Gitealize nicknames from Blender ID

Blender ID allows a wider range of characters in the nickname than Gitea
allows. Before using a Blender ID nickname as a Gitea username, it needs
to be massaged into a valid form.

This was already done in the Blender ID-to-Gitea webhook for username
changes, and this PR introduces it in Gitea itself for the registration
flow.

The implementation follows what the webhook code[1] does, except it's
simpler because it can use built-in Gitea functionality.

This fixes https://projects.blender.org/blender/blender/issues/111937

[1]: https://projects.blender.org/infrastructure/gitea-blenderid-webhook/src/branch/main/gitea_blenderid_webhook/gitea_users.py

Reviewed on: #3

go.mod

@@ -79,6 +79,7 @@ require (
 	github.com/microcosm-cc/bluemonday v1.0.25
 	github.com/minio/minio-go/v7 v7.0.52
 	github.com/minio/sha256-simd v1.0.0
+	github.com/mozillazg/go-unidecode v0.2.0
 	github.com/msteinert/pam v1.1.0
 	github.com/nektos/act v0.2.45
 	github.com/nfnt/resize v0.0.0-20180221191011-83c6a9932646

go.sum

@@ -829,6 +829,8 @@ github.com/modern-go/reflect2 v1.0.2 h1:xBagoLtFs94CBntxluKeaWgTMpvLxC4ur3nMaC9G
 github.com/modern-go/reflect2 v1.0.2/go.mod h1:yWuevngMOJpCy52FWWMvUC8ws7m/LJsjYzDa0/r8luk=
 github.com/modocache/gover v0.0.0-20171022184752-b58185e213c5/go.mod h1:caMODM3PzxT8aQXRPkAt8xlV/e7d7w8GM5g0fa5F0D8=
 github.com/montanaflynn/stats v0.0.0-20171201202039-1bf9dbcd8cbe/go.mod h1:wL8QJuTMNUDYhXwkmfOly8iTdp5TEcJFWZD2D7SIkUc=
+github.com/mozillazg/go-unidecode v0.2.0 h1:vFGEzAH9KSwyWmXCOblazEWDh7fOkpmy/Z4ArmamSUc=
+github.com/mozillazg/go-unidecode v0.2.0/go.mod h1:zB48+/Z5toiRolOZy9ksLryJ976VIwmDmpQ2quyt1aA=
 github.com/mrjones/oauth v0.0.0-20180629183705-f4e24b6d100c/go.mod h1:skjdDftzkFALcuGzYSklqYd8gvat6F1gZJ4YPVbkZpM=
 github.com/mrjones/oauth v0.0.0-20190623134757-126b35219450 h1:j2kD3MT1z4PXCiUllUJF9mWUESr9TWKS7iEKsQ/IipM=
 github.com/mrjones/oauth v0.0.0-20190623134757-126b35219450/go.mod h1:skjdDftzkFALcuGzYSklqYd8gvat6F1gZJ4YPVbkZpM=

services/auth/source/oauth2/blenderid/blenderid.go

@@ -164,7 +164,7 @@ func userFromReader(r io.Reader, user *goth.User) error {
 	}
 	user.Email = u.Email
 	user.Name = u.Name
-	user.NickName = u.NickName
+	user.NickName = gitealizeUsername(u.NickName)
 	user.UserID = strconv.Itoa(u.ID)
 	user.AvatarURL = fmt.Sprintf("https://id.blender.org/api/user/%s/avatar", user.UserID)
 	return nil

services/auth/source/oauth2/blenderid/gitealize_usernames.go

@@ -0,0 +1,65 @@
+// Copyright 2023 The Gitea Authors. All rights reserved.
+// SPDX-License-Identifier: MIT
+package blenderid
+
+import (
+	"regexp"
+	"strings"
+
+	"code.gitea.io/gitea/models/user"
+
+	"github.com/mozillazg/go-unidecode"
+)
+
+var (
+	reInvalidCharsPattern = regexp.MustCompile(`[^\da-zA-Z.\w-]+`)
+
+	// Consecutive non-alphanumeric at start:
+	reConsPrefix = regexp.MustCompile(`^[._-]+`)
+	reConsSuffix = regexp.MustCompile(`[._-]+$`)
+	reConsInfix  = regexp.MustCompile(`[._-]{2,}`)
+)
+
+// gitealizeUsername turns a valid Blender ID nickname into a valid Gitea username.
+func gitealizeUsername(bidNickname string) string {
+	// Remove accents and other non-ASCIIness.
+	asciiUsername := unidecode.Unidecode(bidNickname)
+	asciiUsername = strings.TrimSpace(asciiUsername)
+	asciiUsername = strings.ReplaceAll(asciiUsername, " ", "_")
+
+	err := user.IsUsableUsername(asciiUsername)
+	if err == nil && len(asciiUsername) <= 40 {
+		return asciiUsername
+	}
+
+	newUsername := asciiUsername
+	newUsername = reInvalidCharsPattern.ReplaceAllString(newUsername, "_")
+	newUsername = reConsPrefix.ReplaceAllString(newUsername, "")
+	newUsername = reConsSuffix.ReplaceAllString(newUsername, "")
+	newUsername = reConsInfix.ReplaceAllStringFunc(
+		newUsername,
+		func(match string) string {
+			firstRune := []rune(match)[0]
+			return string(firstRune)
+		})
+
+	if newUsername == "" {
+		// Everything was stripped and nothing was left. Better to keep as-is and
+		// just let Gitea bork on it.
+		return asciiUsername
+	}
+
+	// This includes a test for reserved names, which are easily circumvented by
+	// appending another character.
+	if user.IsUsableUsername(newUsername) != nil {
+		if len(newUsername) > 39 {
+			return newUsername[:39] + "2"
+		}
+		return newUsername + "2"
+	}
+
+	if len(newUsername) > 40 {
+		return newUsername[:40]
+	}
+	return newUsername
+}

services/auth/source/oauth2/blenderid/gitealize_usernames_test.go

@@ -0,0 +1,43 @@
+// Copyright 2023 The Gitea Authors. All rights reserved.
+// SPDX-License-Identifier: MIT
+package blenderid
+
+import "testing"
+
+func Test_gitealizeUsername(t *testing.T) {
+	tests := []struct {
+		name        string
+		bidNickname string
+		want        string
+	}{
+		{"empty", "", ""},
+		{"underscore", "_", "_"},
+		{"reserved-name", "ghost", "ghost2"}, // Reserved name in Gitea.
+		{"short", "x", "x"},
+		{"simple", "simple", "simple"},
+		{"start-bad", "____startbad", "startbad"},
+		{"end-bad", "endbad___", "endbad"},
+		{"mid-bad-1", "mid__bad", "mid_bad"},
+		{"mid-bad-2", "user_.-name", "user_name"},
+		{"plus-mid-single", "RT2+356", "RT2_356"},
+		{"plus-mid-many", "RT2+++356", "RT2_356"},
+		{"plus-end", "RT2356+", "RT2356"},
+		{
+			"too-long", // # Max username length is 40:
+			"aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa",
+			"aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa",
+		},
+		{"accented-latin", "Ümlaut-Đenja", "Umlaut-Denja"},
+		{"thai", "แบบไทย", "aebbaithy"},
+		{"mandarin", "普通话", "Pu_Tong_Hua"},
+		{"cyrillic", "ћирилица", "tshirilitsa"},
+		{"all-bad", "------", "------"},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			if got := gitealizeUsername(tt.bidNickname); got != tt.want {
+				t.Errorf("gitealizeUsername() = %v, want %v", got, tt.want)
+			}
+		})
+	}
+}