From 8661818123c08925dc16c97684259c004e7058ee Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Matthias=20Ri=C3=9Fe?= <m.risse@fz-juelich.de>
Date: Thu, 16 Jan 2025 18:49:09 +0100
Subject: [PATCH] Use annexed content for comparison in diffs

This makes it such that annexed files are treated like plain git files
in comparisons (e.g. the diff of a commit).

It also changes the image diff viewer to show a more reasonable error
message when one of the annexed files under comparison is missing.
---
 modules/annex/annex.go           | 12 ++++++++++++
 routers/web/repo/compare.go      | 27 +++++++++++++++++++++------
 web_src/js/features/imagediff.js | 12 +++++++++++-
 3 files changed, 44 insertions(+), 7 deletions(-)

diff --git a/modules/annex/annex.go b/modules/annex/annex.go
index 9a70612a1e..dee24d21bc 100644
--- a/modules/annex/annex.go
+++ b/modules/annex/annex.go
@@ -22,6 +22,7 @@ import (
 	"code.gitea.io/gitea/modules/git"
 	"code.gitea.io/gitea/modules/log"
 	"code.gitea.io/gitea/modules/setting"
+	"code.gitea.io/gitea/modules/typesniffer"
 )
 
 // ErrBlobIsNotAnnexed occurs if a blob does not contain a valid annex key
@@ -178,3 +179,14 @@ func UUID2RepoPath(uuid string) (string, error) {
 	// Otherwise just return the cached entry
 	return repoPath, nil
 }
+
+// GuessContentType guesses the content type of the annexed blob.
+func GuessContentType(blob *git.Blob) (typesniffer.SniffedType, error) {
+	r, err := Content(blob)
+	if err != nil {
+		return typesniffer.SniffedType{}, err
+	}
+	defer r.Close()
+
+	return typesniffer.DetectContentTypeFromReader(r)
+}
diff --git a/routers/web/repo/compare.go b/routers/web/repo/compare.go
index 03d49fa692..104655e134 100644
--- a/routers/web/repo/compare.go
+++ b/routers/web/repo/compare.go
@@ -23,6 +23,7 @@ import (
 	repo_model "code.gitea.io/gitea/models/repo"
 	"code.gitea.io/gitea/models/unit"
 	user_model "code.gitea.io/gitea/models/user"
+	"code.gitea.io/gitea/modules/annex"
 	"code.gitea.io/gitea/modules/base"
 	"code.gitea.io/gitea/modules/charset"
 	csv_module "code.gitea.io/gitea/modules/csv"
@@ -72,7 +73,21 @@ func setCompareContext(ctx *context.Context, before, head *git.Commit, headOwner
 			return st
 		}
 
-		st, err := blob.GuessContentType()
+		isAnnexed, err := annex.IsAnnexed(blob)
+		if err != nil {
+			log.Error("IsAnnexed failed: %v", err)
+			return st
+		}
+		if isAnnexed {
+			st, err = annex.GuessContentType(blob)
+			if err != nil {
+				log.Error("GuessContentType failed: %v", err)
+				return st
+			}
+			return st
+		}
+
+		st, err = blob.GuessContentType()
 		if err != nil {
 			log.Error("GuessContentType failed: %v", err)
 			return st
@@ -90,18 +105,18 @@ func SourceCommitURL(owner, name string, commit *git.Commit) string {
 	return setting.AppSubURL + "/" + url.PathEscape(owner) + "/" + url.PathEscape(name) + "/src/commit/" + url.PathEscape(commit.ID.String())
 }
 
-// RawCommitURL creates a relative URL for the raw commit in the given repository
-func RawCommitURL(owner, name string, commit *git.Commit) string {
-	return setting.AppSubURL + "/" + url.PathEscape(owner) + "/" + url.PathEscape(name) + "/raw/commit/" + url.PathEscape(commit.ID.String())
+// MediaCommitURL creates a relative URL for the commit media (plain git, LFS, or annex content) in the given repository
+func MediaCommitURL(owner, name string, commit *git.Commit) string {
+	return setting.AppSubURL + "/" + url.PathEscape(owner) + "/" + url.PathEscape(name) + "/media/commit/" + url.PathEscape(commit.ID.String())
 }
 
 // setPathsCompareContext sets context data for source and raw paths
 func setPathsCompareContext(ctx *context.Context, base, head *git.Commit, headOwner, headName string) {
 	ctx.Data["SourcePath"] = SourceCommitURL(headOwner, headName, head)
-	ctx.Data["RawPath"] = RawCommitURL(headOwner, headName, head)
+	ctx.Data["RawPath"] = MediaCommitURL(headOwner, headName, head)
 	if base != nil {
 		ctx.Data["BeforeSourcePath"] = SourceCommitURL(headOwner, headName, base)
-		ctx.Data["BeforeRawPath"] = RawCommitURL(headOwner, headName, base)
+		ctx.Data["BeforeRawPath"] = MediaCommitURL(headOwner, headName, base)
 	}
 }
 
diff --git a/web_src/js/features/imagediff.js b/web_src/js/features/imagediff.js
index d1b139ffde..52c9017b07 100644
--- a/web_src/js/features/imagediff.js
+++ b/web_src/js/features/imagediff.js
@@ -92,7 +92,17 @@ export function initImageDiff() {
         return loadElem(img, info.path);
       }));
       // only the first images is associated with $boundsInfo
-      if (!success) info.$boundsInfo.text('(image error)');
+      if (!success) {
+        const blobContent = await GET(info.path.replace('/media/', '/raw/')).then((response) => response.text());
+        if (blobContent.startsWith('.git/annex/objects')) {
+          for (const item of document.querySelectorAll('.image-diff .overflow-menu-items .item')) {
+            item.style.display = 'none';
+          }
+          info.$boundsInfo[0].parentElement.textContent = 'annexed file is not present on the server';
+        } else {
+          info.$boundsInfo.text('(image error)');
+        }
+      }
       if (info.mime === 'image/svg+xml') {
         const resp = await GET(info.path);
         const text = await resp.text();