[feature] Fetch + display custom emoji in statuses from remote instan…

…ces (#807)

* start implementing remote emoji fetcher

* update status where pk

* aaa

* tidy up a little

* check size limits for emojis

* thank you linter, i love you <3

* update swagger docs

* add emoji dereference test

* make emoji max sizes configurable

* normalize db.ErrAlreadyExists

docs/api/swagger.yaml

@@ -2724,7 +2724,9 @@ paths:
         pattern: \w{2,30}
         required: true
         type: string
-      - description: A png or gif image of the emoji. Animated pngs work too!
+      - description: |-
+          A png or gif image of the emoji. Animated pngs work too!
+          To ensure compatibility with other fedi implementations, emoji size limit is 50kb by default.
         in: formData
         name: image
         required: true

docs/configuration/media.md

@@ -39,4 +39,20 @@ media-description-max-chars: 500
 # Examples: [30, 60, 7, 0]
 # Default: 30
 media-remote-cache-days: 30
+
+# Int. Max size in bytes of emojis uploaded to this instance via the admin API.
+# The default is the same as the Mastodon size limit for emojis (50kb), which allows
+# for good interoperability. Raising this limit may cause issues with federation
+# of your emojis to other instances, so beware.
+# Examples: [51200, 102400]
+# Default: 51200
+media-emoji-local-max-size: 51200
+
+# Int. Max size in bytes of emojis to download from other instances.
+# By default this is 100kb, or twice the size of the default for media-emoji-local-max-size.
+# This strikes a good balance between decent interoperability with instances that have
+# higher emoji size limits, and not taking up too much space in storage.
+# Examples: [51200, 102400]
+# Default: 51200
+media-emoji-remote-max-size: 102400
 ```

example/config.yaml

@@ -211,7 +211,7 @@ accounts-reason-required: true
 ##### MEDIA CONFIG #####
 ########################
 
-# Config pertaining to user media uploads (videos, image, image descriptions).
+# Config pertaining to media uploads (videos, image, image descriptions, emoji).
 
 # Int. Maximum allowed image upload size in bytes.
 # Examples: [2097152, 10485760]
@@ -244,6 +244,22 @@ media-description-max-chars: 500
 # Default: 30
 media-remote-cache-days: 30
 
+# Int. Max size in bytes of emojis uploaded to this instance via the admin API.
+# The default is the same as the Mastodon size limit for emojis (50kb), which allows
+# for good interoperability. Raising this limit may cause issues with federation
+# of your emojis to other instances, so beware.
+# Examples: [51200, 102400]
+# Default: 51200
+media-emoji-local-max-size: 51200
+
+# Int. Max size in bytes of emojis to download from other instances.
+# By default this is 100kb, or twice the size of the default for media-emoji-local-max-size.
+# This strikes a good balance between decent interoperability with instances that have
+# higher emoji size limits, and not taking up too much space in storage.
+# Examples: [51200, 102400]
+# Default: 51200
+media-emoji-remote-max-size: 102400
+
 ##########################
 ##### STORAGE CONFIG #####
 ##########################

internal/api/client/admin/emojicreate.go

@@ -26,6 +26,7 @@ import (
 	"github.com/gin-gonic/gin"
 	"github.com/superseriousbusiness/gotosocial/internal/api"
 	"github.com/superseriousbusiness/gotosocial/internal/api/model"
+	"github.com/superseriousbusiness/gotosocial/internal/config"
 	"github.com/superseriousbusiness/gotosocial/internal/gtserror"
 	"github.com/superseriousbusiness/gotosocial/internal/oauth"
 	"github.com/superseriousbusiness/gotosocial/internal/validate"
@@ -56,7 +57,9 @@ import (
 //   required: true
 // - name: image
 //   in: formData
-//   description: A png or gif image of the emoji. Animated pngs work too!
+//   description: |-
+//     A png or gif image of the emoji. Animated pngs work too!
+//     To ensure compatibility with other fedi implementations, emoji size limit is 50kb by default.
 //   type: file
 //   required: true
 //
@@ -126,5 +129,10 @@ func validateCreateEmoji(form *model.EmojiCreateRequest) error {
 		return errors.New("no emoji given")
 	}
 
+	maxSize := config.GetMediaEmojiLocalMaxSize()
+	if form.Image.Size > int64(maxSize) {
+		return fmt.Errorf("emoji image too large: image is %dKB but size limit for custom emojis is %dKB", form.Image.Size/1024, maxSize/1024)
+	}
+
 	return validate.EmojiShortcode(form.Shortcode)
 }

internal/config/config.go

@@ -79,6 +79,8 @@ type Configuration struct {
 	MediaDescriptionMinChars int `name:"media-description-min-chars" usage:"Min required chars for an image description"`
 	MediaDescriptionMaxChars int `name:"media-description-max-chars" usage:"Max permitted chars for an image description"`
 	MediaRemoteCacheDays     int `name:"media-remote-cache-days" usage:"Number of days to locally cache media from remote instances. If set to 0, remote media will be kept indefinitely."`
+	MediaEmojiLocalMaxSize   int `name:"media-emoji-local-max-size" usage:"Max size in bytes of emojis uploaded to this instance via the admin API."`
+	MediaEmojiRemoteMaxSize  int `name:"media-emoji-remote-max-size" usage:"Max size in bytes of emojis to download from other instances."`
 
 	StorageBackend       string `name:"storage-backend" usage:"Storage backend to use for media attachments"`
 	StorageLocalBasePath string `name:"storage-local-base-path" usage:"Full path to an already-created directory where gts should store/retrieve media files. Subfolders will be created within this dir."`

internal/config/defaults.go

@@ -58,6 +58,8 @@ var Defaults = Configuration{
 	MediaDescriptionMinChars: 0,
 	MediaDescriptionMaxChars: 500,
 	MediaRemoteCacheDays:     30,
+	MediaEmojiLocalMaxSize:   51200,  // 50kb
+	MediaEmojiRemoteMaxSize:  102400, // 100kb
 
 	StorageBackend:       "local",
 	StorageLocalBasePath: "/gotosocial/storage",

internal/config/flags.go

@@ -75,6 +75,8 @@ func AddServerFlags(cmd *cobra.Command) {
 		cmd.Flags().Int(MediaDescriptionMinCharsFlag(), cfg.MediaDescriptionMinChars, fieldtag("MediaDescriptionMinChars", "usage"))
 		cmd.Flags().Int(MediaDescriptionMaxCharsFlag(), cfg.MediaDescriptionMaxChars, fieldtag("MediaDescriptionMaxChars", "usage"))
 		cmd.Flags().Int(MediaRemoteCacheDaysFlag(), cfg.MediaRemoteCacheDays, fieldtag("MediaRemoteCacheDays", "usage"))
+		cmd.Flags().Int(MediaEmojiLocalMaxSizeFlag(), cfg.MediaEmojiLocalMaxSize, fieldtag("MediaEmojiLocalMaxSize", "usage"))
+		cmd.Flags().Int(MediaEmojiRemoteMaxSizeFlag(), cfg.MediaEmojiRemoteMaxSize, fieldtag("MediaEmojiRemoteMaxSize", "usage"))
 
 		// Storage
 		cmd.Flags().String(StorageBackendFlag(), cfg.StorageBackend, fieldtag("StorageBackend", "usage"))

internal/config/helpers.gen.go

@@ -793,6 +793,56 @@ func GetMediaRemoteCacheDays() int { return global.GetMediaRemoteCacheDays() }
 // SetMediaRemoteCacheDays safely sets the value for global configuration 'MediaRemoteCacheDays' field
 func SetMediaRemoteCacheDays(v int) { global.SetMediaRemoteCacheDays(v) }
 
+// GetMediaEmojiLocalMaxSize safely fetches the Configuration value for state's 'MediaEmojiLocalMaxSize' field
+func (st *ConfigState) GetMediaEmojiLocalMaxSize() (v int) {
+	st.mutex.Lock()
+	v = st.config.MediaEmojiLocalMaxSize
+	st.mutex.Unlock()
+	return
+}
+
+// SetMediaEmojiLocalMaxSize safely sets the Configuration value for state's 'MediaEmojiLocalMaxSize' field
+func (st *ConfigState) SetMediaEmojiLocalMaxSize(v int) {
+	st.mutex.Lock()
+	defer st.mutex.Unlock()
+	st.config.MediaEmojiLocalMaxSize = v
+	st.reloadToViper()
+}
+
+// MediaEmojiLocalMaxSizeFlag returns the flag name for the 'MediaEmojiLocalMaxSize' field
+func MediaEmojiLocalMaxSizeFlag() string { return "media-emoji-local-max-size" }
+
+// GetMediaEmojiLocalMaxSize safely fetches the value for global configuration 'MediaEmojiLocalMaxSize' field
+func GetMediaEmojiLocalMaxSize() int { return global.GetMediaEmojiLocalMaxSize() }
+
+// SetMediaEmojiLocalMaxSize safely sets the value for global configuration 'MediaEmojiLocalMaxSize' field
+func SetMediaEmojiLocalMaxSize(v int) { global.SetMediaEmojiLocalMaxSize(v) }
+
+// GetMediaEmojiRemoteMaxSize safely fetches the Configuration value for state's 'MediaEmojiRemoteMaxSize' field
+func (st *ConfigState) GetMediaEmojiRemoteMaxSize() (v int) {
+	st.mutex.Lock()
+	v = st.config.MediaEmojiRemoteMaxSize
+	st.mutex.Unlock()
+	return
+}
+
+// SetMediaEmojiRemoteMaxSize safely sets the Configuration value for state's 'MediaEmojiRemoteMaxSize' field
+func (st *ConfigState) SetMediaEmojiRemoteMaxSize(v int) {
+	st.mutex.Lock()
+	defer st.mutex.Unlock()
+	st.config.MediaEmojiRemoteMaxSize = v
+	st.reloadToViper()
+}
+
+// MediaEmojiRemoteMaxSizeFlag returns the flag name for the 'MediaEmojiRemoteMaxSize' field
+func MediaEmojiRemoteMaxSizeFlag() string { return "media-emoji-remote-max-size" }
+
+// GetMediaEmojiRemoteMaxSize safely fetches the value for global configuration 'MediaEmojiRemoteMaxSize' field
+func GetMediaEmojiRemoteMaxSize() int { return global.GetMediaEmojiRemoteMaxSize() }
+
+// SetMediaEmojiRemoteMaxSize safely sets the value for global configuration 'MediaEmojiRemoteMaxSize' field
+func SetMediaEmojiRemoteMaxSize(v int) { global.SetMediaEmojiRemoteMaxSize(v) }
+
 // GetStorageBackend safely fetches the Configuration value for state's 'StorageBackend' field
 func (st *ConfigState) GetStorageBackend() (v string) {
 	st.mutex.Lock()

internal/config/validate.go

@@ -67,6 +67,14 @@ func Validate() error {
 		errs = append(errs, fmt.Errorf("%s must be set", WebAssetBaseDirFlag()))
 	}
 
+	if m := GetMediaEmojiLocalMaxSize(); m < 0 {
+		errs = append(errs, fmt.Errorf("%s must not be less than 0", MediaEmojiLocalMaxSizeFlag()))
+	}
+
+	if m := GetMediaEmojiRemoteMaxSize(); m < 0 {
+		errs = append(errs, fmt.Errorf("%s must not be less than 0", MediaEmojiRemoteMaxSizeFlag()))
+	}
+
 	if len(errs) > 0 {
 		errStrings := []string{}
 		for _, err := range errs {

internal/config/validate_test.go

@@ -141,6 +141,16 @@ func (suite *ConfigValidateTestSuite) TestValidateConfigBadProtocolNoHost() {
 	suite.EqualError(err, "host must be set; protocol must be set to either http or https, provided value was foo")
 }
 
+func (suite *ConfigValidateTestSuite) TestValidateConfigBadEmojiSizes() {
+	testrig.InitTestConfig()
+
+	config.SetMediaEmojiLocalMaxSize(-10)
+	config.SetMediaEmojiRemoteMaxSize(-50)
+
+	err := config.Validate()
+	suite.EqualError(err, "media-emoji-local-max-size must not be less than 0; media-emoji-remote-max-size must not be less than 0")
+}
+
 func TestConfigValidateTestSuite(t *testing.T) {
 	suite.Run(t, &ConfigValidateTestSuite{})
 }

internal/db/bundb/errors.go

@@ -19,7 +19,7 @@ func processPostgresError(err error) db.Error {
 	// (https://www.postgresql.org/docs/10/errcodes-appendix.html)
 	switch pgErr.Code {
 	case "23505" /* unique_violation */ :
-		return db.NewErrAlreadyExists(pgErr.Message)
+		return db.ErrAlreadyExists
 	default:
 		return err
 	}
@@ -36,7 +36,7 @@ func processSQLiteError(err error) db.Error {
 	// Handle supplied error code:
 	switch sqliteErr.Code() {
 	case sqlite3.SQLITE_CONSTRAINT_UNIQUE, sqlite3.SQLITE_CONSTRAINT_PRIMARYKEY:
-		return db.NewErrAlreadyExists(err.Error())
+		return db.ErrAlreadyExists
 	default:
 		return err
 	}

internal/db/bundb/status.go

@@ -22,6 +22,7 @@ import (
 	"container/list"
 	"context"
 	"database/sql"
+	"errors"
 	"time"
 
 	"github.com/superseriousbusiness/gotosocial/internal/cache"
@@ -175,6 +176,57 @@ func (s *statusDB) PutStatus(ctx context.Context, status *gtsmodel.Status) db.Er
 	})
 }
 
+func (s *statusDB) UpdateStatus(ctx context.Context, status *gtsmodel.Status) (*gtsmodel.Status, db.Error) {
+	err := s.conn.RunInTx(ctx, func(tx bun.Tx) error {
+		// create links between this status and any emojis it uses
+		for _, i := range status.EmojiIDs {
+			if _, err := tx.NewInsert().Model(&gtsmodel.StatusToEmoji{
+				StatusID: status.ID,
+				EmojiID:  i,
+			}).Exec(ctx); err != nil {
+				err = s.conn.errProc(err)
+				if !errors.Is(err, db.ErrAlreadyExists) {
+					return err
+				}
+			}
+		}
+
+		// create links between this status and any tags it uses
+		for _, i := range status.TagIDs {
+			if _, err := tx.NewInsert().Model(&gtsmodel.StatusToTag{
+				StatusID: status.ID,
+				TagID:    i,
+			}).Exec(ctx); err != nil {
+				err = s.conn.errProc(err)
+				if !errors.Is(err, db.ErrAlreadyExists) {
+					return err
+				}
+			}
+		}
+
+		// change the status ID of the media attachments to this status
+		for _, a := range status.Attachments {
+			a.StatusID = status.ID
+			a.UpdatedAt = time.Now()
+			if _, err := tx.NewUpdate().Model(a).
+				Where("id = ?", a.ID).
+				Exec(ctx); err != nil {
+				return err
+			}
+		}
+
+		// Finally, update the status itself
+		if _, err := tx.NewUpdate().Model(status).WherePK().Exec(ctx); err != nil {
+			return err
+		}
+
+		s.cache.Put(status)
+		return nil
+	})
+
+	return status, err
+}
+
 func (s *statusDB) GetStatusParents(ctx context.Context, status *gtsmodel.Status, onlyDirect bool) ([]*gtsmodel.Status, db.Error) {
 	parents := []*gtsmodel.Status{}
 	s.statusParent(ctx, status, &parents, onlyDirect)

internal/db/emoji.go

@@ -35,4 +35,6 @@ type Emoji interface {
 	// GetEmojiByShortcodeDomain gets an emoji based on its shortcode and domain.
 	// For local emoji, domain should be an empty string.
 	GetEmojiByShortcodeDomain(ctx context.Context, shortcode string, domain string) (*gtsmodel.Emoji, Error)
+	// GetEmojiByURI returns one emoji based on its ActivityPub URI.
+	GetEmojiByURI(ctx context.Context, uri string) (*gtsmodel.Emoji, Error)
 }

internal/db/error.go

@@ -28,19 +28,8 @@ var (
 	ErrNoEntries Error = fmt.Errorf("no entries")
 	// ErrMultipleEntries is returned when a caller expected ONE entry for a query, but multiples were found.
 	ErrMultipleEntries Error = fmt.Errorf("multiple entries")
+	// ErrAlreadyExists is returned when a conflict was encountered in the db when doing an insert.
+	ErrAlreadyExists Error = fmt.Errorf("already exists")
 	// ErrUnknown denotes an unknown database error.
 	ErrUnknown Error = fmt.Errorf("unknown error")
 )
-
-// ErrAlreadyExists is returned when a caller tries to insert a database entry that already exists in the db.
-type ErrAlreadyExists struct {
-	message string
-}
-
-func (e *ErrAlreadyExists) Error() string {
-	return e.message
-}
-
-func NewErrAlreadyExists(msg string) error {
-	return &ErrAlreadyExists{message: msg}
-}

internal/db/status.go

@@ -38,6 +38,9 @@ type Status interface {
 	// PutStatus stores one status in the database.
 	PutStatus(ctx context.Context, status *gtsmodel.Status) Error
 
+	// UpdateStatus updates one status in the database and returns it to the caller.
+	UpdateStatus(ctx context.Context, status *gtsmodel.Status) (*gtsmodel.Status, Error)
+
 	// CountStatusReplies returns the amount of replies recorded for a status, or an error if something goes wrong
 	CountStatusReplies(ctx context.Context, status *gtsmodel.Status) (int, Error)
 

internal/federation/dereferencing/dereferencer.go

@@ -41,6 +41,7 @@ type Dereferencer interface {
 	GetRemoteInstance(ctx context.Context, username string, remoteInstanceURI *url.URL) (*gtsmodel.Instance, error)
 
 	GetRemoteMedia(ctx context.Context, requestingUsername string, accountID string, remoteURL string, ai *media.AdditionalMediaInfo) (*media.ProcessingMedia, error)
+	GetRemoteEmoji(ctx context.Context, requestingUsername string, remoteURL string, shortcode string, id string, emojiURI string, ai *media.AdditionalEmojiInfo) (*media.ProcessingEmoji, error)
 
 	DereferenceAnnounce(ctx context.Context, announce *gtsmodel.Status, requestingUsername string) error
 	DereferenceThread(ctx context.Context, username string, statusIRI *url.URL) error

internal/federation/dereferencing/emoji.go

@@ -0,0 +1,51 @@
+/*
+   GoToSocial
+   Copyright (C) 2021-2022 GoToSocial Authors admin@gotosocial.org
+
+   This program is free software: you can redistribute it and/or modify
+   it under the terms of the GNU Affero General Public License as published by
+   the Free Software Foundation, either version 3 of the License, or
+   (at your option) any later version.
+
+   This program is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+   GNU Affero General Public License for more details.
+
+   You should have received a copy of the GNU Affero General Public License
+   along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+package dereferencing
+
+import (
+	"context"
+	"fmt"
+	"io"
+	"net/url"
+
+	"github.com/superseriousbusiness/gotosocial/internal/media"
+)
+
+func (d *deref) GetRemoteEmoji(ctx context.Context, requestingUsername string, remoteURL string, shortcode string, id string, emojiURI string, ai *media.AdditionalEmojiInfo) (*media.ProcessingEmoji, error) {
+	t, err := d.transportController.NewTransportForUsername(ctx, requestingUsername)
+	if err != nil {
+		return nil, fmt.Errorf("GetRemoteEmoji: error creating transport: %s", err)
+	}
+
+	derefURI, err := url.Parse(remoteURL)
+	if err != nil {
+		return nil, fmt.Errorf("GetRemoteEmoji: error parsing url: %s", err)
+	}
+
+	dataFunc := func(innerCtx context.Context) (io.Reader, int, error) {
+		return t.DereferenceMedia(innerCtx, derefURI)
+	}
+
+	processingMedia, err := d.mediaManager.ProcessEmoji(ctx, dataFunc, nil, shortcode, id, emojiURI, ai)
+	if err != nil {
+		return nil, fmt.Errorf("GetRemoteEmoji: error processing emoji: %s", err)
+	}
+
+	return processingMedia, nil
+}