rclone/fs/operations/check.go
nielash fd8faeb0e6 vfs: fix unicode normalization on macOS - fixes #7072
Before this change, the VFS layer did not properly handle unicode normalization,
which caused problems particularly for users of macOS. While attempts were made
to handle it with various `-o modules=iconv` combinations, this was an imperfect
solution, as no one combination allowed both NFC and NFD content to
simultaneously be both visible and editable via Finder.

After this change, the VFS supports `--no-unicode-normalization` (default `false`)
via the existing `--vfs-case-insensitive` logic, which is extended to apply to both
case insensitivity and unicode normalization form.

This change also adds an additional flag, `--vfs-block-norm-dupes`, to address a
probably rare but potentially possible scenario where a directory contains
multiple duplicate filenames after applying case and unicode normalization
settings. In such a scenario, this flag (disabled by default) hides the
duplicates. This comes with a performance tradeoff, as rclone will have to scan
the entire directory for duplicates when listing a directory. For this reason,
it is recommended to leave this disabled if not needed. However, macOS users may
wish to consider using it, as otherwise, if a remote directory contains both NFC
and NFD versions of the same filename, an odd situation will occur: both
versions of the file will be visible in the mount, and both will appear to be
editable, however, editing either version will actually result in only the NFD
version getting edited under the hood. `--vfs-block-norm-dupes` prevents this
confusion by detecting this scenario, hiding the duplicates, and logging an
error, similar to how this is handled in `rclone sync`.
2024-03-06 16:12:13 +00:00

619 lines
17 KiB
Go

package operations
import (
"bufio"
"bytes"
"context"
"errors"
"fmt"
"io"
"os"
"regexp"
"strings"
"sync"
"sync/atomic"
"github.com/rclone/rclone/fs"
"github.com/rclone/rclone/fs/accounting"
"github.com/rclone/rclone/fs/filter"
"github.com/rclone/rclone/fs/fserrors"
"github.com/rclone/rclone/fs/hash"
"github.com/rclone/rclone/fs/march"
"github.com/rclone/rclone/lib/readers"
"golang.org/x/text/unicode/norm"
)
// checkFn is the type of the checking function used in CheckFn()
//
// It should check the two objects (a, b) and return if they differ
// and whether the hash was used.
//
// If there are differences then this should Errorf the difference and
// the reason but return with err = nil. It should not CountError in
// this case.
type checkFn func(ctx context.Context, a, b fs.Object) (differ bool, noHash bool, err error)
// CheckOpt contains options for the Check functions
type CheckOpt struct {
Fdst, Fsrc fs.Fs // fses to check
Check checkFn // function to use for checking
OneWay bool // one way only?
Combined io.Writer // a file with file names with leading sigils
MissingOnSrc io.Writer // files only in the destination
MissingOnDst io.Writer // files only in the source
Match io.Writer // matching files
Differ io.Writer // differing files
Error io.Writer // files with errors of some kind
}
// checkMarch is used to march over two Fses in the same way as
// sync/copy
type checkMarch struct {
ioMu sync.Mutex
wg sync.WaitGroup
tokens chan struct{}
differences atomic.Int32
noHashes atomic.Int32
srcFilesMissing atomic.Int32
dstFilesMissing atomic.Int32
matches atomic.Int32
opt CheckOpt
}
// report outputs the fileName to out if required and to the combined log
func (c *checkMarch) report(o fs.DirEntry, out io.Writer, sigil rune) {
c.reportFilename(o.String(), out, sigil)
}
func (c *checkMarch) reportFilename(filename string, out io.Writer, sigil rune) {
if out != nil {
SyncFprintf(out, "%s\n", filename)
}
if c.opt.Combined != nil {
SyncFprintf(c.opt.Combined, "%c %s\n", sigil, filename)
}
}
// DstOnly have an object which is in the destination only
func (c *checkMarch) DstOnly(dst fs.DirEntry) (recurse bool) {
switch dst.(type) {
case fs.Object:
if c.opt.OneWay {
return false
}
err := fmt.Errorf("file not in %v", c.opt.Fsrc)
fs.Errorf(dst, "%v", err)
_ = fs.CountError(err)
c.differences.Add(1)
c.srcFilesMissing.Add(1)
c.report(dst, c.opt.MissingOnSrc, '-')
case fs.Directory:
// Do the same thing to the entire contents of the directory
if c.opt.OneWay {
return false
}
return true
default:
panic("Bad object in DirEntries")
}
return false
}
// SrcOnly have an object which is in the source only
func (c *checkMarch) SrcOnly(src fs.DirEntry) (recurse bool) {
switch src.(type) {
case fs.Object:
err := fmt.Errorf("file not in %v", c.opt.Fdst)
fs.Errorf(src, "%v", err)
_ = fs.CountError(err)
c.differences.Add(1)
c.dstFilesMissing.Add(1)
c.report(src, c.opt.MissingOnDst, '+')
case fs.Directory:
// Do the same thing to the entire contents of the directory
return true
default:
panic("Bad object in DirEntries")
}
return false
}
// check to see if two objects are identical using the check function
func (c *checkMarch) checkIdentical(ctx context.Context, dst, src fs.Object) (differ bool, noHash bool, err error) {
ci := fs.GetConfig(ctx)
tr := accounting.Stats(ctx).NewCheckingTransfer(src, "checking")
defer func() {
tr.Done(ctx, err)
}()
if sizeDiffers(ctx, src, dst) {
err = fmt.Errorf("sizes differ")
fs.Errorf(src, "%v", err)
return true, false, nil
}
if ci.SizeOnly {
return false, false, nil
}
return c.opt.Check(ctx, dst, src)
}
// Match is called when src and dst are present, so sync src to dst
func (c *checkMarch) Match(ctx context.Context, dst, src fs.DirEntry) (recurse bool) {
switch srcX := src.(type) {
case fs.Object:
dstX, ok := dst.(fs.Object)
if ok {
if SkipDestructive(ctx, src, "check") {
return false
}
c.wg.Add(1)
c.tokens <- struct{}{} // put a token to limit concurrency
go func() {
defer func() {
<-c.tokens // get the token back to free up a slot
c.wg.Done()
}()
differ, noHash, err := c.checkIdentical(ctx, dstX, srcX)
if err != nil {
fs.Errorf(src, "%v", err)
_ = fs.CountError(err)
c.report(src, c.opt.Error, '!')
} else if differ {
c.differences.Add(1)
err := errors.New("files differ")
// the checkFn has already logged the reason
_ = fs.CountError(err)
c.report(src, c.opt.Differ, '*')
} else {
c.matches.Add(1)
c.report(src, c.opt.Match, '=')
if noHash {
c.noHashes.Add(1)
fs.Debugf(dstX, "OK - could not check hash")
} else {
fs.Debugf(dstX, "OK")
}
}
}()
} else {
err := fmt.Errorf("is file on %v but directory on %v", c.opt.Fsrc, c.opt.Fdst)
fs.Errorf(src, "%v", err)
_ = fs.CountError(err)
c.differences.Add(1)
c.dstFilesMissing.Add(1)
c.report(src, c.opt.MissingOnDst, '+')
}
case fs.Directory:
// Do the same thing to the entire contents of the directory
_, ok := dst.(fs.Directory)
if ok {
return true
}
err := fmt.Errorf("is file on %v but directory on %v", c.opt.Fdst, c.opt.Fsrc)
fs.Errorf(dst, "%v", err)
_ = fs.CountError(err)
c.differences.Add(1)
c.srcFilesMissing.Add(1)
c.report(dst, c.opt.MissingOnSrc, '-')
default:
panic("Bad object in DirEntries")
}
return false
}
// CheckFn checks the files in fsrc and fdst according to Size and
// hash using checkFunction on each file to check the hashes.
//
// checkFunction sees if dst and src are identical
//
// it returns true if differences were found
// it also returns whether it couldn't be hashed
func CheckFn(ctx context.Context, opt *CheckOpt) error {
ci := fs.GetConfig(ctx)
if opt.Check == nil {
return errors.New("internal error: nil check function")
}
c := &checkMarch{
tokens: make(chan struct{}, ci.Checkers),
opt: *opt,
}
// set up a march over fdst and fsrc
m := &march.March{
Ctx: ctx,
Fdst: c.opt.Fdst,
Fsrc: c.opt.Fsrc,
Dir: "",
Callback: c,
NoTraverse: ci.NoTraverse,
NoUnicodeNormalization: ci.NoUnicodeNormalization,
}
fs.Debugf(c.opt.Fdst, "Waiting for checks to finish")
err := m.Run(ctx)
c.wg.Wait() // wait for background go-routines
return c.reportResults(ctx, err)
}
func (c *checkMarch) reportResults(ctx context.Context, err error) error {
if c.dstFilesMissing.Load() > 0 {
fs.Logf(c.opt.Fdst, "%d files missing", c.dstFilesMissing.Load())
}
if c.srcFilesMissing.Load() > 0 {
entity := "files"
if c.opt.Fsrc == nil {
entity = "hashes"
}
fs.Logf(c.opt.Fsrc, "%d %s missing", c.srcFilesMissing.Load(), entity)
}
fs.Logf(c.opt.Fdst, "%d differences found", accounting.Stats(ctx).GetErrors())
if errs := accounting.Stats(ctx).GetErrors(); errs > 0 {
fs.Logf(c.opt.Fdst, "%d errors while checking", errs)
}
if c.noHashes.Load() > 0 {
fs.Logf(c.opt.Fdst, "%d hashes could not be checked", c.noHashes.Load())
}
if c.matches.Load() > 0 {
fs.Logf(c.opt.Fdst, "%d matching files", c.matches.Load())
}
if err != nil {
return err
}
if c.differences.Load() > 0 {
// Return an already counted error so we don't double count this error too
err = fserrors.FsError(fmt.Errorf("%d differences found", c.differences.Load()))
fserrors.Count(err)
return err
}
return nil
}
// Check the files in fsrc and fdst according to Size and hash
func Check(ctx context.Context, opt *CheckOpt) error {
optCopy := *opt
optCopy.Check = func(ctx context.Context, dst, src fs.Object) (differ bool, noHash bool, err error) {
same, ht, err := CheckHashes(ctx, src, dst)
if err != nil {
return true, false, err
}
if ht == hash.None {
return false, true, nil
}
if !same {
err = fmt.Errorf("%v differ", ht)
fs.Errorf(src, "%v", err)
return true, false, nil
}
return false, false, nil
}
return CheckFn(ctx, &optCopy)
}
// CheckEqualReaders checks to see if in1 and in2 have the same
// content when read.
//
// it returns true if differences were found
func CheckEqualReaders(in1, in2 io.Reader) (differ bool, err error) {
const bufSize = 64 * 1024
buf1 := make([]byte, bufSize)
buf2 := make([]byte, bufSize)
for {
n1, err1 := readers.ReadFill(in1, buf1)
n2, err2 := readers.ReadFill(in2, buf2)
// check errors
if err1 != nil && err1 != io.EOF {
return true, err1
} else if err2 != nil && err2 != io.EOF {
return true, err2
}
// err1 && err2 are nil or io.EOF here
// process the data
if n1 != n2 || !bytes.Equal(buf1[:n1], buf2[:n2]) {
return true, nil
}
// if both streams finished the we have finished
if err1 == io.EOF && err2 == io.EOF {
break
}
}
return false, nil
}
// CheckIdenticalDownload checks to see if dst and src are identical
// by reading all their bytes if necessary.
//
// it returns true if differences were found
func CheckIdenticalDownload(ctx context.Context, dst, src fs.Object) (differ bool, err error) {
ci := fs.GetConfig(ctx)
err = Retry(ctx, src, ci.LowLevelRetries, func() error {
differ, err = checkIdenticalDownload(ctx, dst, src)
return err
})
return differ, err
}
// Does the work for CheckIdenticalDownload
func checkIdenticalDownload(ctx context.Context, dst, src fs.Object) (differ bool, err error) {
var in1, in2 io.ReadCloser
in1, err = Open(ctx, dst)
if err != nil {
return true, fmt.Errorf("failed to open %q: %w", dst, err)
}
tr1 := accounting.Stats(ctx).NewTransfer(dst, nil)
defer func() {
tr1.Done(ctx, nil) // error handling is done by the caller
}()
in1 = tr1.Account(ctx, in1).WithBuffer() // account and buffer the transfer
in2, err = Open(ctx, src)
if err != nil {
return true, fmt.Errorf("failed to open %q: %w", src, err)
}
tr2 := accounting.Stats(ctx).NewTransfer(dst, nil)
defer func() {
tr2.Done(ctx, nil) // error handling is done by the caller
}()
in2 = tr2.Account(ctx, in2).WithBuffer() // account and buffer the transfer
// To assign err variable before defer.
differ, err = CheckEqualReaders(in1, in2)
return
}
// CheckDownload checks the files in fsrc and fdst according to Size
// and the actual contents of the files.
func CheckDownload(ctx context.Context, opt *CheckOpt) error {
optCopy := *opt
optCopy.Check = func(ctx context.Context, a, b fs.Object) (differ bool, noHash bool, err error) {
differ, err = CheckIdenticalDownload(ctx, a, b)
if err != nil {
return true, true, fmt.Errorf("failed to download: %w", err)
}
return differ, false, nil
}
return CheckFn(ctx, &optCopy)
}
// ApplyTransforms handles --no-unicode-normalization and --ignore-case-sync for CheckSum
// so that it matches behavior of Check (where it's handled by March)
func ApplyTransforms(ctx context.Context, s string) string {
ci := fs.GetConfig(ctx)
return ToNormal(s, !ci.NoUnicodeNormalization, ci.IgnoreCaseSync)
}
// ToNormal normalizes case and unicode form and returns the transformed string.
// It is similar to ApplyTransforms but does not use a context.
// If normUnicode == true, s will be transformed to NFC.
// If normCase == true, s will be transformed to lowercase.
// If both are true, both transformations will be performed.
func ToNormal(s string, normUnicode, normCase bool) string {
if normUnicode {
s = norm.NFC.String(s)
}
if normCase {
s = strings.ToLower(s)
}
return s
}
// CheckSum checks filesystem hashes against a SUM file
func CheckSum(ctx context.Context, fsrc, fsum fs.Fs, sumFile string, hashType hash.Type, opt *CheckOpt, download bool) error {
var options CheckOpt
if opt != nil {
options = *opt
} else {
// default options for hashsum -c
options.Combined = os.Stdout
}
// CheckSum treats Fsrc and Fdst specially:
options.Fsrc = nil // no file system here, corresponds to the sum list
options.Fdst = fsrc // denotes the file system to check
opt = &options // override supplied argument
if !download && (hashType == hash.None || !opt.Fdst.Hashes().Contains(hashType)) {
return fmt.Errorf("%s: hash type is not supported by file system: %s", hashType, opt.Fdst)
}
if sumFile == "" {
return fmt.Errorf("not a sum file: %s", fsum)
}
sumObj, err := fsum.NewObject(ctx, sumFile)
if err != nil {
return fmt.Errorf("cannot open sum file: %w", err)
}
hashes, err := ParseSumFile(ctx, sumObj)
if err != nil {
return fmt.Errorf("failed to parse sum file: %w", err)
}
ci := fs.GetConfig(ctx)
c := &checkMarch{
tokens: make(chan struct{}, ci.Checkers),
opt: *opt,
}
lastErr := ListFn(ctx, opt.Fdst, func(obj fs.Object) {
c.checkSum(ctx, obj, download, hashes, hashType)
})
c.wg.Wait() // wait for background go-routines
// make census of unhandled sums
fi := filter.GetConfig(ctx)
for filename, hash := range hashes {
if hash == "" { // the sum has been successfully consumed
continue
}
if !fi.IncludeRemote(filename) { // the file was filtered out
continue
}
// filesystem missed the file, sum wasn't consumed
err := fmt.Errorf("file not in %v", opt.Fdst)
fs.Errorf(filename, "%v", err)
_ = fs.CountError(err)
if lastErr == nil {
lastErr = err
}
c.dstFilesMissing.Add(1)
c.reportFilename(filename, opt.MissingOnDst, '+')
}
return c.reportResults(ctx, lastErr)
}
// checkSum checks single object against golden hashes
func (c *checkMarch) checkSum(ctx context.Context, obj fs.Object, download bool, hashes HashSums, hashType hash.Type) {
normalizedRemote := ApplyTransforms(ctx, obj.Remote())
c.ioMu.Lock()
sumHash, sumFound := hashes[normalizedRemote]
hashes[normalizedRemote] = "" // mark sum as consumed
c.ioMu.Unlock()
if !sumFound && c.opt.OneWay {
return
}
var err error
tr := accounting.Stats(ctx).NewCheckingTransfer(obj, "hashing")
defer tr.Done(ctx, err)
if !sumFound {
err = errors.New("sum not found")
_ = fs.CountError(err)
fs.Errorf(obj, "%v", err)
c.differences.Add(1)
c.srcFilesMissing.Add(1)
c.report(obj, c.opt.MissingOnSrc, '-')
return
}
if !download {
var objHash string
objHash, err = obj.Hash(ctx, hashType)
c.matchSum(ctx, sumHash, objHash, obj, err, hashType)
return
}
c.wg.Add(1)
c.tokens <- struct{}{} // put a token to limit concurrency
go func() {
var (
objHash string
err error
in io.ReadCloser
)
defer func() {
c.matchSum(ctx, sumHash, objHash, obj, err, hashType)
<-c.tokens // get the token back to free up a slot
c.wg.Done()
}()
if in, err = Open(ctx, obj); err != nil {
return
}
tr := accounting.Stats(ctx).NewTransfer(obj, nil)
in = tr.Account(ctx, in).WithBuffer() // account and buffer the transfer
defer func() {
tr.Done(ctx, nil) // will close the stream
}()
hashVals, err2 := hash.StreamTypes(in, hash.NewHashSet(hashType))
if err2 != nil {
err = err2 // pass to matchSum
return
}
objHash = hashVals[hashType]
}()
}
// matchSum sums up the results of hashsum matching for an object
func (c *checkMarch) matchSum(ctx context.Context, sumHash, objHash string, obj fs.Object, err error, hashType hash.Type) {
switch {
case err != nil:
_ = fs.CountError(err)
fs.Errorf(obj, "Failed to calculate hash: %v", err)
c.report(obj, c.opt.Error, '!')
case sumHash == "":
err = errors.New("duplicate file")
_ = fs.CountError(err)
fs.Errorf(obj, "%v", err)
c.report(obj, c.opt.Error, '!')
case objHash == "":
fs.Debugf(nil, "%v = %s (sum)", hashType, sumHash)
fs.Debugf(obj, "%v - could not check hash (%v)", hashType, c.opt.Fdst)
c.noHashes.Add(1)
c.matches.Add(1)
c.report(obj, c.opt.Match, '=')
case objHash == sumHash:
fs.Debugf(obj, "%v = %s OK", hashType, sumHash)
c.matches.Add(1)
c.report(obj, c.opt.Match, '=')
default:
err = errors.New("files differ")
_ = fs.CountError(err)
fs.Debugf(nil, "%v = %s (sum)", hashType, sumHash)
fs.Debugf(obj, "%v = %s (%v)", hashType, objHash, c.opt.Fdst)
fs.Errorf(obj, "%v", err)
c.differences.Add(1)
c.report(obj, c.opt.Differ, '*')
}
}
// HashSums represents a parsed SUM file
type HashSums map[string]string
// ParseSumFile parses a hash SUM file and returns hashes as a map
func ParseSumFile(ctx context.Context, sumFile fs.Object) (HashSums, error) {
rd, err := Open(ctx, sumFile)
if err != nil {
return nil, err
}
parser := bufio.NewReader(rd)
const maxWarn = 3
numWarn := 0
re := regexp.MustCompile(`^([^ ]+) [ *](.+)$`)
hashes := HashSums{}
for lineNo := 0; true; lineNo++ {
lineBytes, _, err := parser.ReadLine()
if err == io.EOF {
break
}
if err != nil {
return nil, err
}
line := string(lineBytes)
if line == "" {
continue
}
fields := re.FindStringSubmatch(ApplyTransforms(ctx, line))
if fields == nil {
numWarn++
if numWarn <= maxWarn {
fs.Logf(sumFile, "improperly formatted checksum line %d", lineNo)
}
continue
}
sum, file := fields[1], fields[2]
if hashes[file] != "" {
numWarn++
if numWarn <= maxWarn {
fs.Logf(sumFile, "duplicate file on checksum line %d", lineNo)
}
continue
}
// We've standardised on lower case checksums in rclone internals.
hashes[file] = strings.ToLower(sum)
}
if numWarn > maxWarn {
fs.Logf(sumFile, "%d warning(s) suppressed...", numWarn-maxWarn)
}
if err = rd.Close(); err != nil {
return nil, err
}
return hashes, nil
}