Dump: add output format tar and output to stdout (#10376)

* Dump: Use mholt/archive/v3 to support tar including many compressions

Signed-off-by: Philipp Homann <homann.philipp@googlemail.com>

* Dump: Allow dump output to stdout

Signed-off-by: Philipp Homann <homann.philipp@googlemail.com>

* Dump: Fixed bug present since #6677 where SessionConfig.Provider is never "file"

Signed-off-by: Philipp Homann <homann.philipp@googlemail.com>

* Dump: never pack RepoRootPath, LFS.ContentPath and LogRootPath when they are below AppDataPath

Signed-off-by: Philipp Homann <homann.philipp@googlemail.com>

* Dump: also dump LFS (fixes #10058)

Signed-off-by: Philipp Homann <homann.philipp@googlemail.com>

* Dump: never dump CustomPath if CustomPath is a subdir of or equal to AppDataPath (fixes #10365)

Signed-off-by: Philipp Homann <homann.philipp@googlemail.com>

* Use log.Info instead of fmt.Fprintf

Signed-off-by: Philipp Homann <homann.philipp@googlemail.com>

* import ordering

* make fmt

Co-authored-by: zeripath <art27@cantab.net>
Co-authored-by: techknowlogick <techknowlogick@gitea.io>
Co-authored-by: Matti R <matti@mdranta.net>

vendor/github.com/andybalholm/brotli/encoder_dict.go

@@ -0,0 +1,22 @@
+package brotli
+
+/* Dictionary data (words and transforms) for 1 possible context */
+type encoderDictionary struct {
+	words                 *dictionary
+	cutoffTransformsCount uint32
+	cutoffTransforms      uint64
+	hash_table            []uint16
+	buckets               []uint16
+	dict_words            []dictWord
+}
+
+func initEncoderDictionary(dict *encoderDictionary) {
+	dict.words = getDictionary()
+
+	dict.hash_table = kStaticDictionaryHash[:]
+	dict.buckets = kStaticDictionaryBuckets[:]
+	dict.dict_words = kStaticDictionaryWords[:]
+
+	dict.cutoffTransformsCount = kCutoffTransformsCount
+	dict.cutoffTransforms = kCutoffTransforms
+}