mirror of
https://github.com/octoleo/syncthing.git
synced 2024-11-09 14:50:56 +00:00
214 lines
6.7 KiB
Go
214 lines
6.7 KiB
Go
// Copyright (C) 2018 The Syncthing Authors.
|
|
//
|
|
// This Source Code Form is subject to the terms of the Mozilla Public
|
|
// License, v. 2.0. If a copy of the MPL was not distributed with this file,
|
|
// You can obtain one at https://mozilla.org/MPL/2.0/.
|
|
|
|
package model
|
|
|
|
import (
|
|
"sort"
|
|
"time"
|
|
|
|
"github.com/syncthing/syncthing/lib/config"
|
|
"github.com/syncthing/syncthing/lib/db"
|
|
"github.com/syncthing/syncthing/lib/events"
|
|
"github.com/syncthing/syncthing/lib/fs"
|
|
"github.com/syncthing/syncthing/lib/ignore"
|
|
"github.com/syncthing/syncthing/lib/protocol"
|
|
"github.com/syncthing/syncthing/lib/versioner"
|
|
)
|
|
|
|
func init() {
|
|
folderFactories[config.FolderTypeReceiveOnly] = newReceiveOnlyFolder
|
|
}
|
|
|
|
/*
|
|
receiveOnlyFolder is a folder that does not propagate local changes outward.
|
|
It does this by the following general mechanism (not all of which is
|
|
implemted in this file):
|
|
|
|
- Local changes are scanned and versioned as usual, but get the
|
|
FlagLocalReceiveOnly bit set.
|
|
|
|
- When changes are sent to the cluster this bit gets converted to the
|
|
Invalid bit (like all other local flags, currently) and also the Version
|
|
gets set to the empty version. The reason for clearing the Version is to
|
|
ensure that other devices will not consider themselves out of date due to
|
|
our change.
|
|
|
|
- The database layer accounts sizes per flag bit, so we can know how many
|
|
files have been changed locally. We use this to trigger a "Revert" option
|
|
on the folder when the amount of locally changed data is nonzero.
|
|
|
|
- To revert we take the files which have changed and reset their version
|
|
counter down to zero. The next pull will replace our changed version with
|
|
the globally latest. As this is a user-initiated operation we do not cause
|
|
conflict copies when reverting.
|
|
|
|
- When pulling normally (i.e., not in the revert case) with local changes,
|
|
normal conflict resolution will apply. Conflict copies will be created,
|
|
but not propagated outwards (because receive only, right).
|
|
|
|
Implementation wise a receiveOnlyFolder is just a sendReceiveFolder that
|
|
sets an extra bit on local changes and has a Revert method.
|
|
*/
|
|
type receiveOnlyFolder struct {
|
|
*sendReceiveFolder
|
|
}
|
|
|
|
func newReceiveOnlyFolder(model *model, fset *db.FileSet, ignores *ignore.Matcher, cfg config.FolderConfiguration, ver versioner.Versioner, fs fs.Filesystem, evLogger events.Logger, ioLimiter *byteSemaphore) service {
|
|
sr := newSendReceiveFolder(model, fset, ignores, cfg, ver, fs, evLogger, ioLimiter).(*sendReceiveFolder)
|
|
sr.localFlags = protocol.FlagLocalReceiveOnly // gets propagated to the scanner, and set on locally changed files
|
|
return &receiveOnlyFolder{sr}
|
|
}
|
|
|
|
func (f *receiveOnlyFolder) Revert() {
|
|
f.doInSync(func() error { f.revert(); return nil })
|
|
}
|
|
|
|
func (f *receiveOnlyFolder) revert() {
|
|
l.Infof("Reverting folder %v", f.Description)
|
|
|
|
f.setState(FolderScanning)
|
|
defer f.setState(FolderIdle)
|
|
|
|
scanChan := make(chan string)
|
|
go f.pullScannerRoutine(scanChan)
|
|
defer close(scanChan)
|
|
|
|
delQueue := &deleteQueue{
|
|
handler: f, // for the deleteItemOnDisk and deleteDirOnDisk methods
|
|
ignores: f.ignores,
|
|
scanChan: scanChan,
|
|
}
|
|
|
|
batch := make([]protocol.FileInfo, 0, maxBatchSizeFiles)
|
|
batchSizeBytes := 0
|
|
snap := f.fset.Snapshot()
|
|
defer snap.Release()
|
|
snap.WithHave(protocol.LocalDeviceID, func(intf db.FileIntf) bool {
|
|
fi := intf.(protocol.FileInfo)
|
|
if !fi.IsReceiveOnlyChanged() {
|
|
// We're only interested in files that have changed locally in
|
|
// receive only mode.
|
|
return true
|
|
}
|
|
|
|
fi.LocalFlags &^= protocol.FlagLocalReceiveOnly
|
|
if len(fi.Version.Counters) == 1 && fi.Version.Counters[0].ID == f.shortID {
|
|
// We are the only device mentioned in the version vector so the
|
|
// file must originate here. A revert then means to delete it.
|
|
// We'll delete files directly, directories get queued and
|
|
// handled below.
|
|
|
|
handled, err := delQueue.handle(fi, snap)
|
|
if err != nil {
|
|
l.Infof("Revert: deleting %s: %v\n", fi.Name, err)
|
|
return true // continue
|
|
}
|
|
if !handled {
|
|
return true // continue
|
|
}
|
|
|
|
fi.SetDeleted(f.shortID)
|
|
fi.Version = protocol.Vector{} // if this file ever resurfaces anywhere we want our delete to be strictly older
|
|
} else {
|
|
// Revert means to throw away our local changes. We reset the
|
|
// version to the empty vector, which is strictly older than any
|
|
// other existing version. It is not in conflict with anything,
|
|
// either, so we will not create a conflict copy of our local
|
|
// changes.
|
|
fi.Version = protocol.Vector{}
|
|
}
|
|
|
|
batch = append(batch, fi)
|
|
batchSizeBytes += fi.ProtoSize()
|
|
|
|
if len(batch) >= maxBatchSizeFiles || batchSizeBytes >= maxBatchSizeBytes {
|
|
f.updateLocalsFromScanning(batch)
|
|
batch = batch[:0]
|
|
batchSizeBytes = 0
|
|
}
|
|
return true
|
|
})
|
|
if len(batch) > 0 {
|
|
f.updateLocalsFromScanning(batch)
|
|
}
|
|
batch = batch[:0]
|
|
batchSizeBytes = 0
|
|
|
|
// Handle any queued directories
|
|
deleted, err := delQueue.flush(snap)
|
|
if err != nil {
|
|
l.Infoln("Revert:", err)
|
|
}
|
|
now := time.Now()
|
|
for _, dir := range deleted {
|
|
batch = append(batch, protocol.FileInfo{
|
|
Name: dir,
|
|
Type: protocol.FileInfoTypeDirectory,
|
|
ModifiedS: now.Unix(),
|
|
ModifiedBy: f.shortID,
|
|
Deleted: true,
|
|
Version: protocol.Vector{},
|
|
})
|
|
}
|
|
if len(batch) > 0 {
|
|
f.updateLocalsFromScanning(batch)
|
|
}
|
|
|
|
// We will likely have changed our local index, but that won't trigger a
|
|
// pull by itself. Make sure we schedule one so that we start
|
|
// downloading files.
|
|
f.SchedulePull()
|
|
}
|
|
|
|
// deleteQueue handles deletes by delegating to a handler and queuing
|
|
// directories for last.
|
|
type deleteQueue struct {
|
|
handler interface {
|
|
deleteItemOnDisk(item protocol.FileInfo, snap *db.Snapshot, scanChan chan<- string) error
|
|
deleteDirOnDisk(dir string, snap *db.Snapshot, scanChan chan<- string) error
|
|
}
|
|
ignores *ignore.Matcher
|
|
dirs []string
|
|
scanChan chan<- string
|
|
}
|
|
|
|
func (q *deleteQueue) handle(fi protocol.FileInfo, snap *db.Snapshot) (bool, error) {
|
|
// Things that are ignored but not marked deletable are not processed.
|
|
ign := q.ignores.Match(fi.Name)
|
|
if ign.IsIgnored() && !ign.IsDeletable() {
|
|
return false, nil
|
|
}
|
|
|
|
// Directories are queued for later processing.
|
|
if fi.IsDirectory() {
|
|
q.dirs = append(q.dirs, fi.Name)
|
|
return false, nil
|
|
}
|
|
|
|
// Kill it.
|
|
err := q.handler.deleteItemOnDisk(fi, snap, q.scanChan)
|
|
return true, err
|
|
}
|
|
|
|
func (q *deleteQueue) flush(snap *db.Snapshot) ([]string, error) {
|
|
// Process directories from the leaves inward.
|
|
sort.Sort(sort.Reverse(sort.StringSlice(q.dirs)))
|
|
|
|
var firstError error
|
|
var deleted []string
|
|
|
|
for _, dir := range q.dirs {
|
|
if err := q.handler.deleteDirOnDisk(dir, snap, q.scanChan); err == nil {
|
|
deleted = append(deleted, dir)
|
|
} else if err != nil && firstError == nil {
|
|
firstError = err
|
|
}
|
|
}
|
|
|
|
return deleted, firstError
|
|
}
|