0
Fork 0
mirror of https://codeberg.org/forgejo/forgejo.git synced 2024-12-29 10:54:02 -05:00
forgejo/routers/common/middleware.go
Giteabot a2a9b0f977
Support sanitising the URL by removing extra slashes in the URL (#21333) (#23300)
Backport #21333

Changes in this PR :

Strips incoming request URL of additional slashes (/). For example an
input like

`https://git.data.coop//halfd/new-website.git` is translated to
`https://git.data.coop/halfd/new-website.git`

Fixes https://github.com/go-gitea/gitea/issues/20462

Fix #23242

Co-authored-by: Sandeep Bhat <sandyethadka@gmail.com>
Co-authored-by: zeripath <art27@cantab.net>
Co-authored-by: Jason Song <i@wolfogre.com>
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
Co-authored-by: delvh <leon@kske.dev>
2023-03-05 02:14:12 -05:00

114 lines
3.4 KiB
Go

// Copyright 2021 The Gitea Authors. All rights reserved.
// SPDX-License-Identifier: MIT
package common
import (
"fmt"
"net/http"
"strings"
"code.gitea.io/gitea/modules/cache"
"code.gitea.io/gitea/modules/context"
"code.gitea.io/gitea/modules/log"
"code.gitea.io/gitea/modules/process"
"code.gitea.io/gitea/modules/setting"
"code.gitea.io/gitea/modules/web/routing"
"github.com/chi-middleware/proxy"
chi "github.com/go-chi/chi/v5"
)
// Middlewares returns common middlewares
func Middlewares() []func(http.Handler) http.Handler {
handlers := []func(http.Handler) http.Handler{
func(next http.Handler) http.Handler {
return http.HandlerFunc(func(resp http.ResponseWriter, req *http.Request) {
// First of all escape the URL RawPath to ensure that all routing is done using a correctly escaped URL
req.URL.RawPath = req.URL.EscapedPath()
ctx, _, finished := process.GetManager().AddTypedContext(req.Context(), fmt.Sprintf("%s: %s", req.Method, req.RequestURI), process.RequestProcessType, true)
defer finished()
next.ServeHTTP(context.NewResponse(resp), req.WithContext(cache.WithCacheContext(ctx)))
})
},
}
if setting.ReverseProxyLimit > 0 {
opt := proxy.NewForwardedHeadersOptions().
WithForwardLimit(setting.ReverseProxyLimit).
ClearTrustedProxies()
for _, n := range setting.ReverseProxyTrustedProxies {
if !strings.Contains(n, "/") {
opt.AddTrustedProxy(n)
} else {
opt.AddTrustedNetwork(n)
}
}
handlers = append(handlers, proxy.ForwardedHeaders(opt))
}
// Strip slashes.
handlers = append(handlers, stripSlashesMiddleware)
if !setting.Log.DisableRouterLog {
handlers = append(handlers, routing.NewLoggerHandler())
}
if setting.Log.EnableAccessLog {
handlers = append(handlers, context.AccessLogger())
}
handlers = append(handlers, func(next http.Handler) http.Handler {
return http.HandlerFunc(func(resp http.ResponseWriter, req *http.Request) {
// Why we need this? The Recovery() will try to render a beautiful
// error page for user, but the process can still panic again, and other
// middleware like session also may panic then we have to recover twice
// and send a simple error page that should not panic anymore.
defer func() {
if err := recover(); err != nil {
routing.UpdatePanicError(req.Context(), err)
combinedErr := fmt.Sprintf("PANIC: %v\n%s", err, log.Stack(2))
log.Error("%v", combinedErr)
if setting.IsProd {
http.Error(resp, http.StatusText(http.StatusInternalServerError), http.StatusInternalServerError)
} else {
http.Error(resp, combinedErr, http.StatusInternalServerError)
}
}
}()
next.ServeHTTP(resp, req)
})
})
return handlers
}
func stripSlashesMiddleware(next http.Handler) http.Handler {
return http.HandlerFunc(func(resp http.ResponseWriter, req *http.Request) {
var urlPath string
rctx := chi.RouteContext(req.Context())
if rctx != nil && rctx.RoutePath != "" {
urlPath = rctx.RoutePath
} else if req.URL.RawPath != "" {
urlPath = req.URL.RawPath
} else {
urlPath = req.URL.Path
}
sanitizedPath := &strings.Builder{}
prevWasSlash := false
for _, chr := range strings.TrimRight(urlPath, "/") {
if chr != '/' || !prevWasSlash {
sanitizedPath.WriteRune(chr)
}
prevWasSlash = chr == '/'
}
if rctx == nil {
req.URL.Path = sanitizedPath.String()
} else {
rctx.RoutePath = sanitizedPath.String()
}
next.ServeHTTP(resp, req)
})
}