internal/pkg: optionally suppress via assumed checksum

This is quite error-prone and causes cache inconsistency similar to the store inconsistency seen on nix when a similar condition happens. Keep this behind a flag in case it is ever beneficial.

Signed-off-by: Ophestra <cat@gensokyo.uk>

3 files changed, 100 insertions, 11 deletions

diff --git a/internal/pkg/dir_test.go b/internal/pkg/dir_test.go
index cceb89b1..f1a81355 100644
--- a/internal/pkg/dir_test.go
+++ b/internal/pkg/dir_test.go
@@ -184,6 +184,32 @@ func TestFlatten(t *testing.T) {
 			{Mode: fs.ModeDir | 0700, Path: "work"},
 		}, pkg.MustDecode("WVpvsVqVKg9Nsh744x57h51AuWUoUR2nnh8Md-EYBQpk6ziyTuUn6PLtF2e0Eu_d"), nil},
 
+		{"sample no assume checksum", fstest.MapFS{
+			".": {Mode: fs.ModeDir | 0700},
+
+			"checksum": {Mode: fs.ModeDir | 0700},
+			"checksum/Aubi5EG4_Y8DhL9bQ3Q4HFBhLRF7X5gt9D3CNCQfT-TeBtlRXc7Zi_JYZEMoCC7M":       {Mode: fs.ModeDir | 0500},
+			"checksum/Aubi5EG4_Y8DhL9bQ3Q4HFBhLRF7X5gt9D3CNCQfT-TeBtlRXc7Zi_JYZEMoCC7M/check": {Mode: 0400, Data: []byte{}},
+
+			"identifier": {Mode: fs.ModeDir | 0700},
+			"identifier/_wAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA": {Mode: fs.ModeSymlink | 0777, Data: []byte("../checksum/Aubi5EG4_Y8DhL9bQ3Q4HFBhLRF7X5gt9D3CNCQfT-TeBtlRXc7Zi_JYZEMoCC7M")},
+			"identifier/_wEAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA": {Mode: fs.ModeSymlink | 0777, Data: []byte("../checksum/Aubi5EG4_Y8DhL9bQ3Q4HFBhLRF7X5gt9D3CNCQfT-TeBtlRXc7Zi_JYZEMoCC7M")},
+
+			"work": {Mode: fs.ModeDir | 0700},
+		}, []pkg.FlatEntry{
+			{Mode: fs.ModeDir | 0700, Path: "."},
+
+			{Mode: fs.ModeDir | 0700, Path: "checksum"},
+			{Mode: fs.ModeDir | 0500, Path: "checksum/Aubi5EG4_Y8DhL9bQ3Q4HFBhLRF7X5gt9D3CNCQfT-TeBtlRXc7Zi_JYZEMoCC7M"},
+			{Mode: 0400, Path: "checksum/Aubi5EG4_Y8DhL9bQ3Q4HFBhLRF7X5gt9D3CNCQfT-TeBtlRXc7Zi_JYZEMoCC7M/check", Data: []byte{}},
+
+			{Mode: fs.ModeDir | 0700, Path: "identifier"},
+			{Mode: fs.ModeSymlink | 0777, Path: "identifier/_wAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA", Data: []byte("../checksum/Aubi5EG4_Y8DhL9bQ3Q4HFBhLRF7X5gt9D3CNCQfT-TeBtlRXc7Zi_JYZEMoCC7M")},
+			{Mode: fs.ModeSymlink | 0777, Path: "identifier/_wEAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA", Data: []byte("../checksum/Aubi5EG4_Y8DhL9bQ3Q4HFBhLRF7X5gt9D3CNCQfT-TeBtlRXc7Zi_JYZEMoCC7M")},
+
+			{Mode: fs.ModeDir | 0700, Path: "work"},
+		}, pkg.MustDecode("OC290t23aimNo2Rp2pPwan5GI2KRLRdOwYxXQMD9jw0QROgHnNXWodoWdV0hwu2w"), nil},
+
 		{"sample tar step unpack", fstest.MapFS{
 			".": {Mode: fs.ModeDir | 0500},
 
diff --git a/internal/pkg/pkg.go b/internal/pkg/pkg.go
index 0c882bea..ccb7541e 100644
--- a/internal/pkg/pkg.go
+++ b/internal/pkg/pkg.go
@@ -507,6 +507,20 @@ const (
 	// CSchedIdle arranges for the [ext.SCHED_IDLE] scheduling priority to be
 	// set for [KindExec] and [KindExecNet] containers.
 	CSchedIdle
+
+	// CAssumeChecksum enables the use of [KnownChecksum] for duplicate function
+	// call suppression via the on-disk cache.
+	//
+	// This may cause incorrect cure outcome if an impossible checksum is
+	// specified that matches an output already present in the on-disk cache.
+	// This may be avoided by purposefully specifying a statistically
+	// unattainable checksum, like the zero value.
+	//
+	// While this optimisation might seem appealing, it is almost never
+	// applicable in real world use. Almost every time this path was taken, it
+	// was caused by an incorrect checksum accidentally left behind while
+	// bumping a package. Only enable this if you are really sure you need it.
+	CAssumeChecksum
 )
 
 // Cache is a support layer that implementations of [Artifact] can use to store
@@ -1041,7 +1055,7 @@ func (c *Cache) finaliseIdent(
 // [FileArtifact] to the filesystem. If err is nil, the caller is responsible
 // for closing the resulting [io.ReadCloser].
 func (c *Cache) openFile(f FileArtifact) (r io.ReadCloser, err error) {
-	if kc, ok := f.(KnownChecksum); ok {
+	if kc, ok := f.(KnownChecksum); c.flags&CAssumeChecksum != 0 && ok {
 		c.checksumMu.RLock()
 		r, err = os.Open(c.base.Append(
 			dirChecksum,
@@ -1484,16 +1498,18 @@ func (c *Cache) cure(a Artifact, curesExempt bool) (
 			checksums,
 		)
 
-		c.checksumMu.RLock()
-		checksumFi, err = os.Stat(checksumPathname.String())
-		c.checksumMu.RUnlock()
+		if c.flags&CAssumeChecksum != 0 {
+			c.checksumMu.RLock()
+			checksumFi, err = os.Stat(checksumPathname.String())
+			c.checksumMu.RUnlock()
 
-		if err != nil {
-			if !errors.Is(err, os.ErrNotExist) {
-				return
-			}
+			if err != nil {
+				if !errors.Is(err, os.ErrNotExist) {
+					return
+				}
 
-			checksumFi, err = nil, nil
+				checksumFi, err = nil, nil
+			}
 		}
 	}
 
diff --git a/internal/pkg/pkg_test.go b/internal/pkg/pkg_test.go
index 82f260e7..50e2a584 100644
--- a/internal/pkg/pkg_test.go
+++ b/internal/pkg/pkg_test.go
@@ -469,7 +469,7 @@ func TestCache(t *testing.T) {
 	}()
 
 	testCases := []cacheTestCase{
-		{"file", pkg.CValidateKnown, nil, func(t *testing.T, base *check.Absolute, c *pkg.Cache) {
+		{"file", pkg.CValidateKnown | pkg.CAssumeChecksum, nil, func(t *testing.T, base *check.Absolute, c *pkg.Cache) {
 			identifier := (pkg.ID)(bytes.Repeat([]byte{
 				0x75, 0xe6, 0x9d, 0x6d, 0xe7, 0x9f,
 			}, 8))
@@ -626,7 +626,7 @@ func TestCache(t *testing.T) {
 			}
 		}, pkg.MustDecode("St9rlE-mGZ5gXwiv_hzQ_B8bZP-UUvSNmf4nHUZzCMOumb6hKnheZSe0dmnuc4Q2")},
 
-		{"directory", 0, nil, func(t *testing.T, base *check.Absolute, c *pkg.Cache) {
+		{"directory", pkg.CAssumeChecksum, nil, func(t *testing.T, base *check.Absolute, c *pkg.Cache) {
 			id := pkg.MustDecode(
 				"HnySzeLQvSBZuTUcvfmLEX_OmH4yJWWH788NxuLuv7kVn8_uPM6Ks4rqFWM2NZJY",
 			)
@@ -873,6 +873,53 @@ func TestCache(t *testing.T) {
 			<-wCureDone
 		}, pkg.MustDecode("E4vEZKhCcL2gPZ2Tt59FS3lDng-d_2SKa2i5G_RbDfwGn6EemptFaGLPUDiOa94C")},
 
+		{"no assume checksum", 0, nil, func(t *testing.T, base *check.Absolute, c *pkg.Cache) {
+			makeGarbage := func(work *check.Absolute, wantErr error) error {
+				if err := os.Mkdir(work.String(), 0700); err != nil {
+					return err
+				}
+
+				if err := os.WriteFile(work.Append(
+					"check",
+				).String(), nil, 0400); err != nil {
+					return err
+				}
+
+				return wantErr
+			}
+
+			wantChecksum := pkg.MustDecode("Aubi5EG4_Y8DhL9bQ3Q4HFBhLRF7X5gt9D3CNCQfT-TeBtlRXc7Zi_JYZEMoCC7M")
+
+			cureMany(t, c, []cureStep{
+				{"create", overrideChecksum{wantChecksum, overrideIdent{pkg.ID{0xff, 0}, &stubArtifact{
+					kind: pkg.KindTar,
+					cure: func(t *pkg.TContext) error {
+						return makeGarbage(t.GetWorkDir(), nil)
+					},
+				}}}, base.Append(
+					"identifier",
+					pkg.Encode(pkg.ID{0xff, 0}),
+				), wantChecksum, nil},
+
+				{"reject", overrideChecksum{wantChecksum, overrideIdent{pkg.ID{0xfe, 1}, &stubArtifact{
+					kind: pkg.KindTar,
+					cure: func(t *pkg.TContext) error {
+						return makeGarbage(t.GetWorkDir(), stub.UniqueError(0xbad))
+					},
+				}}}, nil, pkg.Checksum{}, stub.UniqueError(0xbad)},
+
+				{"match", overrideChecksum{wantChecksum, overrideIdent{pkg.ID{0xff, 1}, &stubArtifact{
+					kind: pkg.KindTar,
+					cure: func(t *pkg.TContext) error {
+						return makeGarbage(t.GetWorkDir(), nil)
+					},
+				}}}, base.Append(
+					"identifier",
+					pkg.Encode(pkg.ID{0xff, 1}),
+				), wantChecksum, nil},
+			})
+		}, pkg.MustDecode("OC290t23aimNo2Rp2pPwan5GI2KRLRdOwYxXQMD9jw0QROgHnNXWodoWdV0hwu2w")},
+
 		{"scrub", 0, nil, func(t *testing.T, base *check.Absolute, c *pkg.Cache) {
 			cureMany(t, c, []cureStep{
 				{"bad measured file", newStubFile(