refactor(favicon): improve cache and error handling

This commit is contained in:
yusing
2025-10-05 20:37:27 +08:00
parent f4f104d206
commit 6051f75145
11 changed files with 109 additions and 239 deletions

View File

@@ -21,7 +21,7 @@ require (
github.com/stretchr/testify v1.11.1 github.com/stretchr/testify v1.11.1
github.com/yusing/godoxy v0.18.6 github.com/yusing/godoxy v0.18.6
github.com/yusing/godoxy/socketproxy v0.0.0-00010101000000-000000000000 github.com/yusing/godoxy/socketproxy v0.0.0-00010101000000-000000000000
github.com/yusing/goutils v0.5.2 github.com/yusing/goutils v0.6.0
) )
require ( require (
@@ -31,6 +31,7 @@ require (
github.com/andybalholm/cascadia v1.3.3 // indirect github.com/andybalholm/cascadia v1.3.3 // indirect
github.com/bytedance/gopkg v0.1.3 // indirect github.com/bytedance/gopkg v0.1.3 // indirect
github.com/bytedance/sonic/loader v0.3.0 // indirect github.com/bytedance/sonic/loader v0.3.0 // indirect
github.com/cenkalti/backoff/v5 v5.0.3 // indirect
github.com/cloudwego/base64x v0.1.6 // indirect github.com/cloudwego/base64x v0.1.6 // indirect
github.com/containerd/errdefs v1.0.0 // indirect github.com/containerd/errdefs v1.0.0 // indirect
github.com/containerd/errdefs/pkg v0.3.0 // indirect github.com/containerd/errdefs/pkg v0.3.0 // indirect

View File

@@ -208,8 +208,8 @@ github.com/yusing/ds v0.2.0 h1:lPhDU5eA2uvquVrBrzLCrQXRJJgSXlUYA53TbuK2sQY=
github.com/yusing/ds v0.2.0/go.mod h1:XhKV4l7cZwBbbl7lRzNC9zX27zvCM0frIwiuD40ULRk= github.com/yusing/ds v0.2.0/go.mod h1:XhKV4l7cZwBbbl7lRzNC9zX27zvCM0frIwiuD40ULRk=
github.com/yusing/gointernals v0.1.16 h1:GrhZZdxzA+jojLEqankctJrOuAYDb7kY1C93S1pVR34= github.com/yusing/gointernals v0.1.16 h1:GrhZZdxzA+jojLEqankctJrOuAYDb7kY1C93S1pVR34=
github.com/yusing/gointernals v0.1.16/go.mod h1:B/0FVXt4WPmgzVy3ynzkqKi+BSGaJVmwCJBRXYapo34= github.com/yusing/gointernals v0.1.16/go.mod h1:B/0FVXt4WPmgzVy3ynzkqKi+BSGaJVmwCJBRXYapo34=
github.com/yusing/goutils v0.5.2 h1:h3Xbz3buTqH3SaL3LPdd/4wgNIQXcnHjNvCIlhOr5B0= github.com/yusing/goutils v0.6.0 h1:oXdqKlTzaWWzjBPoAeqMVPCrQ4VHLtGMU+VqIS1U02Q=
github.com/yusing/goutils v0.5.2/go.mod h1:3dgYe/A3+8wT88/iAHwXdL44q5bP+qVo2WAOiPBqOrg= github.com/yusing/goutils v0.6.0/go.mod h1:3dgYe/A3+8wT88/iAHwXdL44q5bP+qVo2WAOiPBqOrg=
github.com/yusufpapurcu/wmi v1.2.4 h1:zFUKzehAFReQwLys1b/iSMl+JQGSCSjtVqQn9bBrPo0= github.com/yusufpapurcu/wmi v1.2.4 h1:zFUKzehAFReQwLys1b/iSMl+JQGSCSjtVqQn9bBrPo0=
github.com/yusufpapurcu/wmi v1.2.4/go.mod h1:SBZ9tNy3G9/m5Oi98Zks0QjeHVDvuK0qfxQmPyzfmi0= github.com/yusufpapurcu/wmi v1.2.4/go.mod h1:SBZ9tNy3G9/m5Oi98Zks0QjeHVDvuK0qfxQmPyzfmi0=
go.opentelemetry.io/auto/sdk v1.2.1 h1:jXsnJ4Lmnqd11kwkBV2LgLoFMZKizbCi5fNZ/ipaZ64= go.opentelemetry.io/auto/sdk v1.2.1 h1:jXsnJ4Lmnqd11kwkBV2LgLoFMZKizbCi5fNZ/ipaZ64=

3
go.mod
View File

@@ -49,7 +49,7 @@ require (
github.com/yusing/godoxy/agent v0.0.0-20251005042105-448a2fbd6f50 github.com/yusing/godoxy/agent v0.0.0-20251005042105-448a2fbd6f50
github.com/yusing/godoxy/internal/dnsproviders v0.0.0-20251005040558-74224c8e8784 github.com/yusing/godoxy/internal/dnsproviders v0.0.0-20251005040558-74224c8e8784
github.com/yusing/godoxy/internal/utils v0.1.0 github.com/yusing/godoxy/internal/utils v0.1.0
github.com/yusing/goutils v0.5.2 github.com/yusing/goutils v0.6.0
) )
require ( require (
@@ -217,6 +217,7 @@ require (
github.com/aziontech/azionapi-go-sdk v0.143.0 // indirect github.com/aziontech/azionapi-go-sdk v0.143.0 // indirect
github.com/bytedance/gopkg v0.1.3 // indirect github.com/bytedance/gopkg v0.1.3 // indirect
github.com/bytedance/sonic/loader v0.3.0 // indirect github.com/bytedance/sonic/loader v0.3.0 // indirect
github.com/cenkalti/backoff/v5 v5.0.3 // indirect
github.com/cloudwego/base64x v0.1.6 // indirect github.com/cloudwego/base64x v0.1.6 // indirect
github.com/containerd/errdefs v1.0.0 // indirect github.com/containerd/errdefs v1.0.0 // indirect
github.com/containerd/errdefs/pkg v0.3.0 // indirect github.com/containerd/errdefs/pkg v0.3.0 // indirect

4
go.sum
View File

@@ -1646,8 +1646,8 @@ github.com/yusing/ds v0.2.0 h1:lPhDU5eA2uvquVrBrzLCrQXRJJgSXlUYA53TbuK2sQY=
github.com/yusing/ds v0.2.0/go.mod h1:XhKV4l7cZwBbbl7lRzNC9zX27zvCM0frIwiuD40ULRk= github.com/yusing/ds v0.2.0/go.mod h1:XhKV4l7cZwBbbl7lRzNC9zX27zvCM0frIwiuD40ULRk=
github.com/yusing/gointernals v0.1.16 h1:GrhZZdxzA+jojLEqankctJrOuAYDb7kY1C93S1pVR34= github.com/yusing/gointernals v0.1.16 h1:GrhZZdxzA+jojLEqankctJrOuAYDb7kY1C93S1pVR34=
github.com/yusing/gointernals v0.1.16/go.mod h1:B/0FVXt4WPmgzVy3ynzkqKi+BSGaJVmwCJBRXYapo34= github.com/yusing/gointernals v0.1.16/go.mod h1:B/0FVXt4WPmgzVy3ynzkqKi+BSGaJVmwCJBRXYapo34=
github.com/yusing/goutils v0.5.2 h1:h3Xbz3buTqH3SaL3LPdd/4wgNIQXcnHjNvCIlhOr5B0= github.com/yusing/goutils v0.6.0 h1:oXdqKlTzaWWzjBPoAeqMVPCrQ4VHLtGMU+VqIS1U02Q=
github.com/yusing/goutils v0.5.2/go.mod h1:3dgYe/A3+8wT88/iAHwXdL44q5bP+qVo2WAOiPBqOrg= github.com/yusing/goutils v0.6.0/go.mod h1:3dgYe/A3+8wT88/iAHwXdL44q5bP+qVo2WAOiPBqOrg=
github.com/yusufpapurcu/wmi v1.2.4 h1:zFUKzehAFReQwLys1b/iSMl+JQGSCSjtVqQn9bBrPo0= github.com/yusufpapurcu/wmi v1.2.4 h1:zFUKzehAFReQwLys1b/iSMl+JQGSCSjtVqQn9bBrPo0=
github.com/yusufpapurcu/wmi v1.2.4/go.mod h1:SBZ9tNy3G9/m5Oi98Zks0QjeHVDvuK0qfxQmPyzfmi0= github.com/yusufpapurcu/wmi v1.2.4/go.mod h1:SBZ9tNy3G9/m5Oi98Zks0QjeHVDvuK0qfxQmPyzfmi0=
github.com/zeebo/assert v1.3.0/go.mod h1:Pq9JiuJQpG8JLJdtkwrJESF0Foym2/D9XMU5ciN/wJ0= github.com/zeebo/assert v1.3.0/go.mod h1:Pq9JiuJQpG8JLJdtkwrJESF0Foym2/D9XMU5ciN/wJ0=

View File

@@ -44,9 +44,9 @@ func FavIcon(c *gin.Context) {
c.JSON(http.StatusBadRequest, apitypes.Error("invalid url", err)) c.JSON(http.StatusBadRequest, apitypes.Error("invalid url", err))
return return
} }
fetchResult := homepage.FetchFavIconFromURL(c.Request.Context(), &iconURL) fetchResult, err := homepage.FetchFavIconFromURL(c.Request.Context(), &iconURL)
if !fetchResult.OK() { if err != nil {
c.JSON(fetchResult.StatusCode, apitypes.Error(fetchResult.ErrMsg)) homepage.GinFetchError(c, fetchResult.StatusCode, err)
return return
} }
c.Data(fetchResult.StatusCode, fetchResult.ContentType(), fetchResult.Icon) c.Data(fetchResult.StatusCode, fetchResult.ContentType(), fetchResult.Icon)
@@ -54,38 +54,38 @@ func FavIcon(c *gin.Context) {
} }
// try with alias // try with alias
result := GetFavIconFromAlias(c.Request.Context(), request.Alias) result, err := GetFavIconFromAlias(c.Request.Context(), request.Alias)
if !result.OK() { if err != nil {
c.JSON(result.StatusCode, apitypes.Error(result.ErrMsg)) homepage.GinFetchError(c, result.StatusCode, err)
return return
} }
c.Data(result.StatusCode, result.ContentType(), result.Icon) c.Data(result.StatusCode, result.ContentType(), result.Icon)
} }
func GetFavIconFromAlias(ctx context.Context, alias string) *homepage.FetchResult { func GetFavIconFromAlias(ctx context.Context, alias string) (homepage.FetchResult, error) {
// try with route.Icon // try with route.Icon
r, ok := routes.HTTP.Get(alias) r, ok := routes.HTTP.Get(alias)
if !ok { if !ok {
return &homepage.FetchResult{ return homepage.FetchResultWithErrorf(http.StatusNotFound, "route not found")
StatusCode: http.StatusNotFound,
ErrMsg: "route not found",
}
} }
var result *homepage.FetchResult var (
result homepage.FetchResult
err error
)
hp := r.HomepageItem() hp := r.HomepageItem()
if hp.Icon != nil { if hp.Icon != nil {
if hp.Icon.IconSource == homepage.IconSourceRelative { if hp.Icon.IconSource == homepage.IconSourceRelative {
result = homepage.FindIcon(ctx, r, *hp.Icon.FullURL) result, err = homepage.FindIcon(ctx, r, *hp.Icon.FullURL)
} else { } else {
result = homepage.FetchFavIconFromURL(ctx, hp.Icon) result, err = homepage.FetchFavIconFromURL(ctx, hp.Icon)
} }
} else { } else {
// try extract from "link[rel=icon]" // try extract from "link[rel=icon]"
result = homepage.FindIcon(ctx, r, "/") result, err = homepage.FindIcon(ctx, r, "/")
} }
if result.StatusCode == 0 { if result.StatusCode == 0 {
result.StatusCode = http.StatusOK result.StatusCode = http.StatusOK
} }
return result return result, err
} }

View File

@@ -155,7 +155,7 @@ require (
github.com/youmark/pkcs8 v0.0.0-20240726163527-a2c0da244d78 // indirect github.com/youmark/pkcs8 v0.0.0-20240726163527-a2c0da244d78 // indirect
github.com/yusing/godoxy/internal/utils v0.1.0 // indirect github.com/yusing/godoxy/internal/utils v0.1.0 // indirect
github.com/yusing/gointernals v0.1.16 // indirect github.com/yusing/gointernals v0.1.16 // indirect
github.com/yusing/goutils v0.5.2 // indirect github.com/yusing/goutils v0.6.0 // indirect
go.opentelemetry.io/auto/sdk v1.2.1 // indirect go.opentelemetry.io/auto/sdk v1.2.1 // indirect
go.opentelemetry.io/contrib/instrumentation/net/http/otelhttp v0.63.0 // indirect go.opentelemetry.io/contrib/instrumentation/net/http/otelhttp v0.63.0 // indirect
go.opentelemetry.io/otel v1.38.0 // indirect go.opentelemetry.io/otel v1.38.0 // indirect

View File

@@ -1527,8 +1527,8 @@ github.com/yuin/goldmark v1.4.1/go.mod h1:mwnBkeHKe2W/ZEtQ+71ViKU8L12m81fl3OWwC1
github.com/yuin/goldmark v1.4.13/go.mod h1:6yULJ656Px+3vBD8DxQVa3kxgyrAnzto9xy5taEt/CY= github.com/yuin/goldmark v1.4.13/go.mod h1:6yULJ656Px+3vBD8DxQVa3kxgyrAnzto9xy5taEt/CY=
github.com/yusing/gointernals v0.1.16 h1:GrhZZdxzA+jojLEqankctJrOuAYDb7kY1C93S1pVR34= github.com/yusing/gointernals v0.1.16 h1:GrhZZdxzA+jojLEqankctJrOuAYDb7kY1C93S1pVR34=
github.com/yusing/gointernals v0.1.16/go.mod h1:B/0FVXt4WPmgzVy3ynzkqKi+BSGaJVmwCJBRXYapo34= github.com/yusing/gointernals v0.1.16/go.mod h1:B/0FVXt4WPmgzVy3ynzkqKi+BSGaJVmwCJBRXYapo34=
github.com/yusing/goutils v0.5.2 h1:h3Xbz3buTqH3SaL3LPdd/4wgNIQXcnHjNvCIlhOr5B0= github.com/yusing/goutils v0.6.0 h1:oXdqKlTzaWWzjBPoAeqMVPCrQ4VHLtGMU+VqIS1U02Q=
github.com/yusing/goutils v0.5.2/go.mod h1:3dgYe/A3+8wT88/iAHwXdL44q5bP+qVo2WAOiPBqOrg= github.com/yusing/goutils v0.6.0/go.mod h1:3dgYe/A3+8wT88/iAHwXdL44q5bP+qVo2WAOiPBqOrg=
github.com/zeebo/assert v1.3.0/go.mod h1:Pq9JiuJQpG8JLJdtkwrJESF0Foym2/D9XMU5ciN/wJ0= github.com/zeebo/assert v1.3.0/go.mod h1:Pq9JiuJQpG8JLJdtkwrJESF0Foym2/D9XMU5ciN/wJ0=
github.com/zeebo/xxh3 v1.0.2/go.mod h1:5NWz9Sef7zIDm2JHfFlcQvNekmcEl9ekUZQQKCYaDcA= github.com/zeebo/xxh3 v1.0.2/go.mod h1:5NWz9Sef7zIDm2JHfFlcQvNekmcEl9ekUZQQKCYaDcA=
go.etcd.io/bbolt v1.3.2/go.mod h1:IbVyRI1SCnLcuJnV2u8VeU0CEYM7e686BmAb1XKL+uU= go.etcd.io/bbolt v1.3.2/go.mod h1:IbVyRI1SCnLcuJnV2u8VeU0CEYM7e686BmAb1XKL+uU=

View File

@@ -4,6 +4,7 @@ import (
"bytes" "bytes"
"context" "context"
"errors" "errors"
"fmt"
"io" "io"
"net/http" "net/http"
"net/url" "net/url"
@@ -12,8 +13,11 @@ import (
"time" "time"
"github.com/PuerkitoBio/goquery" "github.com/PuerkitoBio/goquery"
"github.com/gin-gonic/gin"
"github.com/vincent-petithory/dataurl" "github.com/vincent-petithory/dataurl"
apitypes "github.com/yusing/godoxy/internal/api/types"
gphttp "github.com/yusing/godoxy/internal/net/gphttp" gphttp "github.com/yusing/godoxy/internal/net/gphttp"
"github.com/yusing/goutils/cache"
httputils "github.com/yusing/goutils/http" httputils "github.com/yusing/goutils/http"
strutils "github.com/yusing/goutils/strings" strutils "github.com/yusing/goutils/strings"
) )
@@ -21,17 +25,31 @@ import (
type FetchResult struct { type FetchResult struct {
Icon []byte Icon []byte
StatusCode int StatusCode int
ErrMsg string
contentType string contentType string
} }
const faviconFetchTimeout = 3 * time.Second func FetchResultWithErrorf(statusCode int, msgFmt string, args ...any) (FetchResult, error) {
return FetchResult{StatusCode: statusCode}, fmt.Errorf(msgFmt, args...)
func (res *FetchResult) OK() bool {
return len(res.Icon) > 0
} }
func FetchResultOK(icon []byte, contentType string) (FetchResult, error) {
return FetchResult{Icon: icon, contentType: contentType}, nil
}
func GinFetchError(c *gin.Context, statusCode int, err error) {
if statusCode == 0 {
statusCode = http.StatusInternalServerError
}
if statusCode == http.StatusInternalServerError {
c.Error(apitypes.InternalServerError(err, "unexpected error"))
} else {
c.JSON(statusCode, apitypes.Error(err.Error()))
}
}
const faviconFetchTimeout = 3 * time.Second
func (res *FetchResult) ContentType() string { func (res *FetchResult) ContentType() string {
if res.contentType == "" { if res.contentType == "" {
if bytes.HasPrefix(res.Icon, []byte("<svg")) || bytes.HasPrefix(res.Icon, []byte("<?xml")) { if bytes.HasPrefix(res.Icon, []byte("<svg")) || bytes.HasPrefix(res.Icon, []byte("<?xml")) {
@@ -44,56 +62,54 @@ func (res *FetchResult) ContentType() string {
const maxRedirectDepth = 5 const maxRedirectDepth = 5
func FetchFavIconFromURL(ctx context.Context, iconURL *IconURL) *FetchResult { func FetchFavIconFromURL(ctx context.Context, iconURL *IconURL) (FetchResult, error) {
switch iconURL.IconSource { switch iconURL.IconSource {
case IconSourceAbsolute: case IconSourceAbsolute:
return fetchIconAbsolute(ctx, iconURL.URL()) return FetchIconAbsolute(ctx, iconURL.URL())
case IconSourceRelative: case IconSourceRelative:
return &FetchResult{StatusCode: http.StatusBadRequest, ErrMsg: "unexpected relative icon"} return FetchResultWithErrorf(http.StatusBadRequest, "unexpected relative icon")
case IconSourceWalkXCode, IconSourceSelfhSt: case IconSourceWalkXCode, IconSourceSelfhSt:
return fetchKnownIcon(ctx, iconURL) return fetchKnownIcon(ctx, iconURL)
} }
return &FetchResult{StatusCode: http.StatusBadRequest, ErrMsg: "invalid icon source"} return FetchResultWithErrorf(http.StatusBadRequest, "invalid icon source")
} }
func fetchIconAbsolute(ctx context.Context, url string) *FetchResult { var FetchIconAbsolute = cache.NewKeyFunc(func(ctx context.Context, url string) (FetchResult, error) {
if result := loadIconCache(url); result != nil {
return result
}
req, err := http.NewRequestWithContext(ctx, http.MethodGet, url, nil) req, err := http.NewRequestWithContext(ctx, http.MethodGet, url, nil)
if err != nil { if err != nil {
if errors.Is(err, context.DeadlineExceeded) || errors.Is(err, context.Canceled) { return FetchResultWithErrorf(http.StatusInternalServerError, "cannot create request: %w", err)
return &FetchResult{StatusCode: http.StatusBadGateway, ErrMsg: "request timeout"}
}
return &FetchResult{StatusCode: http.StatusInternalServerError, ErrMsg: err.Error()}
} }
resp, err := gphttp.Do(req) resp, err := gphttp.Do(req)
if err == nil { if err == nil {
defer resp.Body.Close() defer resp.Body.Close()
} else {
if errors.Is(err, context.DeadlineExceeded) || errors.Is(err, context.Canceled) {
return FetchResultWithErrorf(http.StatusBadGateway, "request timeout")
}
return FetchResultWithErrorf(http.StatusBadGateway, "connection error: %w", err)
} }
if err != nil || resp.StatusCode != http.StatusOK {
return &FetchResult{StatusCode: http.StatusBadGateway, ErrMsg: "connection error"} if resp.StatusCode != http.StatusOK {
return FetchResultWithErrorf(resp.StatusCode, "upstream error: http %d", resp.StatusCode)
} }
icon, err := io.ReadAll(resp.Body) icon, err := io.ReadAll(resp.Body)
if err != nil { if err != nil {
return &FetchResult{StatusCode: http.StatusInternalServerError, ErrMsg: "internal error"} return FetchResultWithErrorf(http.StatusInternalServerError, "failed to read response body: %w", err)
} }
if len(icon) == 0 { if len(icon) == 0 {
return &FetchResult{StatusCode: http.StatusNotFound, ErrMsg: "empty icon"} return FetchResultWithErrorf(http.StatusNotFound, "empty icon")
} }
res := &FetchResult{Icon: icon} res := FetchResult{Icon: icon}
if contentType := resp.Header.Get("Content-Type"); contentType != "" { if contentType := resp.Header.Get("Content-Type"); contentType != "" {
res.contentType = contentType res.contentType = contentType
} }
// else leave it empty // else leave it empty
storeIconCache(url, res) return res, nil
return res }).WithMaxEntries(200).WithRetriesExponentialBackoff(3).WithTTL(4 * time.Hour).Build()
}
var nameSanitizer = strings.NewReplacer( var nameSanitizer = strings.NewReplacer(
"_", "-", "_", "-",
@@ -106,57 +122,57 @@ func sanitizeName(name string) string {
return strings.ToLower(nameSanitizer.Replace(name)) return strings.ToLower(nameSanitizer.Replace(name))
} }
func fetchKnownIcon(ctx context.Context, url *IconURL) *FetchResult { func fetchKnownIcon(ctx context.Context, url *IconURL) (FetchResult, error) {
// if icon isn't in the list, no need to fetch // if icon isn't in the list, no need to fetch
if !url.HasIcon() { if !url.HasIcon() {
return &FetchResult{StatusCode: http.StatusNotFound, ErrMsg: "no such icon"} return FetchResult{StatusCode: http.StatusNotFound}, errors.New("no such icon")
} }
return fetchIconAbsolute(ctx, url.URL()) return FetchIconAbsolute(ctx, url.URL())
} }
func fetchIcon(ctx context.Context, filename string) *FetchResult { func fetchIcon(ctx context.Context, filename string) (FetchResult, error) {
for _, fileType := range []string{"svg", "webp", "png"} { for _, fileType := range []string{"svg", "webp", "png"} {
result := fetchKnownIcon(ctx, NewSelfhStIconURL(filename, fileType)) result, err := fetchKnownIcon(ctx, NewSelfhStIconURL(filename, fileType))
if result.OK() { if err == nil {
return result return result, err
} }
result = fetchKnownIcon(ctx, NewWalkXCodeIconURL(filename, fileType)) result, err = fetchKnownIcon(ctx, NewWalkXCodeIconURL(filename, fileType))
if result.OK() { if err == nil {
return result return result, err
} }
} }
return &FetchResult{StatusCode: http.StatusNotFound, ErrMsg: "no icon found"} return FetchResultWithErrorf(http.StatusNotFound, "no icon found")
} }
func FindIcon(ctx context.Context, r route, uri string) *FetchResult { func FindIcon(ctx context.Context, r route, uri string) (FetchResult, error) {
if result := loadIconCache(r.Key()); result != nil {
return result
}
for _, ref := range r.References() { for _, ref := range r.References() {
result := fetchIcon(ctx, sanitizeName(ref)) result, err := fetchIcon(ctx, sanitizeName(ref))
if result.OK() { if err == nil {
storeIconCache(r.Key(), result) return result, err
return result
} }
} }
if r, ok := r.(httpRoute); ok { if r, ok := r.(httpRoute); ok {
// fallback to parse html // fallback to parse html
return findIconSlow(ctx, r, uri, nil) return findIconSlowCached(context.WithValue(ctx, "route", r), uri)
} }
return &FetchResult{StatusCode: http.StatusNotFound, ErrMsg: "no icon found"} return FetchResultWithErrorf(http.StatusNotFound, "no icon found")
} }
func findIconSlow(ctx context.Context, r httpRoute, uri string, stack []string) *FetchResult { var findIconSlowCached = cache.NewKeyFunc(func(ctx context.Context, key string) (FetchResult, error) {
r := ctx.Value("route").(httpRoute)
return findIconSlow(ctx, r, key, nil)
}).WithMaxEntries(200).Build() // no retries, no ttl
func findIconSlow(ctx context.Context, r httpRoute, uri string, stack []string) (FetchResult, error) {
select { select {
case <-ctx.Done(): case <-ctx.Done():
return &FetchResult{StatusCode: http.StatusBadGateway, ErrMsg: "request timeout"} return FetchResultWithErrorf(http.StatusBadGateway, "request timeout")
default: default:
} }
if len(stack) > maxRedirectDepth { if len(stack) > maxRedirectDepth {
return &FetchResult{StatusCode: http.StatusBadGateway, ErrMsg: "too many redirects"} return FetchResultWithErrorf(http.StatusBadGateway, "too many redirects")
} }
ctx, cancel := context.WithTimeoutCause(ctx, faviconFetchTimeout, errors.New("favicon request timeout")) ctx, cancel := context.WithTimeoutCause(ctx, faviconFetchTimeout, errors.New("favicon request timeout"))
@@ -164,13 +180,13 @@ func findIconSlow(ctx context.Context, r httpRoute, uri string, stack []string)
newReq, err := http.NewRequestWithContext(ctx, http.MethodGet, r.TargetURL().String(), nil) newReq, err := http.NewRequestWithContext(ctx, http.MethodGet, r.TargetURL().String(), nil)
if err != nil { if err != nil {
return &FetchResult{StatusCode: http.StatusInternalServerError, ErrMsg: "cannot create request"} return FetchResultWithErrorf(http.StatusInternalServerError, "cannot create request: %w", err)
} }
newReq.Header.Set("Accept-Encoding", "identity") // disable compression newReq.Header.Set("Accept-Encoding", "identity") // disable compression
u, err := url.ParseRequestURI(strutils.SanitizeURI(uri)) u, err := url.ParseRequestURI(strutils.SanitizeURI(uri))
if err != nil { if err != nil {
return &FetchResult{StatusCode: http.StatusInternalServerError, ErrMsg: "cannot parse uri"} return FetchResultWithErrorf(http.StatusInternalServerError, "cannot parse uri: %w", err)
} }
newReq.URL.Path = u.Path newReq.URL.Path = u.Path
newReq.URL.RawPath = u.RawPath newReq.URL.RawPath = u.RawPath
@@ -182,48 +198,48 @@ func findIconSlow(ctx context.Context, r httpRoute, uri string, stack []string)
if c.status != http.StatusOK { if c.status != http.StatusOK {
switch c.status { switch c.status {
case 0: case 0:
return &FetchResult{StatusCode: http.StatusBadGateway, ErrMsg: "connection error"} return FetchResultWithErrorf(http.StatusBadGateway, "connection error")
default: default:
if loc := c.Header().Get("Location"); loc != "" { if loc := c.Header().Get("Location"); loc != "" {
loc = strutils.SanitizeURI(loc) loc = strutils.SanitizeURI(loc)
if loc == "/" || loc == newReq.URL.Path || slices.Contains(stack, loc) { if loc == "/" || loc == newReq.URL.Path || slices.Contains(stack, loc) {
return &FetchResult{StatusCode: http.StatusBadGateway, ErrMsg: "circular redirect"} return FetchResultWithErrorf(http.StatusBadGateway, "circular redirect")
} }
// append current path to stack // append current path to stack
// handles redirect to the same path with different query // handles redirect to the same path with different query
return findIconSlow(ctx, r, loc, append(stack, newReq.URL.Path)) return findIconSlow(ctx, r, loc, append(stack, newReq.URL.Path))
} }
} }
return &FetchResult{StatusCode: c.status, ErrMsg: "upstream error: " + string(c.data)} return FetchResultWithErrorf(c.status, "upstream error: %s", c.data)
} }
// return icon data // return icon data
if !httputils.GetContentType(c.header).IsHTML() { if !httputils.GetContentType(c.header).IsHTML() {
return &FetchResult{Icon: c.data, contentType: c.header.Get("Content-Type")} return FetchResultOK(c.data, c.header.Get("Content-Type"))
} }
// try extract from "link[rel=icon]" from path "/" // try extract from "link[rel=icon]" from path "/"
doc, err := goquery.NewDocumentFromReader(bytes.NewBuffer(c.data)) doc, err := goquery.NewDocumentFromReader(bytes.NewBuffer(c.data))
if err != nil { if err != nil {
return &FetchResult{StatusCode: http.StatusInternalServerError, ErrMsg: "failed to parse html"} return FetchResultWithErrorf(http.StatusInternalServerError, "failed to parse html: %w", err)
} }
ele := doc.Find("head > link[rel=icon]").First() ele := doc.Find("head > link[rel=icon]").First()
if ele.Length() == 0 { if ele.Length() == 0 {
return &FetchResult{StatusCode: http.StatusNotFound, ErrMsg: "icon element not found"} return FetchResultWithErrorf(http.StatusNotFound, "icon element not found")
} }
href := ele.AttrOr("href", "") href := ele.AttrOr("href", "")
if href == "" { if href == "" {
return &FetchResult{StatusCode: http.StatusNotFound, ErrMsg: "icon href not found"} return FetchResultWithErrorf(http.StatusNotFound, "icon href not found")
} }
// https://en.wikipedia.org/wiki/Data_URI_scheme // https://en.wikipedia.org/wiki/Data_URI_scheme
if strings.HasPrefix(href, "data:image/") { if strings.HasPrefix(href, "data:image/") {
dataURI, err := dataurl.DecodeString(href) dataURI, err := dataurl.DecodeString(href)
if err != nil { if err != nil {
return &FetchResult{StatusCode: http.StatusInternalServerError, ErrMsg: "failed to decode favicon"} return FetchResultWithErrorf(http.StatusInternalServerError, "failed to decode favicon: %w", err)
} }
return &FetchResult{Icon: dataURI.Data, contentType: dataURI.ContentType()} return FetchResultOK(dataURI.Data, dataURI.ContentType())
} }
switch { switch {
case strings.HasPrefix(href, "http://"), strings.HasPrefix(href, "https://"): case strings.HasPrefix(href, "http://"), strings.HasPrefix(href, "https://"):
return fetchIconAbsolute(ctx, href) return FetchIconAbsolute(ctx, href)
default: default:
return findIconSlow(ctx, r, href, append(stack, newReq.URL.Path)) return findIconSlow(ctx, r, href, append(stack, newReq.URL.Path))
} }

View File

@@ -1,145 +0,0 @@
package homepage
import (
"encoding/base64"
"sync"
"time"
"github.com/bytedance/sonic"
"github.com/rs/zerolog/log"
"github.com/yusing/godoxy/internal/common"
"github.com/yusing/godoxy/internal/jsonstore"
"github.com/yusing/godoxy/internal/utils"
"github.com/yusing/godoxy/internal/utils/atomic"
"github.com/yusing/goutils/task"
)
type cacheEntry struct {
Icon []byte `json:"icon"`
ContentType string `json:"content_type,omitempty"`
LastAccess atomic.Value[time.Time] `json:"last_access"`
}
// cache key can be absolute url or route name.
var (
iconCache = jsonstore.Store[*cacheEntry](common.NamespaceIconCache)
iconMu sync.RWMutex
)
const (
iconCacheTTL = 3 * 24 * time.Hour
cleanUpInterval = time.Minute
maxIconSize = 1024 * 1024 // 1MB
maxCacheEntries = 100
)
func init() {
go func() {
cleanupTicker := time.NewTicker(cleanUpInterval)
defer cleanupTicker.Stop()
for {
select {
case <-task.RootContextCanceled():
return
case <-cleanupTicker.C:
pruneExpiredIconCache()
}
}
}()
}
func pruneExpiredIconCache() {
nPruned := 0
for key, icon := range iconCache.Range {
if icon.IsExpired() {
iconCache.Delete(key)
nPruned++
}
}
if iconCache.Size() > maxCacheEntries {
iconCache.Clear()
newIconCache := make(map[string]*cacheEntry, maxCacheEntries)
i := 0
for key, icon := range iconCache.Range {
if i == maxCacheEntries {
break
}
if !icon.IsExpired() {
newIconCache[key] = icon
i++
}
}
for key, icon := range newIconCache {
iconCache.Store(key, icon)
}
}
if nPruned > 0 {
log.Info().Int("pruned", nPruned).Msg("pruned expired icon cache")
}
}
func PruneRouteIconCache(route route) {
iconCache.Delete(route.Key())
}
func loadIconCache(key string) *FetchResult {
iconMu.RLock()
defer iconMu.RUnlock()
icon, ok := iconCache.Load(key)
if ok && len(icon.Icon) > 0 {
log.Debug().
Str("key", key).
Msg("icon found in cache")
icon.LastAccess.Store(utils.TimeNow())
return &FetchResult{Icon: icon.Icon, contentType: icon.ContentType}
}
return nil
}
func storeIconCache(key string, result *FetchResult) {
icon := result.Icon
if len(icon) > maxIconSize {
log.Debug().Int("size", len(icon)).Msg("icon cache size exceeds max cache size")
return
}
iconMu.Lock()
defer iconMu.Unlock()
entry := &cacheEntry{Icon: icon, ContentType: result.contentType}
entry.LastAccess.Store(time.Now())
iconCache.Store(key, entry)
log.Debug().Str("key", key).Int("size", len(icon)).Msg("stored icon cache")
}
func (e *cacheEntry) IsExpired() bool {
return time.Since(e.LastAccess.Load()) > iconCacheTTL
}
func (e *cacheEntry) UnmarshalJSON(data []byte) error {
var tmp struct {
Icon []byte `json:"icon"`
ContentType string `json:"content_type,omitempty"`
LastAccess time.Time `json:"last_access"`
}
// check if data is json
if sonic.Valid(data) {
err := sonic.Unmarshal(data, &tmp)
// return only if unmarshal is successful
// otherwise fallback to base64
if err == nil {
e.Icon = tmp.Icon
e.ContentType = tmp.ContentType
e.LastAccess.Store(tmp.LastAccess)
return nil
}
}
// fallback to base64
icon, err := base64.StdEncoding.DecodeString(string(data))
if err == nil {
e.Icon = icon
e.LastAccess.Store(time.Now())
return nil
}
return err
}

View File

@@ -62,10 +62,10 @@ func (w *Watcher) wakeFromHTTP(rw http.ResponseWriter, r *http.Request) (shouldN
// handle favicon request // handle favicon request
if isFaviconPath(r.URL.Path) { if isFaviconPath(r.URL.Path) {
result := api.GetFavIconFromAlias(r.Context(), w.route.Name()) result, err := api.GetFavIconFromAlias(r.Context(), w.route.Name())
if !result.OK() { if err != nil {
rw.WriteHeader(result.StatusCode) rw.WriteHeader(result.StatusCode)
fmt.Fprint(rw, result.ErrMsg) fmt.Fprint(rw, err)
return false return false
} }
rw.Header().Set("Content-Type", result.ContentType()) rw.Header().Set("Content-Type", result.ContentType())

View File

@@ -6,7 +6,6 @@ import (
"github.com/yusing/godoxy/agent/pkg/agent" "github.com/yusing/godoxy/agent/pkg/agent"
"github.com/yusing/godoxy/agent/pkg/agentproxy" "github.com/yusing/godoxy/agent/pkg/agentproxy"
"github.com/yusing/godoxy/internal/homepage"
"github.com/yusing/godoxy/internal/idlewatcher" "github.com/yusing/godoxy/internal/idlewatcher"
"github.com/yusing/godoxy/internal/logging/accesslog" "github.com/yusing/godoxy/internal/logging/accesslog"
gphttp "github.com/yusing/godoxy/internal/net/gphttp" gphttp "github.com/yusing/godoxy/internal/net/gphttp"
@@ -150,8 +149,6 @@ func (r *ReveseProxyRoute) Start(parent task.Parent) gperr.Error {
routes.HTTP.Del(r) routes.HTTP.Del(r)
}) })
} }
r.task.OnCancel("reset_favicon", func() { homepage.PruneRouteIconCache(r) })
return nil return nil
} }