[MM-24526] Filter * characters from the search terms in DB (#14884)

store/searchtest/post_layer.go

@@ -126,7 +126,7 @@ var searchPostStoreTests = []searchTest{
 	{
 		Name: "Should discard a wildcard if it's not placed immediately by text",
 		Fn:   testSearchDiscardWildcardAlone,
-		Tags: []string{ENGINE_ELASTICSEARCH},
+		Tags: []string{ENGINE_POSTGRES, ENGINE_MYSQL, ENGINE_ELASTICSEARCH},
 	},
 	{
 		Name: "Should support terms with dash",

store/sqlstore/post_store.go

@@ -1750,11 +1750,9 @@ func (s *SqlPostStore) SearchPostsInTeamForUser(paramsList []*model.SearchParams
 	pchan := make(chan store.StoreResult, len(paramsList))
 
 	for _, params := range paramsList {
-		// Don't allow users to search for everything.
+		// remove any unquoted term that contains only non-alphanumeric chars
-		if params.Terms == "*" {
+		// ex: abcd "**" && abc     >>     abcd "**" abc
-			continue
+		params.Terms = removeNonAlphaNumericUnquotedTerms(params.Terms, " ")
-		}
-
 		params.IncludeDeletedChannels = includeDeletedChannels
 		params.OrTerms = isOrSearch
 

store/sqlstore/utils.go

@@ -7,6 +7,7 @@ import (
 	"database/sql"
 	"strconv"
 	"strings"
+	"unicode"
 
 	"github.com/mattermost/gorp"
 	"github.com/mattermost/mattermost-server/v5/mlog"
@@ -52,3 +53,38 @@ func finalizeTransaction(transaction *gorp.Transaction) {
 		mlog.Error("Failed to rollback transaction", mlog.Err(err))
 	}
 }
+
+// removeNonAlphaNumericUnquotedTerms removes all unquoted words that only contain
+// non-alphanumeric chars from given line
+func removeNonAlphaNumericUnquotedTerms(line, separator string) string {
+	words := strings.Split(line, separator)
+	filteredResult := make([]string, 0, len(words))
+
+	for _, w := range words {
+		if isQuotedWord(w) || containsAlphaNumericChar(w) {
+			filteredResult = append(filteredResult, strings.TrimSpace(w))
+		}
+	}
+	return strings.Join(filteredResult, separator)
+}
+
+// containsAlphaNumericChar returns true in case any letter or digit is present, false otherwise
+func containsAlphaNumericChar(s string) bool {
+	for _, r := range s {
+		if unicode.IsLetter(r) || unicode.IsDigit(r) {
+			return true
+		}
+	}
+	return false
+}
+
+// isQuotedWord return true if the input string is quoted, false otherwise. Ex :-
+// 		"quoted string"  -  will return true
+// 		unquoted string  -  will return false
+func isQuotedWord(s string) bool {
+	if len(s) < 2 {
+		return false
+	}
+
+	return s[0] == '"' && s[len(s)-1] == '"'
+}

store/sqlstore/utils_test.go

@@ -76,3 +76,30 @@ func TestSanitizeSearchTerm(t *testing.T) {
 	result = sanitizeSearchTerm(term, "*")
 	require.Equal(t, result, expected)
 }
+
+func TestRemoveNonAlphaNumericUnquotedTerms(t *testing.T) {
+	const (
+		sep           = " "
+		chineseHello  = "你好"
+		japaneseHello = "こんにちは"
+	)
+	tests := []struct {
+		term string
+		want string
+		name string
+	}{
+		{term: "", want: "", name: "empty"},
+		{term: "h", want: "h", name: "singleChar"},
+		{term: "hello", want: "hello", name: "multiChar"},
+		{term: `hel*lo "**" **& hello`, want: `hel*lo "**" hello`, name: "quoted_unquoted_english"},
+		{term: japaneseHello + chineseHello, want: japaneseHello + chineseHello, name: "japanese_chinese"},
+		{term: japaneseHello + ` "*" ` + chineseHello, want: japaneseHello + ` "*" ` + chineseHello, name: `quoted_japanese_and_chinese`},
+		{term: japaneseHello + ` "*" &&* ` + chineseHello, want: japaneseHello + ` "*" ` + chineseHello, name: "quoted_unquoted_japanese_and_chinese"},
+	}
+	for _, test := range tests {
+		t.Run(test.name, func(t *testing.T) {
+			got := removeNonAlphaNumericUnquotedTerms(test.term, sep)
+			require.Equal(t, test.want, got)
+		})
+	}
+}