adalal/gotenberg
1
0
Fork 0
mirror of https://github.com/gotenberg/gotenberg.git synced 2026-08-08 00:22:14 +01:00
Code Issues Packages Projects Releases Wiki Activity

fix(chromium): better default deny list regexp

Browse Source
This commit is contained in:
Julien Neuhart
2024-02-18 17:18:40 +01:00
parent 7f3a61ef43
commit ad152e62e5
14 changed files with 419 additions and 263 deletions
Download Patch File Download Diff File Expand all files Collapse all files

55
pkg/modules/webhook/middleware.go
View File

@@ -9,14 +9,15 @@ import (
"fmt"
"net/http"
"os"
"regexp"
"strconv"
"strings"
"time"
"github.com/dlclark/regexp2"
"github.com/hashicorp/go-retryablehttp"
"github.com/labstack/echo/v4"
"github.com/gotenberg/gotenberg/v8/pkg/gotenberg"
"github.com/gotenberg/gotenberg/v8/pkg/modules/api"
)
@@ -44,38 +45,38 @@ func webhookMiddleware(w *Webhook) api.Middleware {
)
}
// Let's check if the webhook URLs are acceptable according to our
// allowed/denied lists.
filter := func(URL, header string, allowList, denyList *regexp.Regexp) error {
if !allowList.MatchString(URL) {
return api.WrapError(
fmt.Errorf("'%s' does not match the expression from the allowed list", URL),
api.NewSentinelHttpError(
http.StatusForbidden,
fmt.Sprintf("Invalid '%s' header value: '%s' does not match the authorized URLs", header, URL),
),
)
}
if denyList.String() != "" && denyList.MatchString(URL) {
return api.WrapError(
fmt.Errorf("'%s' matches the expression from the denied list", URL),
api.NewSentinelHttpError(
http.StatusForbidden,
fmt.Sprintf("Invalid '%s' header value: '%s' does not match the authorized URLs", header, URL),
),
)
}
return nil
deadline, ok := ctx.Deadline()
if !ok {
return errors.New("context has no deadline")
}
err := filter(webhookUrl, "Gotenberg-Webhook-Url", w.allowList, w.denyList)
// Let's check if the webhook URLs are acceptable according to our
// allowed/denied lists.
filter := func(url, header string, allowList, denyList *regexp2.Regexp, deadline time.Time) error {
err := gotenberg.FilterDeadline(allowList, denyList, url, deadline)
if err == nil {
return nil
}
if errors.Is(err, gotenberg.ErrFiltered) {
return api.WrapError(
err,
api.NewSentinelHttpError(
http.StatusForbidden,
fmt.Sprintf("Invalid '%s' header value: '%s' does not match the authorized URL", header, url),
),
)
}
return err
}
err := filter(webhookUrl, "Gotenberg-Webhook-Url", w.allowList, w.denyList, deadline)
if err != nil {
return fmt.Errorf("filter webhook URL: %w", err)
}
err = filter(webhookErrorUrl, "Gotenberg-Webhook-Error-Url", w.errorAllowList, w.errorDenyList)
err = filter(webhookErrorUrl, "Gotenberg-Webhook-Error-Url", w.errorAllowList, w.errorDenyList, deadline)
if err != nil {
return fmt.Errorf("filter webhook error URL: %w", err)
}

85
pkg/modules/webhook/middleware_test.go
View File

@@ -11,11 +11,11 @@ import (
"mime/multipart"
"net/http"
"net/http/httptest"
"regexp"
"strings"
"testing"
"time"
"github.com/dlclark/regexp2"
"github.com/labstack/echo/v4"
"go.uber.org/zap"
@@ -47,10 +47,10 @@ func TestWebhookMiddlewareGuards(t *testing.T) {
buildWebhookModule := func() *Webhook {
return &Webhook{
allowList: regexp.MustCompile(""),
denyList: regexp.MustCompile(""),
errorAllowList: regexp.MustCompile(""),
errorDenyList: regexp.MustCompile(""),
allowList: regexp2.MustCompile("", 0),
denyList: regexp2.MustCompile("", 0),
errorAllowList: regexp2.MustCompile("", 0),
errorDenyList: regexp2.MustCompile("", 0),
maxRetry: 0,
retryMinWait: 0,
retryMaxWait: 0,
@@ -63,6 +63,7 @@ func TestWebhookMiddlewareGuards(t *testing.T) {
request *http.Request
mod *Webhook
next echo.HandlerFunc
noDeadline bool
expectError bool
expectHttpError bool
expectHttpStatus int
@@ -76,6 +77,7 @@ func TestWebhookMiddlewareGuards(t *testing.T) {
return nil
}
}(),
noDeadline: false,
expectError: false,
expectHttpError: false,
},
@@ -87,10 +89,23 @@ func TestWebhookMiddlewareGuards(t *testing.T) {
return req
}(),
mod: buildWebhookModule(),
noDeadline: false,
expectError: true,
expectHttpError: true,
expectHttpStatus: http.StatusBadRequest,
},
{
scenario: "context has no deadline",
request: func() *http.Request {
req := buildMultipartFormDataRequest()
req.Header.Set("Gotenberg-Webhook-Url", "foo")
req.Header.Set("Gotenberg-Webhook-Error-Url", "bar")
return req
}(),
mod: buildWebhookModule(),
noDeadline: true,
expectError: true,
},
{
scenario: "webhook URL is not allowed",
request: func() *http.Request {
@@ -101,9 +116,10 @@ func TestWebhookMiddlewareGuards(t *testing.T) {
}(),
mod: func() *Webhook {
mod := buildWebhookModule()
mod.allowList = regexp.MustCompile("bar")
mod.allowList = regexp2.MustCompile("bar", 0)
return mod
}(),
noDeadline: false,
expectError: true,
expectHttpError: true,
expectHttpStatus: http.StatusForbidden,
@@ -118,9 +134,10 @@ func TestWebhookMiddlewareGuards(t *testing.T) {
}(),
mod: func() *Webhook {
mod := buildWebhookModule()
mod.denyList = regexp.MustCompile("foo")
mod.denyList = regexp2.MustCompile("foo", 0)
return mod
}(),
noDeadline: false,
expectError: true,
expectHttpError: true,
expectHttpStatus: http.StatusForbidden,
@@ -135,9 +152,10 @@ func TestWebhookMiddlewareGuards(t *testing.T) {
}(),
mod: func() *Webhook {
mod := buildWebhookModule()
mod.errorAllowList = regexp.MustCompile("foo")
mod.errorAllowList = regexp2.MustCompile("foo", 0)
return mod
}(),
noDeadline: false,
expectError: true,
expectHttpError: true,
expectHttpStatus: http.StatusForbidden,
@@ -152,9 +170,10 @@ func TestWebhookMiddlewareGuards(t *testing.T) {
}(),
mod: func() *Webhook {
mod := buildWebhookModule()
mod.errorDenyList = regexp.MustCompile("bar")
mod.errorDenyList = regexp2.MustCompile("bar", 0)
return mod
}(),
noDeadline: false,
expectError: true,
expectHttpError: true,
expectHttpStatus: http.StatusForbidden,
@@ -169,6 +188,7 @@ func TestWebhookMiddlewareGuards(t *testing.T) {
return req
}(),
mod: buildWebhookModule(),
noDeadline: false,
expectError: true,
expectHttpError: true,
expectHttpStatus: http.StatusBadRequest,
@@ -183,6 +203,7 @@ func TestWebhookMiddlewareGuards(t *testing.T) {
return req
}(),
mod: buildWebhookModule(),
noDeadline: false,
expectError: true,
expectHttpError: true,
expectHttpStatus: http.StatusBadRequest,
@@ -198,6 +219,7 @@ func TestWebhookMiddlewareGuards(t *testing.T) {
return req
}(),
mod: buildWebhookModule(),
noDeadline: false,
expectError: true,
expectHttpError: true,
expectHttpStatus: http.StatusBadRequest,
@@ -213,6 +235,7 @@ func TestWebhookMiddlewareGuards(t *testing.T) {
return req
}(),
mod: buildWebhookModule(),
noDeadline: false,
expectError: true,
expectHttpError: true,
expectHttpStatus: http.StatusBadRequest,
@@ -228,6 +251,7 @@ func TestWebhookMiddlewareGuards(t *testing.T) {
return req
}(),
mod: buildWebhookModule(),
noDeadline: false,
expectError: true,
expectHttpError: true,
expectHttpStatus: http.StatusBadRequest,
@@ -242,6 +266,7 @@ func TestWebhookMiddlewareGuards(t *testing.T) {
return req
}(),
mod: buildWebhookModule(),
noDeadline: false,
expectError: true,
expectHttpError: true,
expectHttpStatus: http.StatusBadRequest,
@@ -254,15 +279,20 @@ func TestWebhookMiddlewareGuards(t *testing.T) {
c := srv.NewContext(tc.request, httptest.NewRecorder())
ctx := &api.ContextMock{Context: &api.Context{}}
ctx.SetEchoContext(c)
c.Set("context", ctx.Context)
c.Set("cancel", func() context.CancelFunc {
return func() {
return
}
}())
if tc.noDeadline {
ctx := &api.ContextMock{Context: &api.Context{Context: context.Background()}}
ctx.SetEchoContext(c)
c.Set("context", ctx.Context)
c.Set("cancel", func() context.CancelFunc {
return nil
}())
} else {
timeoutCtx, cancel := context.WithTimeout(context.Background(), time.Duration(10)*time.Second)
ctx := &api.ContextMock{Context: &api.Context{Context: timeoutCtx}}
ctx.SetEchoContext(c)
c.Set("context", ctx.Context)
c.Set("cancel", cancel)
}
err := webhookMiddleware(tc.mod).Handler(tc.next)(c)
@@ -320,10 +350,10 @@ func TestWebhookMiddlewareAsynchronousProcess(t *testing.T) {
buildWebhookModule := func() *Webhook {
return &Webhook{
allowList: regexp.MustCompile(""),
denyList: regexp.MustCompile(""),
errorAllowList: regexp.MustCompile(""),
errorDenyList: regexp.MustCompile(""),
allowList: regexp2.MustCompile("", 0),
denyList: regexp2.MustCompile("", 0),
errorAllowList: regexp2.MustCompile("", 0),
errorDenyList: regexp2.MustCompile("", 0),
maxRetry: 0,
retryMinWait: 0,
retryMaxWait: 0,
@@ -426,22 +456,17 @@ func TestWebhookMiddlewareAsynchronousProcess(t *testing.T) {
c.Set("trace", "foo")
c.Set("startTime", time.Now())
ctx := &api.ContextMock{Context: &api.Context{}}
timeoutCtx, cancel := context.WithTimeout(context.Background(), time.Duration(10)*time.Second)
ctx := &api.ContextMock{Context: &api.Context{Context: timeoutCtx}}
ctx.SetLogger(zap.NewNop())
ctx.SetEchoContext(c)
c.Set("context", ctx.Context)
c.Set("cancel", func() context.CancelFunc {
return func() {
return
}
}())
c.Set("cancel", cancel)
webhook := echo.New()
webhook.HideBanner = true
webhook.HidePort = true
rand.Seed(time.Now().UnixNano())
webhookPort := rand.Intn(65535-1025+1) + 1025
c.Request().Header.Set("Gotenberg-Webhook-Url", fmt.Sprintf("http://localhost:%d/", webhookPort))

10
pkg/modules/webhook/webhook.go
View File

@@ -1,9 +1,9 @@
package webhook
import (
"regexp"
"time"
"github.com/dlclark/regexp2"
flag "github.com/spf13/pflag"
"github.com/gotenberg/gotenberg/v8/pkg/gotenberg"
@@ -17,10 +17,10 @@ func init() {
// Webhook is a module which provides a middleware for uploading output files
// to any destinations in an asynchronous fashion.
type Webhook struct {
allowList *regexp.Regexp
denyList *regexp.Regexp
errorAllowList *regexp.Regexp
errorDenyList *regexp.Regexp
allowList *regexp2.Regexp
denyList *regexp2.Regexp
errorAllowList *regexp2.Regexp
errorDenyList *regexp2.Regexp
maxRetry int
retryMinWait time.Duration
retryMaxWait time.Duration