Add ArchGitChecksum template command in bashbrew cat

This also finally adds `bashbrew context` as an explicit subcommand so that issues with this code are easier to test/debug (so we can generate the actual tarball and compare it to previous versions of it, versions generated by `git archive`, etc).

As-is, this currently generates verbatim identical checksums to https://github.com/docker-library/meta-scripts/blob/0cde8de57dfe411ed5578feffe1b10f811e11dc2/sources.sh#L90-L96 (by design).  We'll wait to do any cache bust there until we implement `Dockerfile`/context filtering:

```console
$ bashbrew cat varnish:stable --format '{{ .TagEntry.GitCommit }} {{ .TagEntry.Directory }}'
0c295b528f28a98650fb2580eab6d34b30b165c4 stable/debian
$ git -C "$BASHBREW_CACHE/git" archive 0c295b528f28a98650fb2580eab6d34b30b165c4:stable/debian/ | ./tar-scrubber | sha256sum
3aef5ac859b23d65dfe5e9f2a47750e9a32852222829cfba762a870c1473fad6
$ bashbrew cat --format '{{ .ArchGitChecksum arch .TagEntry }}' varnish:stable
3aef5ac859b23d65dfe5e9f2a47750e9a32852222829cfba762a870c1473fad6
```

(Choosing `varnish:stable` there because it currently has [some 100% valid dangling symlinks](https://github.com/varnish/docker-varnish/tree/6b1c6ffedcfececac71e46a85122c1adaef25868/stable/debian/scripts) that tripped up my code beautifully 💕)

From a performance perspective (which was the original reason for looking into / implementing this), running the `meta-scripts/sources.sh` script against `--all` vs this, my local system gets ~18.5m vs ~4.5m (faster being this new pure-Go implementation).

Author: tianon

cmd/bashbrew/git.go

@@ -96,6 +96,20 @@ func getGitCommit(commit string) (string, error) {
 	return h.String(), nil
 }
 
+func (r Repo) archGitFS(arch string, entry *manifest.Manifest2822Entry) (fs.FS, error) {
+	commit, err := r.fetchGitRepo(arch, entry)
+	if err != nil {
+		return nil, fmt.Errorf("failed fetching %q: %w", r.EntryIdentifier(entry), err)
+	}
+
+	gitFS, err := gitCommitFS(commit)
+	if err != nil {
+		return nil, err
+	}
+
+	return fs.Sub(gitFS, entry.ArchDirectory(arch))
+}
+
 func gitCommitFS(commit string) (fs.FS, error) {
 	if err := ensureGitInit(); err != nil {
 		return nil, err

cmd/bashbrew/main.go

@@ -3,11 +3,13 @@ package main
 import (
 	"fmt"
 	"os"
+	"path"
 	"path/filepath"
 	"strings"
 
 	"github.com/sirupsen/logrus" // this is used by containerd libraries, so we need to set the default log level for it
 	"github.com/urfave/cli"
+	xTerm "golang.org/x/term"
 
 	"github.com/docker-library/bashbrew/architecture"
 	"github.com/docker-library/bashbrew/manifest"
@@ -421,6 +423,61 @@ func main() {
 
 			Category: "plumbing",
 		},
+		{
+			Name:  "context",
+			Usage: "(eventually Dockerfile-filtered) git archive",
+			Flags: []cli.Flag{
+				cli.BoolFlag{
+					Name:  "sha256",
+					Usage: `print sha256 instead of raw tar`,
+				},
+				// TODO "unfiltered" or something for not applying Dockerfile filtering (once that's implemented)
+			},
+			Before: subcommandBeforeFactory("context"),
+			Action: func(c *cli.Context) error {
+				repos, err := repos(false, c.Args()...)
+				if err != nil {
+					return err
+				}
+				if len(repos) != 1 {
+					return fmt.Errorf("'context' expects to act on exactly one architecture of one entry of one repo (got %d repos)", len(repos))
+				}
+
+				r, err := fetch(repos[0])
+				if err != nil {
+					return err
+				}
+
+				// TODO technically something like "hello-world:latest" *could* be relaxed a little if it resolves via architecture to one and only one entry 🤔 (but that's a little hard to implement with the existing internal data structures -- see TODO at the top of "sort.go")
+
+				if r.TagEntry == nil {
+					return fmt.Errorf("'context' expects to act on exactly one architecture of one entry of one repo (no specific entry of %q selected)", r.RepoName)
+				}
+				if len(r.TagEntries) != 1 {
+					return fmt.Errorf("'context' expects to act on exactly one architecture of one entry of one repo (got %d entires)", len(r.TagEntries))
+				}
+
+				if !r.TagEntry.HasArchitecture(arch) {
+					return fmt.Errorf("%q does not include architecture %q", path.Join(namespace, r.RepoName)+":"+r.TagEntry.Tags[0], arch)
+				}
+
+				if c.Bool("sha256") {
+					sum, err := r.ArchGitChecksum(arch, r.TagEntry)
+					if err != nil {
+						return err
+					}
+					fmt.Println(sum)
+					return nil
+				} else {
+					if xTerm.IsTerminal(int(os.Stdout.Fd())) {
+						return fmt.Errorf("cowardly refusing to output a tar to a terminal")
+					}
+					return r.archContextTar(arch, r.TagEntry, os.Stdout)
+				}
+			},
+
+			Category: "plumbing",
+		},
 		{
 			Name:     "remote",
 			Usage:    "query registries for bashbrew-related data",

cmd/bashbrew/oci-builder.go

@@ -89,6 +89,7 @@ func importOCIBlob(ctx context.Context, cs content.Store, fs iofs.FS, descriptor
 
 // this is "docker build" but for "Builder: oci-import"
 func ociImportBuild(tags []string, commit, dir, file string) (*imagespec.Descriptor, error) {
+	// TODO use r.archGitFS (we have no r or arch or entry here 😅)
 	fs, err := gitCommitFS(commit)
 	if err != nil {
 		return nil, err

cmd/bashbrew/sort.go

@@ -5,6 +5,8 @@ import (
 	"pault.ag/go/topsort"
 )
 
+// TODO unify archFilter and applyConstraints handling by pre-filtering the full list of Repo objects such that all that remains are things we should process (thus removing all "if" statements throughout the various loops); re-doing the Architectures and Entries lists to only include ones we should process, etc
+
 func sortRepos(repos []string, applyConstraints bool) ([]string, error) {
 	rs := []*Repo{}
 	rsMap := map[*Repo]string{}
@@ -103,10 +105,10 @@ func sortRepoObjects(rs []*Repo, applyConstraints bool) ([]*Repo, error) {
 				continue
 			}
 			/*
-			// TODO need archFilter here :(
-			if archFilter && !entry.HasArchitecture(arch) {
-				continue
-			}
+				// TODO need archFilter here :(
+				if archFilter && !entry.HasArchitecture(arch) {
+					continue
+				}
 			*/
 
 			entryArches := []string{arch}

cmd/bashbrew/tar.go

@@ -0,0 +1,28 @@
+package main
+
+import (
+	"crypto/sha256"
+	"fmt"
+	"io"
+
+	"github.com/docker-library/bashbrew/manifest"
+	"github.com/docker-library/bashbrew/pkg/tarscrub"
+)
+
+func (r Repo) archContextTar(arch string, entry *manifest.Manifest2822Entry, w io.Writer) error {
+	f, err := r.archGitFS(arch, entry)
+	if err != nil {
+		return err
+	}
+
+	return tarscrub.WriteTar(f, w)
+}
+
+func (r Repo) ArchGitChecksum(arch string, entry *manifest.Manifest2822Entry) (string, error) {
+	h := sha256.New()
+	err := r.archContextTar(arch, entry, h)
+	if err != nil {
+		return "", err
+	}
+	return fmt.Sprintf("%x", h.Sum(nil)), nil
+}

go.mod

@@ -9,6 +9,7 @@ require (
 	github.com/sirupsen/logrus v1.9.0
 	github.com/urfave/cli v1.22.10
 	go.etcd.io/bbolt v1.3.7
+	golang.org/x/term v0.5.0
 	pault.ag/go/debian v0.12.0
 	pault.ag/go/topsort v0.1.1
 )

go.sum

@@ -940,6 +940,7 @@ golang.org/x/term v0.0.0-20210927222741-03fcf44c2211/go.mod h1:jbD1KX2456YbFQfuX
 golang.org/x/term v0.0.0-20220722155259-a9ba230a4035/go.mod h1:jbD1KX2456YbFQfuXm/mYQcufACuNUgVhRMnK/tPxf8=
 golang.org/x/term v0.2.0/go.mod h1:TVmDHMZPmdnySmBfhjOoOdhjzdE1h4u1VwSiw2l1Nuc=
 golang.org/x/term v0.5.0 h1:n2a8QNdAb0sZNpU9R1ALUXBbY+w51fCQDN+7EdxNBsY=
+golang.org/x/term v0.5.0/go.mod h1:jMB1sMXY+tzblOD4FWmEbocvup2/aLOaQEp7JmGp78k=
 golang.org/x/text v0.0.0-20170915032832-14c0d48ead0c/go.mod h1:NqM8EUOU14njkJ3fqMW+pc6Ldnwhi/IjpwHt7yyuwOQ=
 golang.org/x/text v0.3.0/go.mod h1:NqM8EUOU14njkJ3fqMW+pc6Ldnwhi/IjpwHt7yyuwOQ=
 golang.org/x/text v0.3.1-0.20180807135948-17ff2d5776d2/go.mod h1:NqM8EUOU14njkJ3fqMW+pc6Ldnwhi/IjpwHt7yyuwOQ=