ftps://gitea.com
`) + + t.Run("LinkEllipsis", func(t *testing.T) { + input := util.EllipsisDisplayString("http://10.1.2.3", 12) + assert.Equal(t, "http://10…", input) + test(input, "http://10…
") + + input = util.EllipsisDisplayString("http://10.1.2.3", 13) + assert.Equal(t, "http://10.…", input) + test(input, "http://10.…
") + }) } func TestRender_email(t *testing.T) { diff --git a/modules/setting/ui.go b/modules/setting/ui.go index db0fe9ef79..20fc612b43 100644 --- a/modules/setting/ui.go +++ b/modules/setting/ui.go @@ -63,6 +63,7 @@ var UI = struct { } `ini:"ui.admin"` User struct { RepoPagingNum int + OrgPagingNum int } `ini:"ui.user"` Meta struct { Author string @@ -127,8 +128,10 @@ var UI = struct { }, User: struct { RepoPagingNum int + OrgPagingNum int }{ RepoPagingNum: 15, + OrgPagingNum: 15, }, Meta: struct { Author string diff --git a/modules/templates/mailer.go b/modules/templates/mailer.go index ace81bf4a5..310d645328 100644 --- a/modules/templates/mailer.go +++ b/modules/templates/mailer.go @@ -11,9 +11,9 @@ import ( "strings" texttmpl "text/template" - "code.gitea.io/gitea/modules/base" "code.gitea.io/gitea/modules/log" "code.gitea.io/gitea/modules/setting" + "code.gitea.io/gitea/modules/util" ) var mailSubjectSplit = regexp.MustCompile(`(?m)^-{3,}\s*$`) @@ -24,7 +24,7 @@ func mailSubjectTextFuncMap() texttmpl.FuncMap { "dict": dict, "Eval": evalTokens, - "EllipsisString": base.EllipsisString, + "EllipsisString": util.EllipsisDisplayString, "AppName": func() string { return setting.AppName }, diff --git a/modules/templates/util_string.go b/modules/templates/util_string.go index 382e2de13f..683c77a870 100644 --- a/modules/templates/util_string.go +++ b/modules/templates/util_string.go @@ -8,7 +8,7 @@ import ( "html/template" "strings" - "code.gitea.io/gitea/modules/base" + "code.gitea.io/gitea/modules/util" ) type StringUtils struct{} @@ -54,7 +54,7 @@ func (su *StringUtils) Cut(s, sep string) []any { } func (su *StringUtils) EllipsisString(s string, maxLength int) string { - return base.EllipsisString(s, maxLength) + return util.EllipsisDisplayString(s, maxLength) } func (su *StringUtils) ToUpper(s string) string { diff --git a/modules/util/string.go b/modules/util/string.go index cf50f591c6..19cf75b8b3 100644 --- a/modules/util/string.go +++ b/modules/util/string.go @@ -3,7 +3,10 @@ package util -import "unsafe" +import ( + "strings" + "unsafe" +) func isSnakeCaseUpper(c byte) bool { return 'A' <= c && c <= 'Z' @@ -95,3 +98,15 @@ func UnsafeBytesToString(b []byte) string { func UnsafeStringToBytes(s string) []byte { return unsafe.Slice(unsafe.StringData(s), len(s)) } + +// SplitTrimSpace splits the string at given separator and trims leading and trailing space +func SplitTrimSpace(input, sep string) []string { + input = strings.TrimSpace(input) + var stringList []string + for _, s := range strings.Split(input, sep) { + if s = strings.TrimSpace(s); s != "" { + stringList = append(stringList, s) + } + } + return stringList +} diff --git a/modules/util/string_test.go b/modules/util/string_test.go index 0a4a8bbcfb..ff67b5c7d4 100644 --- a/modules/util/string_test.go +++ b/modules/util/string_test.go @@ -45,3 +45,8 @@ func TestToSnakeCase(t *testing.T) { assert.Equal(t, expected, ToSnakeCase(input)) } } + +func TestSplitTrimSpace(t *testing.T) { + assert.Equal(t, []string{"a", "b", "c"}, SplitTrimSpace("a\nb\nc", "\n")) + assert.Equal(t, []string{"a", "b"}, SplitTrimSpace("\r\na\n\r\nb\n\n", "\n")) +} diff --git a/modules/util/truncate.go b/modules/util/truncate.go index f2edbdc673..2bce248281 100644 --- a/modules/util/truncate.go +++ b/modules/util/truncate.go @@ -5,6 +5,7 @@ package util import ( "strings" + "unicode" "unicode/utf8" ) @@ -14,43 +15,110 @@ const ( asciiEllipsis = "..." ) -// SplitStringAtByteN splits a string at byte n accounting for rune boundaries. (Combining characters are not accounted for.) -func SplitStringAtByteN(input string, n int) (left, right string) { - if len(input) <= n { - return input, "" +func IsLikelyEllipsisLeftPart(s string) bool { + return strings.HasSuffix(s, utf8Ellipsis) || strings.HasSuffix(s, asciiEllipsis) +} + +func ellipsisGuessDisplayWidth(r rune) int { + // To make the truncated string as long as possible, + // CJK/emoji chars are considered as 2-ASCII width but not 3-4 bytes width. + // Here we only make the best guess (better than counting them in bytes), + // it's impossible to 100% correctly determine the width of a rune without a real font and render. + // + // ATTENTION: the guessed width can't be zero, more details in ellipsisDisplayString's comment + if r <= 255 { + return 1 } - if !utf8.ValidString(input) { - if n-3 < 0 { - return input, "" + switch { + case r == '\u3000': /* ideographic (CJK) characters, still use 2 */ + return 2 + case unicode.Is(unicode.M, r), /* (Mark) */ + unicode.Is(unicode.Cf, r), /* (Other, format) */ + unicode.Is(unicode.Cs, r), /* (Other, surrogate) */ + unicode.Is(unicode.Z /* (Space) */, r): + return 1 + default: + return 2 + } +} + +// EllipsisDisplayString returns a truncated short string for display purpose. +// The length is the approximate number of ASCII-width in the string (CJK/emoji are 2-ASCII width) +// It appends "…" or "..." at the end of truncated string. +// It guarantees the length of the returned runes doesn't exceed the limit. +func EllipsisDisplayString(str string, limit int) string { + s, _, _, _ := ellipsisDisplayString(str, limit) + return s +} + +// EllipsisDisplayStringX works like EllipsisDisplayString while it also returns the right part +func EllipsisDisplayStringX(str string, limit int) (left, right string) { + left, offset, truncated, encounterInvalid := ellipsisDisplayString(str, limit) + if truncated { + right = str[offset:] + r, _ := utf8.DecodeRune(UnsafeStringToBytes(right)) + encounterInvalid = encounterInvalid || r == utf8.RuneError + ellipsis := utf8Ellipsis + if encounterInvalid { + ellipsis = asciiEllipsis } - return input[:n-3] + asciiEllipsis, asciiEllipsis + input[n-3:] + right = ellipsis + right + } + return left, right +} + +func ellipsisDisplayString(str string, limit int) (res string, offset int, truncated, encounterInvalid bool) { + if len(str) <= limit { + return str, len(str), false, false } - end := 0 - for end <= n-3 { - _, size := utf8.DecodeRuneInString(input[end:]) - if end+size > n-3 { + // To future maintainers: this logic must guarantee that the length of the returned runes doesn't exceed the limit, + // because the returned string will also be used as database value. UTF-8 VARCHAR(10) could store 10 rune characters, + // So each rune must be countered as at least 1 width. + // Even if there are some special Unicode characters (zero-width, combining, etc.), they should NEVER be counted as zero. + pos, used := 0, 0 + for i, r := range str { + encounterInvalid = encounterInvalid || r == utf8.RuneError + pos = i + runeWidth := ellipsisGuessDisplayWidth(r) + if used+runeWidth+3 > limit { break } - end += size + used += runeWidth + offset += utf8.RuneLen(r) } - return input[:end] + utf8Ellipsis, utf8Ellipsis + input[end:] -} - -// SplitTrimSpace splits the string at given separator and trims leading and trailing space -func SplitTrimSpace(input, sep string) []string { - // Trim initial leading & trailing space - input = strings.TrimSpace(input) - // replace CRLF with LF - input = strings.ReplaceAll(input, "\r\n", "\n") - - var stringList []string - for _, s := range strings.Split(input, sep) { - // trim leading and trailing space - stringList = append(stringList, strings.TrimSpace(s)) + // if the remaining are fewer than 3 runes, then maybe we could add them, no need to ellipse + if len(str)-pos <= 12 { + var nextCnt, nextWidth int + for _, r := range str[pos:] { + if nextCnt >= 4 { + break + } + nextWidth += ellipsisGuessDisplayWidth(r) + nextCnt++ + } + if nextCnt <= 3 && used+nextWidth <= limit { + return str, len(str), false, false + } } - - return stringList + if limit < 3 { + // if the limit is so small, do not add ellipsis + return str[:offset], offset, true, false + } + ellipsis := utf8Ellipsis + if encounterInvalid { + ellipsis = asciiEllipsis + } + return str[:offset] + ellipsis, offset, true, encounterInvalid +} + +// TruncateRunes returns a truncated string with given rune limit, +// it returns input string if its rune length doesn't exceed the limit. +func TruncateRunes(str string, limit int) string { + if utf8.RuneCountInString(str) < limit { + return str + } + return string([]rune(str)[:limit]) } diff --git a/modules/util/truncate_test.go b/modules/util/truncate_test.go index dfe1230fd4..8789c824f5 100644 --- a/modules/util/truncate_test.go +++ b/modules/util/truncate_test.go @@ -4,43 +4,127 @@ package util import ( + "fmt" + "strings" "testing" "github.com/stretchr/testify/assert" ) -func TestSplitString(t *testing.T) { - type testCase struct { - input string - n int - leftSub string - ellipsis string +func TestEllipsisGuessDisplayWidth(t *testing.T) { + cases := []struct { + r string + want int + }{ + {r: "a", want: 1}, + {r: "é", want: 1}, + {r: "测", want: 2}, + {r: "⚽", want: 2}, + {r: "☁️", want: 3}, // 2 runes, it has a mark + {r: "\u200B", want: 1}, // ZWSP + {r: "\u3000", want: 2}, // ideographic space } - - test := func(tc []*testCase, f func(input string, n int) (left, right string)) { - for _, c := range tc { - l, r := f(c.input, c.n) - if c.ellipsis != "" { - assert.Equal(t, c.leftSub+c.ellipsis, l, "test split %q at %d, expected leftSub: %q", c.input, c.n, c.leftSub) - assert.Equal(t, c.ellipsis+c.input[len(c.leftSub):], r, "test split %s at %d, expected rightSub: %q", c.input, c.n, c.input[len(c.leftSub):]) - } else { - assert.Equal(t, c.leftSub, l, "test split %q at %d, expected leftSub: %q", c.input, c.n, c.leftSub) - assert.Empty(t, r, "test split %q at %d, expected rightSub: %q", c.input, c.n, "") + for _, c := range cases { + t.Run(c.r, func(t *testing.T) { + w := 0 + for _, r := range c.r { + w += ellipsisGuessDisplayWidth(r) } - } + assert.Equal(t, c.want, w, "hex=% x", []byte(c.r)) + }) + } +} + +func TestEllipsisString(t *testing.T) { + cases := []struct { + limit int + + input, left, right string + }{ + {limit: 0, input: "abcde", left: "", right: "…abcde"}, + {limit: 1, input: "abcde", left: "", right: "…abcde"}, + {limit: 2, input: "abcde", left: "", right: "…abcde"}, + {limit: 3, input: "abcde", left: "…", right: "…abcde"}, + {limit: 4, input: "abcde", left: "a…", right: "…bcde"}, + {limit: 5, input: "abcde", left: "abcde", right: ""}, + {limit: 6, input: "abcde", left: "abcde", right: ""}, + {limit: 7, input: "abcde", left: "abcde", right: ""}, + + // a CJK char or emoji is considered as 2-ASCII width, the ellipsis is 3-ASCII width + {limit: 0, input: "测试文本", left: "", right: "…测试文本"}, + {limit: 1, input: "测试文本", left: "", right: "…测试文本"}, + {limit: 2, input: "测试文本", left: "", right: "…测试文本"}, + {limit: 3, input: "测试文本", left: "…", right: "…测试文本"}, + {limit: 4, input: "测试文本", left: "…", right: "…测试文本"}, + {limit: 5, input: "测试文本", left: "测…", right: "…试文本"}, + {limit: 6, input: "测试文本", left: "测…", right: "…试文本"}, + {limit: 7, input: "测试文本", left: "测试…", right: "…文本"}, + {limit: 8, input: "测试文本", left: "测试文本", right: ""}, + {limit: 9, input: "测试文本", left: "测试文本", right: ""}, + + {limit: 6, input: "测试abc", left: "测…", right: "…试abc"}, + {limit: 7, input: "测试abc", left: "测试abc", right: ""}, // exactly 7-width + {limit: 8, input: "测试abc", left: "测试abc", right: ""}, + + {limit: 7, input: "测abc试啊", left: "测ab…", right: "…c试啊"}, + {limit: 8, input: "测abc试啊", left: "测abc…", right: "…试啊"}, + {limit: 9, input: "测abc试啊", left: "测abc试啊", right: ""}, // exactly 9-width + {limit: 10, input: "测abc试啊", left: "测abc试啊", right: ""}, + } + for _, c := range cases { + t.Run(fmt.Sprintf("%s(%d)", c.input, c.limit), func(t *testing.T) { + left, right := EllipsisDisplayStringX(c.input, c.limit) + assert.Equal(t, c.left, left, "left") + assert.Equal(t, c.right, right, "right") + }) } - tc := []*testCase{ - {"abc123xyz", 0, "", utf8Ellipsis}, - {"abc123xyz", 1, "", utf8Ellipsis}, - {"abc123xyz", 4, "a", utf8Ellipsis}, - {"啊bc123xyz", 4, "", utf8Ellipsis}, - {"啊bc123xyz", 6, "啊", utf8Ellipsis}, - {"啊bc", 5, "啊bc", ""}, - {"啊bc", 6, "啊bc", ""}, - {"abc\xef\x03\xfe", 3, "", asciiEllipsis}, - {"abc\xef\x03\xfe", 4, "a", asciiEllipsis}, - {"\xef\x03", 1, "\xef\x03", ""}, - } - test(tc, SplitStringAtByteN) + t.Run("LongInput", func(t *testing.T) { + left, right := EllipsisDisplayStringX(strings.Repeat("abc", 240), 90) + assert.Equal(t, strings.Repeat("abc", 29)+"…", left) + assert.Equal(t, "…"+strings.Repeat("abc", 211), right) + }) + + t.Run("InvalidUtf8", func(t *testing.T) { + invalidCases := []struct { + limit int + left, right string + }{ + {limit: 0, left: "", right: "...\xef\x03\xfe\xef\x03\xfe"}, + {limit: 1, left: "", right: "...\xef\x03\xfe\xef\x03\xfe"}, + {limit: 2, left: "", right: "...\xef\x03\xfe\xef\x03\xfe"}, + {limit: 3, left: "...", right: "...\xef\x03\xfe\xef\x03\xfe"}, + {limit: 4, left: "...", right: "...\xef\x03\xfe\xef\x03\xfe"}, + {limit: 5, left: "\xef\x03\xfe...", right: "...\xef\x03\xfe"}, + {limit: 6, left: "\xef\x03\xfe\xef\x03\xfe", right: ""}, + {limit: 7, left: "\xef\x03\xfe\xef\x03\xfe", right: ""}, + } + for _, c := range invalidCases { + t.Run(fmt.Sprintf("%d", c.limit), func(t *testing.T) { + left, right := EllipsisDisplayStringX("\xef\x03\xfe\xef\x03\xfe", c.limit) + assert.Equal(t, c.left, left, "left") + assert.Equal(t, c.right, right, "right") + }) + } + }) + + t.Run("IsLikelyEllipsisLeftPart", func(t *testing.T) { + assert.True(t, IsLikelyEllipsisLeftPart("abcde…")) + assert.True(t, IsLikelyEllipsisLeftPart("abcde...")) + }) +} + +func TestTruncateRunes(t *testing.T) { + assert.Equal(t, "", TruncateRunes("", 0)) + assert.Equal(t, "", TruncateRunes("", 1)) + + assert.Equal(t, "", TruncateRunes("ab", 0)) + assert.Equal(t, "a", TruncateRunes("ab", 1)) + assert.Equal(t, "ab", TruncateRunes("ab", 2)) + assert.Equal(t, "ab", TruncateRunes("ab", 3)) + + assert.Equal(t, "", TruncateRunes("测试", 0)) + assert.Equal(t, "测", TruncateRunes("测试", 1)) + assert.Equal(t, "测试", TruncateRunes("测试", 2)) + assert.Equal(t, "测试", TruncateRunes("测试", 3)) } diff --git a/modules/web/route.go b/modules/web/router.go similarity index 66% rename from modules/web/route.go rename to modules/web/router.go index 729ac46cdc..da06b955b1 100644 --- a/modules/web/route.go +++ b/modules/web/router.go @@ -4,18 +4,14 @@ package web import ( - "fmt" "net/http" "net/url" "reflect" - "regexp" "strings" - "code.gitea.io/gitea/modules/container" "code.gitea.io/gitea/modules/htmlutil" "code.gitea.io/gitea/modules/reqctx" "code.gitea.io/gitea/modules/setting" - "code.gitea.io/gitea/modules/util" "code.gitea.io/gitea/modules/web/middleware" "gitea.com/go-chi/binding" @@ -45,7 +41,7 @@ func GetForm(dataStore reqctx.RequestDataStore) any { // Router defines a route based on chi's router type Router struct { - chiRouter chi.Router + chiRouter *chi.Mux curGroupPrefix string curMiddlewares []any } @@ -97,16 +93,21 @@ func isNilOrFuncNil(v any) bool { return r.Kind() == reflect.Func && r.IsNil() } -func (r *Router) wrapMiddlewareAndHandler(h []any) ([]func(http.Handler) http.Handler, http.HandlerFunc) { - handlerProviders := make([]func(http.Handler) http.Handler, 0, len(r.curMiddlewares)+len(h)+1) - for _, m := range r.curMiddlewares { +func wrapMiddlewareAndHandler(curMiddlewares, h []any) ([]func(http.Handler) http.Handler, http.HandlerFunc) { + handlerProviders := make([]func(http.Handler) http.Handler, 0, len(curMiddlewares)+len(h)+1) + for _, m := range curMiddlewares { if !isNilOrFuncNil(m) { handlerProviders = append(handlerProviders, toHandlerProvider(m)) } } - for _, m := range h { + if len(h) == 0 { + panic("no endpoint handler provided") + } + for i, m := range h { if !isNilOrFuncNil(m) { handlerProviders = append(handlerProviders, toHandlerProvider(m)) + } else if i == len(h)-1 { + panic("endpoint handler can't be nil") } } middlewares := handlerProviders[:len(handlerProviders)-1] @@ -121,7 +122,7 @@ func (r *Router) wrapMiddlewareAndHandler(h []any) ([]func(http.Handler) http.Ha // Methods adds the same handlers for multiple http "methods" (separated by ","). // If any method is invalid, the lower level router will panic. func (r *Router) Methods(methods, pattern string, h ...any) { - middlewares, handlerFunc := r.wrapMiddlewareAndHandler(h) + middlewares, handlerFunc := wrapMiddlewareAndHandler(r.curMiddlewares, h) fullPattern := r.getPattern(pattern) if strings.Contains(methods, ",") { methods := strings.Split(methods, ",") @@ -141,7 +142,7 @@ func (r *Router) Mount(pattern string, subRouter *Router) { // Any delegate requests for all methods func (r *Router) Any(pattern string, h ...any) { - middlewares, handlerFunc := r.wrapMiddlewareAndHandler(h) + middlewares, handlerFunc := wrapMiddlewareAndHandler(r.curMiddlewares, h) r.chiRouter.With(middlewares...).HandleFunc(r.getPattern(pattern), handlerFunc) } @@ -185,17 +186,6 @@ func (r *Router) NotFound(h http.HandlerFunc) { r.chiRouter.NotFound(h) } -type pathProcessorParam struct { - name string - captureGroup int -} - -type PathProcessor struct { - methods container.Set[string] - re *regexp.Regexp - params []pathProcessorParam -} - func (r *Router) normalizeRequestPath(resp http.ResponseWriter, req *http.Request, next http.Handler) { normalized := false normalizedPath := req.URL.EscapedPath() @@ -253,121 +243,16 @@ func (r *Router) normalizeRequestPath(resp http.ResponseWriter, req *http.Reques next.ServeHTTP(resp, req) } -func (p *PathProcessor) ProcessRequestPath(chiCtx *chi.Context, path string) bool { - if !p.methods.Contains(chiCtx.RouteMethod) { - return false - } - if !strings.HasPrefix(path, "/") { - path = "/" + path - } - pathMatches := p.re.FindStringSubmatchIndex(path) // Golang regexp match pairs [start, end, start, end, ...] - if pathMatches == nil { - return false - } - var paramMatches [][]int - for i := 2; i < len(pathMatches); { - paramMatches = append(paramMatches, []int{pathMatches[i], pathMatches[i+1]}) - pmIdx := len(paramMatches) - 1 - end := pathMatches[i+1] - i += 2 - for ; i < len(pathMatches); i += 2 { - if pathMatches[i] >= end { - break - } - paramMatches[pmIdx] = append(paramMatches[pmIdx], pathMatches[i], pathMatches[i+1]) - } - } - for i, pm := range paramMatches { - groupIdx := p.params[i].captureGroup * 2 - chiCtx.URLParams.Add(p.params[i].name, path[pm[groupIdx]:pm[groupIdx+1]]) - } - return true -} - -func NewPathProcessor(methods, pattern string) *PathProcessor { - p := &PathProcessor{methods: make(container.Set[string])} - for _, method := range strings.Split(methods, ",") { - p.methods.Add(strings.TrimSpace(method)) - } - re := []byte{'^'} - lastEnd := 0 - for lastEnd < len(pattern) { - start := strings.IndexByte(pattern[lastEnd:], '<') - if start == -1 { - re = append(re, pattern[lastEnd:]...) - break - } - end := strings.IndexByte(pattern[lastEnd+start:], '>') - if end == -1 { - panic(fmt.Sprintf("invalid pattern: %s", pattern)) - } - re = append(re, pattern[lastEnd:lastEnd+start]...) - partName, partExp, _ := strings.Cut(pattern[lastEnd+start+1:lastEnd+start+end], ":") - lastEnd += start + end + 1 - - // TODO: it could support to specify a "capture group" for the name, for example: "/