alice.pds.demo.boltless.dev
This repository has no description
3.9 kB
157 lines
1package main
2
3import (
4 "context"
5 "crypto/sha1"
6 "encoding/json"
7 "fmt"
8 "io"
9 "net/url"
10 "os"
11 "path/filepath"
12
13 "github.com/bluesky-social/indigo/atproto/identity"
14 "github.com/bluesky-social/indigo/atproto/syntax"
15 indigoxrpc "github.com/bluesky-social/indigo/xrpc"
16 "github.com/sourcegraph/zoekt"
17 "tangled.org/core/api/tangled"
18)
19
20// 1 MB; match https://sourcegraph.sourcegraph.com/r/github.com/sourcegraph/sourcegraph/-/blob/cmd/searcher/internal/search/store.go?L32
21const MaxFileSize = 1 << 20
22
23func gitIndex(ctx context.Context, cfg *Config, dir identity.Directory, req indexRequest) error {
24 ctx, cancel := context.WithTimeout(ctx, cfg.IndexTimeout)
25 defer cancel()
26
27 repo, err := loadRepo(ctx, dir, req.Repo)
28 if err != nil {
29 return nil
30 }
31 repo.Branches = req.Branches
32
33 gitDir, err := tmpGitDir(repo.Did.String())
34 if err != nil {
35 return err
36 }
37 defer os.RemoveAll(gitDir) // best-effort cleanup
38
39 if err := fetchRepo(ctx, gitDir, repo.CloneURL(), req.Branches); err != nil {
40 return err
41 }
42
43 if err := indexRepo(ctx, cfg, gitDir, *repo); err != nil {
44 return err
45 }
46
47 return nil
48}
49
50func loadRepo(ctx context.Context, dir identity.Directory, repoDID syntax.DID) (*Repo, error) {
51 ident, err := dir.LookupDID(ctx, repoDID)
52 if err != nil {
53 return nil, err
54 }
55
56 knot := ident.PDSEndpoint()
57
58 xrpcc := &indigoxrpc.Client{Host: knot}
59 out, err := tangled.RepoDescribeRepo(ctx, xrpcc, repoDID.String())
60 if err != nil {
61 return nil, err
62 }
63
64 return &Repo{
65 Did: repoDID,
66 Owner: syntax.DID(out.OwnerDid),
67 Slug: syntax.RecordKey(out.Rkey),
68 Knot: knot,
69 }, nil
70}
71
72func fetchRepo(ctx context.Context, gitDir, cloneUrl string, branches []zoekt.RepositoryBranch) error {
73 // Create a repo to fetch into
74 if err := executeCmd(ctx,
75 "git",
76 // use a random default branch. This is so that HEAD isn't a symref to a
77 // branch that is indexed. For example if you are indexing
78 // HEAD,master. Then HEAD would be pointing to master by default.
79 "-c", "init.defaultBranch=nonExistentBranchBB0FOFCH32",
80 "init",
81 // we don't need a working copy
82 "--bare",
83 gitDir,
84 ); err != nil {
85 return err
86 }
87
88 fetchArgs := []string{
89 "-C", gitDir,
90 "-c", "protocol.version=2",
91 "fetch", "--depth=1", "--no-tags",
92 }
93 // Git's blob:limit filter excludes blobs whose size is >= the given limit,
94 // while zoekt indexes files up to and including FileLimit bytes.
95 fetchArgs = append(fetchArgs, fmt.Sprintf("--filter=blob:limit=%d", int64(MaxFileSize)+1))
96
97 fetchArgs = append(fetchArgs, cloneUrl)
98
99 var commits []string
100 for _, b := range branches {
101 commits = append(commits, b.Version)
102 }
103 fetchArgs = append(fetchArgs, commits...)
104
105 if err := executeCmd(ctx, "git", fetchArgs...); err != nil {
106 return err
107 }
108
109 for _, b := range branches {
110 ref := b.Name
111 if ref != "HEAD" {
112 ref = "refs/heads/" + ref
113 }
114 if err := executeCmd(ctx, "git", "-C", gitDir, "update-ref", ref, b.Version); err != nil {
115 return fmt.Errorf("failed update-ref %s to %s: %w", ref, b.Version, err)
116 }
117 }
118
119 return nil
120}
121
122func indexRepo(ctx context.Context, cfg *Config, gitDir string, repo Repo) error {
123 executablePath, err := os.Executable()
124 if err != nil {
125 return err
126 }
127
128 repoJson, err := json.Marshal(repo)
129 if err != nil {
130 return err
131 }
132
133 args := []string{"index"}
134 args = append(args, "-index-dir", cfg.IndexDir)
135 args = append(args, "-appview-url", cfg.AppviewUrl)
136 args = append(args, gitDir, string(repoJson))
137 if err := executeCmd(ctx, executablePath, args...); err != nil {
138 return err
139 }
140 return nil
141}
142
143func tmpGitDir(name string) (string, error) {
144 abs := url.QueryEscape(name)
145 if len(abs) > 200 {
146 h := sha1.New()
147 _, _ = io.WriteString(h, abs)
148 abs = abs[:200] + fmt.Sprintf("%x", h.Sum(nil))[:8]
149 }
150 dir := filepath.Join(os.TempDir(), abs+".git")
151 if _, err := os.Stat(dir); err == nil {
152 if err := os.RemoveAll(dir); err != nil {
153 return "", err
154 }
155 }
156 return dir, nil
157}