(gopts, []string{"add"}))
+
+	repo, err := OpenRepository(gopts)
+	rtest.OK(t, err)
+	key, err := repository.SearchKey(gopts.ctx, repo, testKeyNewPassword, 1, "")
+	rtest.OK(t, err)
+
+	rtest.Equals(t, "john", key.Username)
+	rtest.Equals(t, "example.com", key.Hostname)
+}
+
 func testRunKeyPasswd(t testing.TB, newPassword string, gopts GlobalOptions) {
 	testKeyNewPassword = newPassword
 	defer func() {
@@ -681,6 +929,8 @@ func TestKeyAddRemove(t *testing.T) {
 	t.Logf("testing access with last password %q\n", env.gopts.password)
 	rtest.OK(t, runKey(env.gopts, []string{"list"}))
 	testRunCheck(t, env.gopts)
+
+	testRunKeyAddNewKeyUserHost(t, env.gopts)
 }
 
 func testFileSize(filename string, size int64) error {
@@ -767,8 +1017,8 @@ func TestRestore(t *testing.T) {
 	restoredir := filepath.Join(env.base, "restore")
 	testRunRestoreLatest(t, env.gopts, restoredir, nil, nil)
 
-	rtest.Assert(t, directoriesEqualContents(env.testdata, filepath.Join(restoredir, filepath.Base(env.testdata))),
-		"directories are not equal")
+	diff := directoriesContentsDiff(env.testdata, filepath.Join(restoredir, filepath.Base(env.testdata)))
+	rtest.Assert(t, diff == "", "directories are not equal %v", diff)
 }
 
 func TestRestoreLatest(t *testing.T) {
@@ -901,14 +1151,14 @@ func TestRestoreNoMetadataOnIgnoredIntermediateDirs(t *testing.T) {
 	testRunRestoreIncludes(t, env.gopts, filepath.Join(env.base, "restore0"), snapshotID, []string{"*.ext"})
 
 	f1 := filepath.Join(env.base, "restore0", "testdata", "subdir1", "subdir2")
-	fi, err := os.Stat(f1)
+	_, err := os.Stat(f1)
 	rtest.OK(t, err)
 
 	// restore with filter "*", this should restore meta data on everything.
 	testRunRestoreIncludes(t, env.gopts, filepath.Join(env.base, "restore1"), snapshotID, []string{"*"})
 
 	f2 := filepath.Join(env.base, "restore1", "testdata", "subdir1", "subdir2")
-	fi, err = os.Stat(f2)
+	fi, err := os.Stat(f2)
 	rtest.OK(t, err)
 
 	rtest.Assert(t, fi.ModTime() == time.Unix(0, 0),
@@ -919,10 +1169,7 @@ func TestFind(t *testing.T) {
 	env, cleanup := withTestEnvironment(t)
 	defer cleanup()
 
-	datafile := filepath.Join("testdata", "backup-data.tar.gz")
-	testRunInit(t, env.gopts)
-	rtest.SetupTarTestFixture(t, env.testdata, datafile)
-
+	datafile := testSetupBackupData(t, env)
 	opts := BackupOptions{}
 
 	testRunBackup(t, "", []string{env.testdata}, opts, env.gopts)
@@ -959,10 +1206,7 @@ func TestFindJSON(t *testing.T) {
 	env, cleanup := withTestEnvironment(t)
 	defer cleanup()
 
-	datafile := filepath.Join("testdata", "backup-data.tar.gz")
-	testRunInit(t, env.gopts)
-	rtest.SetupTarTestFixture(t, env.testdata, datafile)
-
+	datafile := testSetupBackupData(t, env)
 	opts := BackupOptions{}
 
 	testRunBackup(t, "", []string{env.testdata}, opts, env.gopts)
@@ -1023,6 +1267,37 @@ func TestRebuildIndexAlwaysFull(t *testing.T) {
 	TestRebuildIndex(t)
 }
 
+type appendOnlyBackend struct {
+	restic.Backend
+}
+
+// called via repo.Backend().Remove()
+func (b *appendOnlyBackend) Remove(ctx context.Context, h restic.Handle) error {
+	return errors.Errorf("Failed to remove %v", h)
+}
+
+func TestRebuildIndexFailsOnAppendOnly(t *testing.T) {
+	env, cleanup := withTestEnvironment(t)
+	defer cleanup()
+
+	datafile := filepath.Join("..", "..", "internal", "checker", "testdata", "duplicate-packs-in-index-test-repo.tar.gz")
+	rtest.SetupTarTestFixture(t, env.base, datafile)
+
+	globalOptions.stdout = ioutil.Discard
+	defer func() {
+		globalOptions.stdout = os.Stdout
+	}()
+
+	env.gopts.backendTestHook = func(r restic.Backend) (restic.Backend, error) {
+		return &appendOnlyBackend{r}, nil
+	}
+	err := runRebuildIndex(env.gopts)
+	if err == nil {
+		t.Error("expected rebuildIndex to fail")
+	}
+	t.Log(err)
+}
+
 func TestCheckRestoreNoLock(t *testing.T) {
 	env, cleanup := withTestEnvironment(t)
 	defer cleanup()
@@ -1054,18 +1329,7 @@ func TestPrune(t *testing.T) {
 	env, cleanup := withTestEnvironment(t)
 	defer cleanup()
 
-	datafile := filepath.Join("testdata", "backup-data.tar.gz")
-	fd, err := os.Open(datafile)
-	if os.IsNotExist(errors.Cause(err)) {
-		t.Skipf("unable to find data file %q, skipping", datafile)
-		return
-	}
-	rtest.OK(t, err)
-	rtest.OK(t, fd.Close())
-
-	testRunInit(t, env.gopts)
-
-	rtest.SetupTarTestFixture(t, env.testdata, datafile)
+	testSetupBackupData(t, env)
 	opts := BackupOptions{}
 
 	testRunBackup(t, "", []string{filepath.Join(env.testdata, "0", "0", "9")}, opts, env.gopts)
@@ -1086,6 +1350,58 @@ func TestPrune(t *testing.T) {
 	testRunCheck(t, env.gopts)
 }
 
+func listPacks(gopts GlobalOptions, t *testing.T) restic.IDSet {
+	r, err := OpenRepository(gopts)
+	rtest.OK(t, err)
+
+	packs := restic.NewIDSet()
+
+	rtest.OK(t, r.List(gopts.ctx, restic.PackFile, func(id restic.ID, size int64) error {
+		packs.Insert(id)
+		return nil
+	}))
+	return packs
+}
+
+func TestPruneWithDamagedRepository(t *testing.T) {
+	env, cleanup := withTestEnvironment(t)
+	defer cleanup()
+
+	datafile := filepath.Join("testdata", "backup-data.tar.gz")
+	testRunInit(t, env.gopts)
+
+	rtest.SetupTarTestFixture(t, env.testdata, datafile)
+	opts := BackupOptions{}
+
+	// create and delete snapshot to create unused blobs
+	testRunBackup(t, "", []string{filepath.Join(env.testdata, "0", "0", "9", "2")}, opts, env.gopts)
+	firstSnapshot := testRunList(t, "snapshots", env.gopts)
+	rtest.Assert(t, len(firstSnapshot) == 1,
+		"expected one snapshot, got %v", firstSnapshot)
+	testRunForget(t, env.gopts, firstSnapshot[0].String())
+
+	oldPacks := listPacks(env.gopts, t)
+
+	// create new snapshot, but lose all data
+	testRunBackup(t, "", []string{filepath.Join(env.testdata, "0", "0", "9", "3")}, opts, env.gopts)
+	snapshotIDs := testRunList(t, "snapshots", env.gopts)
+
+	removeDataPacksExcept(env.gopts, t, oldPacks)
+
+	rtest.Assert(t, len(snapshotIDs) == 1,
+		"expected one snapshot, got %v", snapshotIDs)
+
+	// prune should fail
+	err := runPrune(env.gopts)
+	if err == nil {
+		t.Fatalf("expected prune to fail")
+	}
+	if !strings.Contains(err.Error(), "blobs seem to be missing") {
+		t.Fatalf("did not find hint for missing blobs")
+	}
+	t.Log(err)
+}
+
 func TestHardLink(t *testing.T) {
 	// this test assumes a test set with a single directory containing hard linked files
 	env, cleanup := withTestEnvironment(t)
@@ -1120,9 +1436,9 @@ func TestHardLink(t *testing.T) {
 	for i, snapshotID := range snapshotIDs {
 		restoredir := filepath.Join(env.base, fmt.Sprintf("restore%d", i))
 		t.Logf("restoring snapshot %v to %v", snapshotID.Str(), restoredir)
-		testRunRestore(t, env.gopts, restoredir, snapshotIDs[0])
-		rtest.Assert(t, directoriesEqualContents(env.testdata, filepath.Join(restoredir, "testdata")),
-			"directories are not equal")
+		testRunRestore(t, env.gopts, restoredir, snapshotID)
+		diff := directoriesContentsDiff(env.testdata, filepath.Join(restoredir, "testdata"))
+		rtest.Assert(t, diff == "", "directories are not equal %v", diff)
 
 		linkResults := createFileSetPerHardlink(filepath.Join(restoredir, "testdata"))
 		rtest.Assert(t, linksEqual(linkTests, linkResults),
@@ -1147,11 +1463,7 @@ func linksEqual(source, dest map[uint64][]string) bool {
 		}
 	}
 
-	if len(dest) != 0 {
-		return false
-	}
-
-	return true
+	return len(dest) == 0
 }
 
 func linkEqual(source, dest []string) bool {
@@ -1188,18 +1500,7 @@ func TestQuietBackup(t *testing.T) {
 	env, cleanup := withTestEnvironment(t)
 	defer cleanup()
 
-	datafile := filepath.Join("testdata", "backup-data.tar.gz")
-	fd, err := os.Open(datafile)
-	if os.IsNotExist(errors.Cause(err)) {
-		t.Skipf("unable to find data file %q, skipping", datafile)
-		return
-	}
-	rtest.OK(t, err)
-	rtest.OK(t, fd.Close())
-
-	testRunInit(t, env.gopts)
-
-	rtest.SetupTarTestFixture(t, env.testdata, datafile)
+	testSetupBackupData(t, env)
 	opts := BackupOptions{}
 
 	env.gopts.Quiet = false
@@ -1218,3 +1519,91 @@ func TestQuietBackup(t *testing.T) {
 
 	testRunCheck(t, env.gopts)
 }
+
+func copyFile(dst string, src string) error {
+	srcFile, err := os.Open(src)
+	if err != nil {
+		return err
+	}
+	defer srcFile.Close()
+
+	dstFile, err := os.Create(dst)
+	if err != nil {
+		return err
+	}
+	defer dstFile.Close()
+
+	_, err = io.Copy(dstFile, srcFile)
+	return err
+}
+
+var diffOutputRegexPatterns = []string{
+	"-.+modfile",
+	"M.+modfile1",
+	"\\+.+modfile2",
+	"\\+.+modfile3",
+	"\\+.+modfile4",
+	"-.+submoddir",
+	"-.+submoddir.subsubmoddir",
+	"\\+.+submoddir2",
+	"\\+.+submoddir2.subsubmoddir",
+	"Files: +2 new, +1 removed, +1 changed",
+	"Dirs: +3 new, +2 removed",
+	"Data Blobs: +2 new, +1 removed",
+	"Added: +7[0-9]{2}\\.[0-9]{3} KiB",
+	"Removed: +2[0-9]{2}\\.[0-9]{3} KiB",
+}
+
+func TestDiff(t *testing.T) {
+	env, cleanup := withTestEnvironment(t)
+	defer cleanup()
+
+	testRunInit(t, env.gopts)
+
+	datadir := filepath.Join(env.base, "testdata")
+	testdir := filepath.Join(datadir, "testdir")
+	subtestdir := filepath.Join(testdir, "subtestdir")
+	testfile := filepath.Join(testdir, "testfile")
+
+	rtest.OK(t, os.Mkdir(testdir, 0755))
+	rtest.OK(t, os.Mkdir(subtestdir, 0755))
+	rtest.OK(t, appendRandomData(testfile, 256*1024))
+
+	moddir := filepath.Join(datadir, "moddir")
+	submoddir := filepath.Join(moddir, "submoddir")
+	subsubmoddir := filepath.Join(submoddir, "subsubmoddir")
+	modfile := filepath.Join(moddir, "modfile")
+	rtest.OK(t, os.Mkdir(moddir, 0755))
+	rtest.OK(t, os.Mkdir(submoddir, 0755))
+	rtest.OK(t, os.Mkdir(subsubmoddir, 0755))
+	rtest.OK(t, copyFile(modfile, testfile))
+	rtest.OK(t, appendRandomData(modfile+"1", 256*1024))
+
+	snapshots := make(map[string]struct{})
+	opts := BackupOptions{}
+	testRunBackup(t, "", []string{datadir}, opts, env.gopts)
+	snapshots, firstSnapshotID := lastSnapshot(snapshots, loadSnapshotMap(t, env.gopts))
+
+	rtest.OK(t, os.Rename(modfile, modfile+"3"))
+	rtest.OK(t, os.Rename(submoddir, submoddir+"2"))
+	rtest.OK(t, appendRandomData(modfile+"1", 256*1024))
+	rtest.OK(t, appendRandomData(modfile+"2", 256*1024))
+	rtest.OK(t, os.Mkdir(modfile+"4", 0755))
+
+	testRunBackup(t, "", []string{datadir}, opts, env.gopts)
+	snapshots, secondSnapshotID := lastSnapshot(snapshots, loadSnapshotMap(t, env.gopts))
+
+	_, err := testRunDiffOutput(env.gopts, "", secondSnapshotID)
+	rtest.Assert(t, err != nil, "expected error on invalid snapshot id")
+
+	out, err := testRunDiffOutput(env.gopts, firstSnapshotID, secondSnapshotID)
+	if err != nil {
+		t.Fatalf("expected no error from diff for test repository, got %v", err)
+	}
+
+	for _, pattern := range diffOutputRegexPatterns {
+		r, err := regexp.Compile(pattern)
+		rtest.Assert(t, err == nil, "failed to compile regexp %v", pattern)
+		rtest.Assert(t, r.MatchString(out), "expected pattern %v in output, got\n%v", pattern, out)
+	}
+}

cmd/restic/lock.go

@@ -2,8 +2,6 @@ package main
 
 import (
 	"context"
-	"fmt"
-	"os"
 	"sync"
 	"time"
 
@@ -36,7 +34,7 @@ func lockRepository(repo *repository.Repository, exclusive bool) (*restic.Lock,
 
 	lock, err := lockFn(context.TODO(), repo)
 	if err != nil {
-		return nil, errors.Fatalf("unable to create lock in backend: %v", err)
+		return nil, errors.WithMessage(err, "unable to create lock in backend")
 	}
 	debug.Log("create lock %p (exclusive %v)", lock, exclusive)
 
@@ -79,7 +77,7 @@ func refreshLocks(wg *sync.WaitGroup, done <-chan struct{}) {
 			for _, lock := range globalLocks.locks {
 				err := lock.Refresh(context.TODO())
 				if err != nil {
-					fmt.Fprintf(os.Stderr, "unable to refresh lock: %v\n", err)
+					Warnf("unable to refresh lock: %v\n", err)
 				}
 			}
 			globalLocks.Unlock()

cmd/restic/main.go

@@ -54,7 +54,7 @@ directories in an encrypted repository stored on different backends.
 		if c.Name() == "version" {
 			return nil
 		}
-		pwd, err := resolvePassword(globalOptions)
+		pwd, err := resolvePassword(globalOptions, "RESTIC_PASSWORD")
 		if err != nil {
 			fmt.Fprintf(os.Stderr, "Resolving password failed: %v\n", err)
 			Exit(1)
@@ -88,6 +88,8 @@ func main() {
 	switch {
 	case restic.IsAlreadyLocked(errors.Cause(err)):
 		fmt.Fprintf(os.Stderr, "%v\nthe `unlock` command can be used to remove stale locks\n", err)
+	case err == ErrInvalidSourceData:
+		fmt.Fprintf(os.Stderr, "Warning: %v\n", err)
 	case errors.IsFatal(errors.Cause(err)):
 		fmt.Fprintf(os.Stderr, "%v\n", err)
 	case err != nil:
@@ -103,9 +105,13 @@ func main() {
 	}
 
 	var exitCode int
-	if err != nil {
+	switch err {
+	case nil:
+		exitCode = 0
+	case ErrInvalidSourceData:
+		exitCode = 3
+	default:
 		exitCode = 1
 	}
-
 	Exit(exitCode)
 }

cmd/restic/progress.go

@@ -0,0 +1,36 @@
+package main
+
+import (
+	"fmt"
+	"time"
+
+	"github.com/restic/restic/internal/restic"
+)
+
+// newProgressMax returns a progress that counts blobs.
+func newProgressMax(show bool, max uint64, description string) *restic.Progress {
+	if !show {
+		return nil
+	}
+
+	p := restic.NewProgress()
+
+	p.OnUpdate = func(s restic.Stat, d time.Duration, ticker bool) {
+		status := fmt.Sprintf("[%s] %s  %d / %d %s",
+			formatDuration(d),
+			formatPercent(s.Blobs, max),
+			s.Blobs, max, description)
+
+		if w := stdoutTerminalWidth(); w > 0 {
+			status = shortenStatus(w, status)
+		}
+
+		PrintProgress("%s", status)
+	}
+
+	p.OnDone = func(s restic.Stat, d time.Duration, ticker bool) {
+		fmt.Printf("\n")
+	}
+
+	return p
+}

cmd/restic/secondary_repo.go

@@ -0,0 +1,48 @@
+package main
+
+import (
+	"os"
+
+	"github.com/restic/restic/internal/errors"
+	"github.com/spf13/pflag"
+)
+
+type secondaryRepoOptions struct {
+	Repo            string
+	password        string
+	PasswordFile    string
+	PasswordCommand string
+	KeyHint         string
+}
+
+func initSecondaryRepoOptions(f *pflag.FlagSet, opts *secondaryRepoOptions, repoPrefix string, repoUsage string) {
+	f.StringVarP(&opts.Repo, "repo2", "", os.Getenv("RESTIC_REPOSITORY2"), repoPrefix+" repository "+repoUsage+" (default: $RESTIC_REPOSITORY2)")
+	f.StringVarP(&opts.PasswordFile, "password-file2", "", os.Getenv("RESTIC_PASSWORD_FILE2"), "`file` to read the "+repoPrefix+" repository password from (default: $RESTIC_PASSWORD_FILE2)")
+	f.StringVarP(&opts.KeyHint, "key-hint2", "", os.Getenv("RESTIC_KEY_HINT2"), "key ID of key to try decrypting the "+repoPrefix+" repository first (default: $RESTIC_KEY_HINT2)")
+	f.StringVarP(&opts.PasswordCommand, "password-command2", "", os.Getenv("RESTIC_PASSWORD_COMMAND2"), "shell `command` to obtain the "+repoPrefix+" repository password from (default: $RESTIC_PASSWORD_COMMAND2)")
+}
+
+func fillSecondaryGlobalOpts(opts secondaryRepoOptions, gopts GlobalOptions, repoPrefix string) (GlobalOptions, error) {
+	if opts.Repo == "" {
+		return GlobalOptions{}, errors.Fatal("Please specify a " + repoPrefix + " repository location (--repo2)")
+	}
+	var err error
+	dstGopts := gopts
+	dstGopts.Repo = opts.Repo
+	dstGopts.PasswordFile = opts.PasswordFile
+	dstGopts.PasswordCommand = opts.PasswordCommand
+	dstGopts.KeyHint = opts.KeyHint
+	if opts.password != "" {
+		dstGopts.password = opts.password
+	} else {
+		dstGopts.password, err = resolvePassword(dstGopts, "RESTIC_PASSWORD2")
+		if err != nil {
+			return GlobalOptions{}, err
+		}
+	}
+	dstGopts.password, err = ReadPassword(dstGopts, "enter password for "+repoPrefix+" repository: ")
+	if err != nil {
+		return GlobalOptions{}, err
+	}
+	return dstGopts, nil
+}

doc/020_installation.rst

@@ -245,7 +245,7 @@ From Source
 ***********
 
 restic is written in the Go programming language and you need at least
-Go version 1.11. Building restic may also work with older versions of Go,
+Go version 1.13. Building restic may also work with older versions of Go,
 but that's not supported. See the `Getting
 started <https://golang.org/doc/install>`__ guide of the Go project for
 instructions how to install Go.
@@ -292,7 +292,7 @@ Restic can write out man pages and bash/zsh compatible autocompletion scripts:
     and the auto-completion files for bash and zsh).
 
     Usage:
-      restic generate [command] [flags]
+      restic generate [flags] [command]
 
     Flags:
           --bash-completion file   write bash completion file

doc/030_preparing_a_new_repo.rst

@@ -43,9 +43,9 @@ command and enter the same password twice:
 .. code-block:: console
 
     $ restic init --repo /srv/restic-repo
-    enter password for new backend:
+    enter password for new repository:
     enter password again:
-    created restic backend 085b3c76b9 at /srv/restic-repo
+    created restic repository 085b3c76b9 at /srv/restic-repo
     Please note that knowledge of your password is required to access the repository.
     Losing your password means that your data is irrecoverably lost.
 
@@ -75,9 +75,9 @@ simply be achieved by changing the URL scheme in the ``init`` command:
 .. code-block:: console
 
     $ restic -r sftp:user@host:/srv/restic-repo init
-    enter password for new backend:
+    enter password for new repository:
     enter password again:
-    created restic backend f1c6108821 at sftp:user@host:/srv/restic-repo
+    created restic repository f1c6108821 at sftp:user@host:/srv/restic-repo
     Please note that knowledge of your password is required to access the repository.
     Losing your password means that your data is irrecoverably lost.
 
@@ -212,9 +212,9 @@ default location:
 .. code-block:: console
 
     $ restic -r s3:s3.amazonaws.com/bucket_name init
-    enter password for new backend:
+    enter password for new repository:
     enter password again:
-    created restic backend eefee03bbd at s3:s3.amazonaws.com/bucket_name
+    created restic repository eefee03bbd at s3:s3.amazonaws.com/bucket_name
     Please note that knowledge of your password is required to access the repository.
     Losing your password means that your data is irrecoverably lost.
 
@@ -262,9 +262,9 @@ this command.
 .. code-block:: console
 
     $ ./restic -r s3:http://localhost:9000/restic init
-    enter password for new backend:
+    enter password for new repository:
     enter password again:
-    created restic backend 6ad29560f5 at s3:http://localhost:9000/restic1
+    created restic repository 6ad29560f5 at s3:http://localhost:9000/restic1
     Please note that knowledge of your password is required to access
     the repository. Losing your password means that your data is irrecoverably lost.
 
@@ -291,9 +291,9 @@ this command.
 .. code-block:: console
 
     $ ./restic -r s3:https://<WASABI-SERVICE-URL>/<WASABI-BUCKET-NAME> init
-    enter password for new backend:
+    enter password for new repository:
     enter password again:
-    created restic backend xxxxxxxxxx at s3:https://<WASABI-SERVICE-URL>/<WASABI-BUCKET-NAME>
+    created restic repository xxxxxxxxxx at s3:https://<WASABI-SERVICE-URL>/<WASABI-BUCKET-NAME>
     Please note that knowledge of your password is required to access
     the repository. Losing your password means that your data is irrecoverably lost.
 
@@ -357,9 +357,9 @@ the container does not exist, it will be created automatically:
 .. code-block:: console
 
    $ restic -r swift:container_name:/path init   # path is optional
-   enter password for new backend:
+   enter password for new repository:
    enter password again:
-   created restic backend eefee03bbd at swift:container_name:/path
+   created restic repository eefee03bbd at swift:container_name:/path
    Please note that knowledge of your password is required to access the repository.
    Losing your password means that your data is irrecoverably lost.
 
@@ -391,9 +391,9 @@ privilege to create buckets, it will be created automatically:
 .. code-block:: console
 
     $ restic -r b2:bucketname:path/to/repo init
-    enter password for new backend:
+    enter password for new repository:
     enter password again:
-    created restic backend eefee03bbd at b2:bucketname:path/to/repo
+    created restic repository eefee03bbd at b2:bucketname:path/to/repo
     Please note that knowledge of your password is required to access the repository.
     Losing your password means that your data is irrecoverably lost.
 
@@ -420,10 +420,10 @@ root path like this:
 .. code-block:: console
 
     $ restic -r azure:foo:/ init
-    enter password for new backend:
+    enter password for new repository:
     enter password again:
 
-    created restic backend a934bac191 at azure:foo:/
+    created restic repository a934bac191 at azure:foo:/
     [...]
 
 The number of concurrent connections to the Azure Blob Storage service can be set with the
@@ -464,10 +464,10 @@ repository in the bucket ``foo`` at the root path:
 .. code-block:: console
 
     $ restic -r gs:foo:/ init
-    enter password for new backend:
+    enter password for new repository:
     enter password again:
 
-    created restic backend bde47d6254 at gs:foo2/
+    created restic repository bde47d6254 at gs:foo2/
     [...]
 
 The number of concurrent connections to the GCS service can be set with the

doc/040_backup.rst

@@ -83,7 +83,7 @@ You can even backup individual files in the same repository (not passing
     snapshot 249d0210 saved
 
 If you're interested in what restic does, pass ``--verbose`` twice (or
-``--verbose 2``) to display detailed information about each file and directory
+``--verbose=2``) to display detailed information about each file and directory
 restic encounters:
 
 .. code-block:: console
@@ -142,7 +142,9 @@ the exclude options are:
 -  ``--iexclude`` Same as ``--exclude`` but ignores the case of paths
 -  ``--exclude-caches`` Specified once to exclude folders containing a special file
 -  ``--exclude-file`` Specified one or more times to exclude items listed in a given file
+-  ``--iexclude-file`` Same as ``exclude-file`` but ignores cases like in ``--iexclude``
 -  ``--exclude-if-present foo`` Specified one or more times to exclude a folder's content if it contains a file called ``foo`` (optionally having a given header, no wildcards for the file name supported)
+-  ``--exclude-larger-than size`` Specified once to excludes files larger than the given size
 
 Please see ``restic help backup`` for more specific information about each exclude option.
 
@@ -213,16 +215,47 @@ On most Unixy shells, you can either quote or use backslashes. For example:
 
 By specifying the option ``--one-file-system`` you can instruct restic
 to only backup files from the file systems the initially specified files
-or directories reside on. For example, calling restic like this won't
-backup ``/sys`` or ``/dev`` on a Linux system:
+or directories reside on. In other words, it will prevent restic from crossing
+filesystem boundaries when performing a backup.
+
+For example, if you backup ``/`` with this option and you have external
+media mounted under ``/media/usb`` then restic will not back up ``/media/usb``
+at all because this is a different filesystem than ``/``. Virtual filesystems
+such as ``/proc`` are also considered different and thereby excluded when
+using ``--one-file-system``:
 
 .. code-block:: console
 
     $ restic -r /srv/restic-repo backup --one-file-system /
 
+Please note that this does not prevent you from specifying multiple filesystems
+on the command line, e.g:
+
+.. code-block:: console
+
+    $ restic -r /srv/restic-repo backup --one-file-system / /media/usb
+
+will back up both the ``/`` and ``/media/usb`` filesystems, but will not
+include other filesystems like ``/sys`` and ``/proc``.
+
 .. note:: ``--one-file-system`` is currently unsupported on Windows, and will
     cause the backup to immediately fail with an error.
 
+Files larger than a given size can be excluded using the `--exclude-larger-than`
+option:
+
+.. code-block:: console
+
+    $ restic -r /srv/restic-repo backup ~/work --exclude-larger-than 1M
+
+This excludes files in ``~/work`` which are larger than 1 MB from the backup.
+
+The default unit for the size value is bytes, so e.g. ``--exclude-larger-than 2048``
+would exclude files larger than 2048 bytes (2 kilobytes). To specify other units,
+suffix the size value with one of ``k``/``K`` for kilobytes, ``m``/``M`` for megabytes,
+``g``/``G`` for gigabytes and ``t``/``T`` for terabytes (e.g. ``1k``, ``10K``, ``20m``,
+``20M``,  ``30g``, ``30G``, ``2t`` or ``2T``).
+
 Including Files
 ***************
 
@@ -366,12 +399,11 @@ created as it would only be written at the very (successful) end of
 the backup operation.  Previous snapshots will still be there and will still
 work.
 
-
 Environment Variables
 *********************
 
 In addition to command-line options, restic supports passing various options in
-environment variables. The following list of environment variables:
+environment variables. The following lists these environment variables:
 
 .. code-block:: console
 
@@ -379,9 +411,13 @@ environment variables. The following list of environment variables:
     RESTIC_PASSWORD_FILE                Location of password file (replaces --password-file)
     RESTIC_PASSWORD                     The actual password for the repository
     RESTIC_PASSWORD_COMMAND             Command printing the password for the repository to stdout
+    RESTIC_KEY_HINT                     ID of key to try decrypting first, before other keys
+    RESTIC_CACHE_DIR                    Location of the cache directory
+    RESTIC_PROGRESS_FPS                 Frames per second by which the progress bar is updated
 
     AWS_ACCESS_KEY_ID                   Amazon S3 access key ID
     AWS_SECRET_ACCESS_KEY               Amazon S3 secret access key
+    AWS_DEFAULT_REGION                  Amazon S3 default region
 
     ST_AUTH                             Auth URL for keystone v1 authentication
     ST_USER                             Username for keystone v1 authentication
@@ -416,5 +452,33 @@ environment variables. The following list of environment variables:
 
     RCLONE_BWLIMIT                      rclone bandwidth limit
 
+In addition to restic-specific environment variables, the following system-wide environment variables
+are taken into account for various operations:
+
+ * ``$XDG_CACHE_HOME/restic``, ``$HOME/.cache/restic``: :ref:`caching`.
+ * ``$TMPDIR``: :ref:`temporary_files`.
+ * ``$PATH/fusermount``: Binary for ``restic mount``.
 
 
+Exit status codes
+*****************
+
+Restic returns one of the following exit status codes after the backup command is run:
+
+ * 0 when the backup was successful (snapshot with all source files created)
+ * 1 when there was a fatal error (no snapshot created)
+ * 3 when some source files could not be read (incomplete snapshot with remaining files created)
+
+Fatal errors occur for example when restic is unable to write to the backup destination, when
+there are network connectivity issues preventing successful communication, or when an invalid
+password or command line argument is provided. When restic returns this exit status code, one
+should not expect a snapshot to have been created.
+
+Source file read errors occur when restic fails to read one or more files or directories that
+it was asked to back up, e.g. due to permission problems. Restic displays the number of source
+file read errors that occurred while running the backup. If there are errors of this type,
+restic will still try to complete the backup run with all the other files, and create a
+snapshot that then contains all but the unreadable files.
+
+One can use these exit status codes in scripts and other automation tools, to make them aware of
+the outcome of the backup run. To manually inspect the exit code in e.g. Linux, run ``echo $?``.

doc/045_working_with_repos.rst

@@ -82,6 +82,89 @@ Furthermore you can group the output by the same filters (host, paths, tags):
     1 snapshots
 
 
+Copying snapshots between repositories
+======================================
+
+In case you want to transfer snapshots between two repositories, for
+example from a local to a remote repository, you can use the ``copy`` command:
+
+.. code-block:: console
+
+    $ restic -r /srv/restic-repo copy --repo2 /srv/restic-repo-copy
+    repository d6504c63 opened successfully, password is correct
+    repository 3dd0878c opened successfully, password is correct
+
+    snapshot 410b18a2 of [/home/user/work] at 2020-06-09 23:15:57.305305 +0200 CEST)
+      copy started, this may take a while...
+    snapshot 7a746a07 saved
+
+    snapshot 4e5d5487 of [/home/user/work] at 2020-05-01 22:44:07.012113 +0200 CEST)
+    skipping snapshot 4e5d5487, was already copied to snapshot 50eb62b7
+
+The example command copies all snapshots from the source repository
+``/srv/restic-repo`` to the destination repository ``/srv/restic-repo-copy``.
+Snapshots which have previously been copied between repositories will
+be skipped by later copy runs.
+
+.. note:: Note that this process will have to read (download) and write (upload) the
+    entire snapshot(s) due to the different encryption keys used in the source and
+    destination repository. Also, the transferred files are not re-chunked, which
+    may break deduplication between files already stored in the destination repo
+    and files copied there using this command. See the next section for how to avoid
+    this problem.
+
+For the destination repository ``--repo2`` the password can be read from
+a file ``--password-file2`` or from a command ``--password-command2``.
+Alternatively the environment variables ``$RESTIC_PASSWORD_COMMAND2`` and
+``$RESTIC_PASSWORD_FILE2`` can be used. It is also possible to directly
+pass the password via ``$RESTIC_PASSWORD2``. The key which should be used
+for decryption can be selected by passing its ID via the flag ``--key-hint2``
+or the environment variable ``$RESTIC_KEY_HINT2``.
+
+In case the source and destination repository use the same backend, then
+configuration options and environment variables to configure the backend
+apply to both repositories. For example it might not be possible to specify
+different accounts for the source and destination repository. You can
+avoid this limitation by using the rclone backend along with remotes which
+are configured in rclone.
+
+The list of snapshots to copy can be filtered by host, path in the backup
+and / or a comma-separated tag list:
+
+.. code-block:: console
+
+    $ restic -r /srv/restic-repo copy --repo2 /srv/restic-repo-copy --host luigi --path /srv --tag foo,bar
+
+It is also possible to explicitly specify the list of snapshots to copy, in
+which case only these instead of all snapshots will be copied:
+
+.. code-block:: console
+
+    $ restic -r /srv/restic-repo copy --repo2 /srv/restic-repo-copy 410b18a2 4e5d5487 latest
+
+
+Ensuring deduplication for copied snapshots
+-------------------------------------------
+
+Even though the copy command can transfer snapshots between arbitrary repositories,
+deduplication between snapshots from the source and destination repository may not work.
+To ensure proper deduplication, both repositories have to use the same parameters for
+splitting large files into smaller chunks, which requires additional setup steps. With
+the same parameters restic will for both repositories split identical files into
+identical chunks and therefore deduplication also works for snapshots copied between
+these repositories.
+
+The chunker parameters are generated once when creating a new (destination) repository.
+That is for a copy destination repository we have to instruct restic to initialize it
+using the same chunker parameters as the source repository:
+
+.. code-block:: console
+
+    $ restic -r /srv/restic-repo-copy init --repo2 /srv/restic-repo --copy-chunker-params
+
+Note that it is not possible to change the chunker parameters of an existing repository.
+
+
 Checking integrity and consistency
 ==================================
 
@@ -134,10 +217,10 @@ If the repository structure is intact, restic will show that no errors were foun
     check snapshots, trees and blobs
     no errors were found
 
-By default, the ``check`` command does not verify that the actual data files
+By default, the ``check`` command does not verify that the actual pack files
 on disk in the repository are unmodified, because doing so requires reading
-a copy of every data file in the repository. To tell restic to also verify the
-integrity of the data files in the repository, use the ``--read-data`` flag:
+a copy of every pack file in the repository. To tell restic to also verify the
+integrity of the pack files in the repository, use the ``--read-data`` flag:
 
 .. code-block:: console
 
@@ -151,16 +234,16 @@ integrity of the data files in the repository, use the ``--read-data`` flag:
     duration: 0:00
     no errors were found
 
-.. note:: Since ``--read-data`` has to download all data files in the
+.. note:: Since ``--read-data`` has to download all pack files in the
     repository, beware that it might incur higher bandwidth costs than usual
     and also that it takes more time than the default ``check``.
 
 Alternatively, use the ``--read-data-subset=n/t`` parameter to check only a
-subset of the repository data files at a time. The parameter takes two values,
-``n`` and ``t``. When the check command runs, all data files in the repository
+subset of the repository pack files at a time. The parameter takes two values,
+``n`` and ``t``. When the check command runs, all pack files in the repository
 are logically divided in ``t`` (roughly equal) groups, and only files that
 belong to group number ``n`` are checked. For example, the following commands
-check all repository data files over 5 separate invocations:
+check all repository pack files over 5 separate invocations:
 
 .. code-block:: console
 

doc/050_restore.rst

@@ -52,7 +52,7 @@ You can use the command ``restic ls latest`` or ``restic find foo`` to find the
 path to the file within the snapshot. This path you can then pass to
 ``--include`` in verbatim to only restore the single file or directory.
 
-There are case insensitive variants of of ``--exclude`` and ``--include`` called
+There are case insensitive variants of ``--exclude`` and ``--include`` called
 ``--iexclude`` and ``--iinclude``. These options will behave the same way but
 ignore the casing of paths.
 

doc/060_forget.rst

@@ -176,7 +176,7 @@ Multiple policies will be ORed together so as to be as inclusive as possible
 for keeping snapshots.
 
 Additionally, you can restrict removing snapshots to those which have a
-particular hostname with the ``--hostname`` parameter, or tags with the
+particular hostname with the ``--host`` parameter, or tags with the
 ``--tag`` option. When multiple tags are specified, only the snapshots
 which have all the tags are considered. For example, the following command
 removes all but the latest snapshot of all snapshots that have the tag ``foo``:

doc/080_examples.rst

@@ -311,13 +311,13 @@ the backups:
 
    root@a3e580b6369d:/# useradd -m restic
 
-Then we download and install the restic binary into the user's home
-directory.
+Then we download and install the restic binary into the user's home 
+directory (please adjust the URL to refer to the latest restic version).
 
 .. code-block:: console
 
    root@a3e580b6369d:/# mkdir ~restic/bin
-   root@a3e580b6369d:/# curl -L https://github.com/restic/restic/releases/download/v0.9.1/restic_0.9.1_linux_amd64.bz2 | bunzip2 > ~restic/bin/restic
+   root@a3e580b6369d:/# curl -L https://github.com/restic/restic/releases/download/v0.9.6/restic_0.9.6_linux_amd64.bz2 | bunzip2 > ~restic/bin/restic
 
 Before we assign any special capability to the restic binary we
 restrict its permissions so that only root and the newly created