improve & refactor & return specific error pages

server/gitea/cache.go

@@ -2,6 +2,7 @@ package gitea
 
 import (
 	"bytes"
+	"fmt"
 	"io"
 	"net/http"
 	"time"
@@ -55,6 +56,8 @@ func (f FileResponse) createHttpResponse() *http.Response {
 	}
 	resp.Header.Set(eTagHeader, f.ETag)
 	resp.Header.Set(contentTypeHeader, f.MimeType)
+	resp.Header.Set(contentLengthHeader, fmt.Sprint(len(f.Body)))
+	resp.Header.Set(pagesCacheIndicator, "true")
 
 	return resp
 }

server/gitea/client.go

@@ -27,6 +27,9 @@ const (
 	defaultBranchCacheKeyPrefix   = "defaultBranch"
 	rawContentCacheKeyPrefix      = "rawContent"
 
+	// pages server
+	pagesCacheIndicator = "X-Pages-Cache"
+
 	// gitea
 	giteaObjectTypeHeader = "X-Gitea-Object-Type"
 	objTypeSymlink        = "symlink"

server/handler.go

@@ -1,7 +1,9 @@
 package server
 
 import (
+	"fmt"
 	"net/http"
+	"path"
 	"strings"
 
 	"github.com/rs/zerolog"
@@ -58,7 +60,7 @@ func Handler(mainDomainSuffix, rawDomain string,
 		// Block blacklisted paths (like ACME challenges)
 		for _, blacklistedPath := range blacklistedPaths {
 			if strings.HasPrefix(ctx.Path(), blacklistedPath) {
-				html.ReturnErrorPage(ctx, "", http.StatusForbidden)
+				html.ReturnErrorPage(ctx, "requested blacklisted path", http.StatusForbidden)
 				return
 			}
 		}
@@ -90,7 +92,7 @@ func Handler(mainDomainSuffix, rawDomain string,
 
 		// tryBranch checks if a branch exists and populates the target variables. If canonicalLink is non-empty, it will
 		// also disallow search indexing and add a Link header to the canonical URL.
-		tryBranch := func(log zerolog.Logger, repo, branch string, path []string, canonicalLink string) bool {
+		tryBranch := func(log zerolog.Logger, repo, branch string, _path []string, canonicalLink string) bool {
 			if repo == "" {
 				log.Debug().Msg("tryBranch: repo == ''")
 				return false
@@ -109,7 +111,7 @@ func Handler(mainDomainSuffix, rawDomain string,
 
 			// Branch exists, use it
 			targetRepo = repo
-			targetPath = strings.Trim(strings.Join(path, "/"), "/")
+			targetPath = path.Join(_path...)
 			targetBranch = branchTimestampResult.Branch
 
 			targetOptions.BranchTimestamp = branchTimestampResult.Timestamp
@@ -157,8 +159,8 @@ func Handler(mainDomainSuffix, rawDomain string,
 						canonicalDomainCache)
 					return
 				}
-				log.Debug().Msg("missing branch")
-				html.ReturnErrorPage(ctx, "", http.StatusFailedDependency)
+				log.Debug().Msg("missing branch info")
+				html.ReturnErrorPage(ctx, "missing branch info", http.StatusFailedDependency)
 				return
 			}
 
@@ -198,8 +200,9 @@ func Handler(mainDomainSuffix, rawDomain string,
 				}
 
 				log.Debug().Msg("main domain preparations, now trying with specified repo & branch")
+				branch := pathElements[1][1:]
 				if tryBranch(log,
-					pathElements[0], pathElements[1][1:], pathElements[2:],
+					pathElements[0], branch, pathElements[2:],
 					"/"+pathElements[0]+"/%p",
 				) {
 					log.Debug().Msg("tryBranch, now trying upstream 3")
@@ -207,7 +210,9 @@ func Handler(mainDomainSuffix, rawDomain string,
 						targetOptions, targetOwner, targetRepo, targetBranch, targetPath,
 						canonicalDomainCache)
 				} else {
-					html.ReturnErrorPage(ctx, "", http.StatusFailedDependency)
+					html.ReturnErrorPage(ctx,
+						fmt.Sprintf("explizite set branch '%s' do not exist at '%s/%s'", branch, targetOwner, targetRepo),
+						http.StatusFailedDependency)
 				}
 				return
 			}
@@ -216,14 +221,17 @@ func Handler(mainDomainSuffix, rawDomain string,
 			// example.codeberg.page/@main/index.html
 			if strings.HasPrefix(pathElements[0], "@") {
 				log.Debug().Msg("main domain preparations, now trying with specified branch")
+				branch := pathElements[0][1:]
 				if tryBranch(log,
-					"pages", pathElements[0][1:], pathElements[1:], "/%p") {
+					"pages", branch, pathElements[1:], "/%p") {
 					log.Debug().Msg("tryBranch, now trying upstream 4")
 					tryUpstream(ctx, giteaClient, mainDomainSuffix, trimmedHost,
-						targetOptions, targetOwner, targetRepo, targetBranch, targetPath,
+						targetOptions, targetOwner, "pages", targetBranch, targetPath,
 						canonicalDomainCache)
 				} else {
-					html.ReturnErrorPage(ctx, "", http.StatusFailedDependency)
+					html.ReturnErrorPage(ctx,
+						fmt.Sprintf("explizite set branch '%s' do not exist at '%s/%s'", branch, targetOwner, "pages"),
+						http.StatusFailedDependency)
 				}
 				return
 			}
@@ -254,15 +262,19 @@ func Handler(mainDomainSuffix, rawDomain string,
 			}
 
 			// Couldn't find a valid repo/branch
-			html.ReturnErrorPage(ctx, "", http.StatusFailedDependency)
+			html.ReturnErrorPage(ctx,
+				fmt.Sprintf("couldn't find a valid repo[%s]/branch[%s]", targetRepo, targetBranch),
+				http.StatusFailedDependency)
 			return
 		} else {
 			trimmedHostStr := string(trimmedHost)
 
-			// Serve pages from external domains
+			// Serve pages from custom domains
 			targetOwner, targetRepo, targetBranch = dns.GetTargetFromDNS(trimmedHostStr, string(mainDomainSuffix), dnsLookupCache)
 			if targetOwner == "" {
-				html.ReturnErrorPage(ctx, "", http.StatusFailedDependency)
+				html.ReturnErrorPage(ctx,
+					"could not obtain repo owner from custom domain",
+					http.StatusFailedDependency)
 				return
 			}
 
@@ -280,7 +292,7 @@ func Handler(mainDomainSuffix, rawDomain string,
 				targetRepo, targetBranch, pathElements, canonicalLink) {
 				canonicalDomain, valid := upstream.CheckCanonicalDomain(giteaClient, targetOwner, targetRepo, targetBranch, trimmedHostStr, string(mainDomainSuffix), canonicalDomainCache)
 				if !valid {
-					html.ReturnErrorPage(ctx, "", http.StatusMisdirectedRequest)
+					html.ReturnErrorPage(ctx, "domain not specified in <code>.domains</code> file", http.StatusMisdirectedRequest)
 					return
 				} else if canonicalDomain != trimmedHostStr {
 					// only redirect if the target is also a codeberg page!
@@ -290,7 +302,7 @@ func Handler(mainDomainSuffix, rawDomain string,
 						return
 					}
 
-					html.ReturnErrorPage(ctx, "", http.StatusFailedDependency)
+					html.ReturnErrorPage(ctx, "target is no codeberg page", http.StatusFailedDependency)
 					return
 				}
 
@@ -301,7 +313,7 @@ func Handler(mainDomainSuffix, rawDomain string,
 				return
 			}
 
-			html.ReturnErrorPage(ctx, "", http.StatusFailedDependency)
+			html.ReturnErrorPage(ctx, "could not find target for custom domain", http.StatusFailedDependency)
 			return
 		}
 	}

server/upstream/upstream.go

@@ -2,6 +2,7 @@ package upstream
 
 import (
 	"errors"
+	"fmt"
 	"io"
 	"net/http"
 	"strings"
@@ -56,23 +57,25 @@ type Options struct {
 func (o *Options) Upstream(ctx *context.Context, giteaClient *gitea.Client) (final bool) {
 	log := log.With().Strs("upstream", []string{o.TargetOwner, o.TargetRepo, o.TargetBranch, o.TargetPath}).Logger()
 
+	if o.TargetOwner == "" || o.TargetRepo == "" {
+		html.ReturnErrorPage(ctx, "either repo owner or name info is missing", http.StatusBadRequest)
+		return true
+	}
+
 	// Check if the branch exists and when it was modified
 	if o.BranchTimestamp.IsZero() {
 		branch := GetBranchTimestamp(giteaClient, o.TargetOwner, o.TargetRepo, o.TargetBranch)
 
-		if branch == nil {
-			html.ReturnErrorPage(ctx, "", http.StatusFailedDependency)
+		if branch == nil || branch.Branch == "" {
+			html.ReturnErrorPage(ctx,
+				fmt.Sprintf("could not get timestamp of branch '%s'", o.TargetBranch),
+				http.StatusFailedDependency)
 			return true
 		}
 		o.TargetBranch = branch.Branch
 		o.BranchTimestamp = branch.Timestamp
 	}
 
-	if o.TargetOwner == "" || o.TargetRepo == "" || o.TargetBranch == "" {
-		html.ReturnErrorPage(ctx, "", http.StatusBadRequest)
-		return true
-	}
-
 	// Check if the browser has a cached version
 	if ifModifiedSince, err := time.Parse(time.RFC1123, string(ctx.Response().Header.Get(headerIfModifiedSince))); err == nil {
 		if !ifModifiedSince.Before(o.BranchTimestamp) {
@@ -127,7 +130,7 @@ func (o *Options) Upstream(ctx *context.Context, giteaClient *gitea.Client) (fin
 	}
 	if res != nil && (err != nil || res.StatusCode != http.StatusOK) {
 		log.Printf("Couldn't fetch contents (status code %d): %v\n", res.StatusCode, err)
-		html.ReturnErrorPage(ctx, "", http.StatusInternalServerError)
+		html.ReturnErrorPage(ctx, fmt.Sprintf("%v", err), http.StatusInternalServerError)
 		return true
 	}