This repository has no description
4.9 kB
168 lines
1package xrpc
2
3import (
4 "context"
5 "crypto/sha256"
6 "fmt"
7 "io"
8 "net/http"
9 "path/filepath"
10 "slices"
11 "strings"
12
13 "github.com/bluesky-social/indigo/atproto/atclient"
14 "github.com/bluesky-social/indigo/atproto/syntax"
15 "github.com/go-git/go-git/v5/plumbing/object"
16 "tangled.org/core/knotmirror/xrpc/gitea"
17)
18
19func (x *Xrpc) GetBlob(w http.ResponseWriter, r *http.Request) {
20 var (
21 repoQuery = r.URL.Query().Get("repo")
22 ref = r.URL.Query().Get("ref") // ref can be empty (git.Open handles this)
23 path = r.URL.Query().Get("path")
24 )
25
26 repo, err := syntax.ParseDID(repoQuery)
27 if err != nil {
28 writeJson(w, http.StatusBadRequest, atclient.ErrorBody{Name: "BadRequest", Message: fmt.Sprintf("repo parameter invalid: %s", repoQuery)})
29 return
30 }
31
32 l := x.logger.With("method", "git.getBlob", "repo", repo, "ref", ref, "path", path)
33 l.Debug("request")
34
35 if path == "" {
36 writeJson(w, http.StatusBadRequest, atclient.ErrorBody{Name: "BadRequest", Message: "missing path parameter"})
37 return
38 }
39
40 ctx := r.Context()
41
42 repoPath, err := x.makeRepoPath(ctx, repo)
43 if err != nil {
44 writeJson(w, http.StatusNotFound, atclient.ErrorBody{Name: "RepoNotFound", Message: fmt.Sprintf("unknown repository: %s", repo)})
45 return
46 }
47
48 entry, err := x.getFile(ctx, repoPath, ref, path)
49 if err != nil {
50 l.Warn("local mirror failed", "err", err)
51 writeJson(w, http.StatusInternalServerError, atclient.ErrorBody{Name: "InternalServerError", Message: "failed to get blob"})
52 return
53 }
54 size, reader, err := gitea.ReadBlob(ctx, repoPath, entry.Hash)
55 if err != nil {
56 l.Warn("local mirror failed", "err", err)
57 writeJson(w, http.StatusInternalServerError, atclient.ErrorBody{Name: "InternalServerError", Message: "failed to get blob"})
58 return
59 }
60 defer reader.Close()
61
62 // default to octet-stream for large blobs
63 if size > 1000*1000 { // 1MB
64 w.Header().Set("Content-Type", "application/octet-stream")
65 if _, err := io.Copy(w, reader); err != nil {
66 l.Error("failed to serve the blob", "err", err)
67 }
68 return
69 }
70
71 contents, err := io.ReadAll(reader)
72 if err != nil {
73 l.Error("failed to read blob content", "err", err)
74 writeJson(w, http.StatusInternalServerError, atclient.ErrorBody{Name: "InternalServerError", Message: "failed to read the blob"})
75 return
76 }
77
78 mimeType := http.DetectContentType(contents)
79 // override MIME types for formats that http.DetectContentType does not recognize
80 switch filepath.Ext(path) {
81 case ".svg":
82 mimeType = "image/svg+xml"
83 case ".avif":
84 mimeType = "image/avif"
85 case ".jxl":
86 mimeType = "image/jxl"
87 case ".heic", ".heif":
88 mimeType = "image/heif"
89 }
90
91 switch {
92 case strings.HasPrefix(mimeType, "image/"), strings.HasPrefix(mimeType, "video/"):
93 eTag := fmt.Sprintf("\"%x\"", sha256.Sum256(contents))
94 if clientETag := r.Header.Get("If-None-Match"); clientETag == eTag {
95 w.WriteHeader(http.StatusNotModified)
96 return
97 }
98 w.Header().Set("ETag", eTag)
99 w.Header().Set("Content-Type", mimeType)
100
101 case strings.HasPrefix(mimeType, "text/") || isTextualMimeType(mimeType):
102 w.Header().Set("Cache-Control", "public, no-cache")
103 // serve all text content as text/plain
104 w.Header().Set("Content-Type", "text/plain; charset=utf-8")
105
106 default:
107 l.Error("attempted to serve disallowed file type", "mimetype", mimeType)
108 writeJson(w, http.StatusInternalServerError, atclient.ErrorBody{Name: "InvalidRequest", Message: "only image, video, and text files can be accessed directly"})
109 return
110 }
111 w.Write(contents)
112}
113
114func (x *Xrpc) getFile(ctx context.Context, repoPath, ref, path string) (*object.TreeEntry, error) {
115 rev := ref
116 if rev == "" {
117 rev = "HEAD"
118 }
119
120 head, err := gitea.GetCommit(ctx, repoPath, rev)
121 if err != nil {
122 return nil, fmt.Errorf("get head commit: %w", err)
123 }
124
125 treePath := filepath.Dir(path)
126 name := filepath.Base(path)
127
128 // find subTree
129 subRev := head.Hash.String() + "^{tree}"
130 if treePath != "." {
131 subRev = head.Hash.String() + ":" + treePath
132 }
133 subTree, err := gitea.GetTree(ctx, repoPath, subRev)
134 if err != nil {
135 return nil, fmt.Errorf("get subtree %s: %w", subRev, err)
136 }
137
138 // find entry
139 entry, err := func(subTree *object.Tree) (*object.TreeEntry, error) {
140 for _, entry := range subTree.Entries {
141 if entry.Name == name {
142 return &entry, nil
143 }
144 }
145 return nil, fmt.Errorf("object doesn't exist")
146 }(subTree)
147 if err != nil {
148 return nil, fmt.Errorf("get file: %w", err)
149 }
150
151 return entry, nil
152}
153
154var textualMimeTypes = []string{
155 "application/json",
156 "application/xml",
157 "application/yaml",
158 "application/x-yaml",
159 "application/toml",
160 "application/javascript",
161 "application/ecmascript",
162}
163
164// isTextualMimeType returns true if the MIME type represents textual content
165// that should be served as text/plain for security reasons
166func isTextualMimeType(mimeType string) bool {
167 return slices.Contains(textualMimeTypes, mimeType)
168}