storage/url: allow wildcards to match new line characters (peak#505)

This commit adds `s` flag [^1] to URL's filter Regexp to "let `.` match
`\n`".

[^1]:
https://github.com/google/re2/wiki/Syntax#:~:text=text%20(default%20false)-,s,-let%20.%20match

Co-authored-by: İlkin Balkanay <[email protected]>

command/exclude.go

@@ -4,23 +4,19 @@ import (
 	"path/filepath"
 	"regexp"
 	"strings"
-)
 
-func wildCardToRegexp(pattern string) string {
-	patternRegex := regexp.QuoteMeta(pattern)
-	patternRegex = strings.Replace(patternRegex, "\\?", ".", -1)
-	patternRegex = strings.Replace(patternRegex, "\\*", ".*", -1)
-	patternRegex = "^" + patternRegex + "$"
-	return patternRegex
-}
+	"github.com/peak/s5cmd/strutil"
+)
 
 // createExcludesFromWildcard creates regex strings from wildcard.
 func createExcludesFromWildcard(inputExcludes []string) ([]*regexp.Regexp, error) {
 	var result []*regexp.Regexp
 	for _, input := range inputExcludes {
 		if input != "" {
-			regexVersion := wildCardToRegexp(input)
-			regexpCompiled, err := regexp.Compile(regexVersion)
+			regex := strutil.WildCardToRegexp(input)
+			regex = strutil.MatchFromStartToEnd(regex)
+			regex = strutil.AddNewLineFlag(regex)
+			regexpCompiled, err := regexp.Compile(regex)
 			if err != nil {
 				return nil, err
 			}

e2e/ls_test.go

@@ -144,6 +144,32 @@ func TestListSingleWildcardS3Object(t *testing.T) {
 	}, alignment(true))
 }
 
+func TestListWildcardS3ObjectWithNewLineInName(t *testing.T) {
+	t.Parallel()
+
+	bucket := s3BucketFromTestName(t)
+
+	s3client, s5cmd := setup(t)
+
+	createBucket(t, s3client, bucket)
+	putFile(t, s3client, bucket, "normal.txt", "this is a file content")
+	putFile(t, s3client, bucket, "another.txt", "this is another file content")
+	putFile(t, s3client, bucket, "newli\ne.txt", "this is yet another file content")
+	putFile(t, s3client, bucket, "nap.txt", "this, too, is a file content")
+
+	cmd := s5cmd("ls", "s3://"+bucket+"/n*.txt")
+	result := icmd.RunCmd(cmd)
+
+	result.Assert(t, icmd.Success)
+
+	assertLines(t, result.Stdout(), map[int]compareFunc{
+		0: suffix("28 nap.txt"),
+		1: suffix("32 newli"),
+		2: equals("e.txt"),
+		3: suffix("22 normal.txt"),
+	})
+}
+
 // ls -s bucket/object
 func TestListS3ObjectsWithDashS(t *testing.T) {
 	t.Parallel()
@@ -399,6 +425,8 @@ func TestListS3ObjectsWithExcludeFilter(t *testing.T) {
 		"a/file.c",
 		"file2.txt",
 		"file2.txt.extension", // this should not be excluded.
+		"newli\ne",
+		"newli\ne.txt",
 	}
 
 	s3client, s5cmd := setup(t)
@@ -419,7 +447,9 @@ func TestListS3ObjectsWithExcludeFilter(t *testing.T) {
 		1: match(`a/try.py`),
 		2: match(`file.py`),
 		3: match(`file2.txt.extension`),
-	}, trimMatch(dateRe), alignment(true))
+		4: match("newli"),
+		5: match("e"),
+	}, trimMatch(dateRe), alignment(false))
 }
 
 // ls --exclude ".txt" --exclude ".py" s3://bucket

e2e/util_test.go

@@ -23,6 +23,9 @@ import (
 	"testing"
 	"time"
 
+	"github.com/peak/s5cmd/storage"
+	"github.com/peak/s5cmd/strutil"
+
 	"github.com/aws/aws-sdk-go/aws"
 	"github.com/aws/aws-sdk-go/aws/awserr"
 	"github.com/aws/aws-sdk-go/aws/client"
@@ -34,7 +37,6 @@ import (
 	"github.com/google/go-cmp/cmp"
 	"github.com/iancoleman/strcase"
 	"github.com/igungor/gofakes3"
-	"github.com/peak/s5cmd/storage"
 	"gotest.tools/v3/assert"
 	"gotest.tools/v3/fs"
 	"gotest.tools/v3/icmd"
@@ -554,7 +556,7 @@ func replaceMatchWithSpace(input string, match ...string) string {
 		if m == "" {
 			continue
 		}
-		re := regexp.MustCompile(m)
+		re := regexp.MustCompile(strutil.AddNewLineFlag(m))
 		input = re.ReplaceAllString(input, " ")
 	}
 
@@ -735,7 +737,7 @@ func alignment(v bool) func(*assertOpts) {
 }
 
 func trimMatch(match string) func(*assertOpts) {
-	re := regexp.MustCompile(match)
+	re := regexp.MustCompile(strutil.AddNewLineFlag(match))
 	return func(opts *assertOpts) {
 		opts.trimRegexes = append(opts.trimRegexes, re)
 	}
@@ -863,7 +865,7 @@ func checkLineAlignments(actual string) error {
 }
 
 func match(expected string) compareFunc {
-	re := regexp.MustCompile(expected)
+	re := regexp.MustCompile(strutil.AddNewLineFlag(expected))
 	return func(actual string) error {
 		if re.MatchString(actual) {
 			return nil

storage/url/url.go

@@ -10,6 +10,8 @@ import (
 	"regexp"
 	"runtime"
 	"strings"
+
+	"github.com/peak/s5cmd/strutil"
 )
 
 const (
@@ -249,9 +251,7 @@ func (u *URL) setPrefixAndFilter() error {
 		return nil
 	}
 
-	loc := strings.IndexAny(u.Path, globCharacters)
-	wildOperation := loc > -1
-	if !wildOperation {
+	if loc := strings.IndexAny(u.Path, globCharacters); loc < 0 {
 		u.Delimiter = s3Separator
 		u.Prefix = u.Path
 	} else {
@@ -261,12 +261,12 @@ func (u *URL) setPrefixAndFilter() error {
 
 	filterRegex := matchAllRe
 	if u.filter != "" {
-		filterRegex = regexp.QuoteMeta(u.filter)
-		filterRegex = strings.Replace(filterRegex, "\\?", ".", -1)
-		filterRegex = strings.Replace(filterRegex, "\\*", ".*?", -1)
+		filterRegex = strutil.WildCardToRegexp(u.filter)
 	}
 	filterRegex = regexp.QuoteMeta(u.Prefix) + filterRegex
-	r, err := regexp.Compile("^" + filterRegex + "$")
+	filterRegex = strutil.MatchFromStartToEnd(filterRegex)
+	filterRegex = strutil.AddNewLineFlag(filterRegex)
+	r, err := regexp.Compile(filterRegex)
 	if err != nil {
 		return err
 	}

storage/url/url_test.go

@@ -8,6 +8,7 @@ import (
 
 	"github.com/google/go-cmp/cmp"
 	"github.com/google/go-cmp/cmp/cmpopts"
+	"github.com/peak/s5cmd/strutil"
 )
 
 func TestHasWild(t *testing.T) {
@@ -70,7 +71,7 @@ func TestNew(t *testing.T) {
 				Prefix:    "key",
 				Delimiter: "/",
 			},
-			wantFilterRe: regexp.MustCompile(`^key.*$`).String(),
+			wantFilterRe: regexp.MustCompile(strutil.AddNewLineFlag(`^key.*$`)).String(),
 		},
 		{
 			name:   "url_with_no_wildcard_end_with_slash",
@@ -82,7 +83,7 @@ func TestNew(t *testing.T) {
 				Prefix:    "key/",
 				Delimiter: "/",
 			},
-			wantFilterRe: regexp.MustCompile(`^key/.*$`).String(),
+			wantFilterRe: regexp.MustCompile(strutil.AddNewLineFlag(`^key/.*$`)).String(),
 		},
 		{
 			name:   "url_with_wildcard",
@@ -92,10 +93,10 @@ func TestNew(t *testing.T) {
 				Bucket:      "bucket",
 				Path:        "key/a/?/test/*",
 				Prefix:      "key/a/",
-				filterRegex: regexp.MustCompile(`^key/a/./test/.*?$`),
+				filterRegex: regexp.MustCompile(strutil.AddNewLineFlag(`^key/a/./test/.*$`)),
 				Delimiter:   "",
 			},
-			wantFilterRe: regexp.MustCompile(`^key/a/./test/.*?$`).String(),
+			wantFilterRe: regexp.MustCompile(strutil.AddNewLineFlag(`^key/a/./test/.*$`)).String(),
 		},
 	}
 	for _, tc := range tests {
@@ -112,7 +113,6 @@ func TestNew(t *testing.T) {
 			if tc.wantFilterRe != "" {
 				if diff := cmp.Diff(tc.wantFilterRe, got.filterRegex.String()); diff != "" {
 					t.Errorf("test case %q: URL.filterRegex mismatch (-want +got):\n%v", tc.name, diff)
-
 				}
 			}
 		})
@@ -228,7 +228,7 @@ func TestURLSetPrefixAndFilter(t *testing.T) {
 				Prefix:      "a/b_c/",
 				Delimiter:   "",
 				filter:      "*/de/*/test",
-				filterRegex: regexp.MustCompile("^a/b_c/.*?/de/.*?/test$"),
+				filterRegex: regexp.MustCompile(strutil.AddNewLineFlag("^a/b_c/.*/de/.*/test$")),
 			},
 		},
 		{
@@ -241,7 +241,7 @@ func TestURLSetPrefixAndFilter(t *testing.T) {
 				Prefix:      "a/b_c/d/e",
 				Delimiter:   "/",
 				filter:      "",
-				filterRegex: regexp.MustCompile("^a/b_c/d/e.*$"),
+				filterRegex: regexp.MustCompile(strutil.AddNewLineFlag("^a/b_c/d/e.*$")),
 			},
 		},
 	}

strutil/strutil.go

@@ -3,7 +3,9 @@ package strutil
 import (
 	"encoding/json"
 	"fmt"
+	"regexp"
 	"strconv"
+	"strings"
 )
 
 var humanDivisors = [...]struct {
@@ -40,3 +42,21 @@ func JSON(v interface{}) string {
 	bytes, _ := json.Marshal(v)
 	return string(bytes)
 }
+
+// AddNewLineFlag adds a flag that allows . to match new line character "\n".
+// It assumes that the pattern does not have any flags.
+func AddNewLineFlag(pattern string) string {
+	return "(?s)" + pattern
+}
+
+// WildCardToRegexp converts a wildcarded expresiion to equivalent regular expression
+func WildCardToRegexp(pattern string) string {
+	patternRegex := regexp.QuoteMeta(pattern)
+	patternRegex = strings.Replace(patternRegex, "\\?", ".", -1)
+	return strings.Replace(patternRegex, "\\*", ".*", -1)
+}
+
+// MatchFromStartToEnd enforces that the regex will match the full string
+func MatchFromStartToEnd(pattern string) string {
+	return "^" + pattern + "$"
+}

command/exclude_test.go → strutil/strutil_test.go

@@ -1,8 +1,8 @@
-package command
+package strutil
 
 import "testing"
 
-func Test_wildCardToRegexp(t *testing.T) {
+func Test_WildCardToRegexp(t *testing.T) {
 	t.Parallel()
 	tests := []struct {
 		name    string
@@ -12,22 +12,22 @@ func Test_wildCardToRegexp(t *testing.T) {
 		{
 			name:    "main*",
 			pattern: "main*",
-			wanted:  "^main.*$",
+			wanted:  "main.*",
 		},
 		{
 			name:    "*.txt",
 			pattern: "*.txt",
-			wanted:  "^.*\\.txt$",
+			wanted:  ".*\\.txt",
 		},
 		{
 			name:    "?_main*.txt",
 			pattern: "?_main*.txt",
-			wanted:  "^._main.*\\.txt$",
+			wanted:  "._main.*\\.txt",
 		},
 	}
 	for _, tt := range tests {
 		t.Run(tt.name, func(t *testing.T) {
-			if got := wildCardToRegexp(tt.pattern); got != tt.wanted {
+			if got := WildCardToRegexp(tt.pattern); got != tt.wanted {
 				t.Errorf("wildCardToRegexp() = %v, want %v", got, tt.wanted)
 			}
 		})