From 6cf2b6eb0af726b1ba6fdf24372182faae1edea3 Mon Sep 17 00:00:00 2001
From: bryan newbold <bnewbold@robocracy.org>
Date: Wed, 16 Aug 2023 19:22:46 -0700
Subject: [PATCH] syntax: add RecordKey type

---
 atproto/syntax/recordkey.go                   | 31 ++++++++++
 atproto/syntax/recordkey_test.go              | 59 +++++++++++++++++++
 .../testdata/recordkey_syntax_invalid.txt     | 14 +++++
 .../testdata/recordkey_syntax_valid.txt       |  5 ++
 4 files changed, 109 insertions(+)
 create mode 100644 atproto/syntax/recordkey.go
 create mode 100644 atproto/syntax/recordkey_test.go
 create mode 100644 atproto/syntax/testdata/recordkey_syntax_invalid.txt
 create mode 100644 atproto/syntax/testdata/recordkey_syntax_valid.txt

diff --git a/atproto/syntax/recordkey.go b/atproto/syntax/recordkey.go
new file mode 100644
index 00000000..5de4724a
--- /dev/null
+++ b/atproto/syntax/recordkey.go
@@ -0,0 +1,31 @@
+package syntax
+
+import (
+	"fmt"
+	"regexp"
+)
+
+// String type which represents a syntaxtually valid RecordKey identifier, as could be included in an AT URI
+//
+// Always use [ParseRecordKey] instead of wrapping strings directly, especially when working with input.
+//
+// Syntax specification: https://atproto.com/specs/record-key
+type RecordKey string
+
+func ParseRecordKey(raw string) (RecordKey, error) {
+	if len(raw) > 253 {
+		return "", fmt.Errorf("recordkey is too long (512 chars max)")
+	}
+	if raw == "" || raw == "." || raw == ".." {
+		return "", fmt.Errorf("recordkey can not be empty, '.', or '..'")
+	}
+	var recordkeyRegex = regexp.MustCompile(`^[a-zA-Z0-9_~.-]{1,512}$`)
+	if !recordkeyRegex.MatchString(raw) {
+		return "", fmt.Errorf("recordkey syntax didn't validate via regex")
+	}
+	return RecordKey(raw), nil
+}
+
+func (h RecordKey) String() string {
+	return string(h)
+}
diff --git a/atproto/syntax/recordkey_test.go b/atproto/syntax/recordkey_test.go
new file mode 100644
index 00000000..f25cb36c
--- /dev/null
+++ b/atproto/syntax/recordkey_test.go
@@ -0,0 +1,59 @@
+package syntax
+
+import (
+	"bufio"
+	"fmt"
+	"os"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+)
+
+func TestInteropRecordKeysValid(t *testing.T) {
+	assert := assert.New(t)
+	file, err := os.Open("testdata/recordkey_syntax_valid.txt")
+	assert.NoError(err)
+	defer file.Close()
+	scanner := bufio.NewScanner(file)
+	for scanner.Scan() {
+		line := scanner.Text()
+		if len(line) == 0 || line[0] == '#' {
+			continue
+		}
+		_, err := ParseRecordKey(line)
+		if err != nil {
+			fmt.Println("GOOD: " + line)
+		} else {
+			fmt.Println("FAIL: " + line)
+		}
+		assert.NoError(err)
+	}
+	assert.NoError(scanner.Err())
+}
+
+func TestInteropRecordKeysInvalid(t *testing.T) {
+	assert := assert.New(t)
+	file, err := os.Open("testdata/recordkey_syntax_invalid.txt")
+	assert.NoError(err)
+	defer file.Close()
+	scanner := bufio.NewScanner(file)
+	for scanner.Scan() {
+		line := scanner.Text()
+		if len(line) == 0 || line[0] == '#' {
+			continue
+		}
+		_, err := ParseRecordKey(line)
+		if err == nil {
+			fmt.Println("BAD: " + line)
+		}
+		assert.Error(err)
+	}
+	assert.NoError(scanner.Err())
+}
+
+func TestRecordKeyNoPanic(t *testing.T) {
+	for _, s := range []string{"", "a", ".", ".."} {
+		bad := RecordKey(s)
+		_ = bad.String()
+	}
+}
diff --git a/atproto/syntax/testdata/recordkey_syntax_invalid.txt b/atproto/syntax/testdata/recordkey_syntax_invalid.txt
new file mode 100644
index 00000000..1da3d1e7
--- /dev/null
+++ b/atproto/syntax/testdata/recordkey_syntax_invalid.txt
@@ -0,0 +1,14 @@
+# specs
+literal:self
+alpha/beta
+.
+..
+#extra
+@handle
+any space
+any+space
+number[3]
+number(3)
+"quote"
+pre:fix
+dHJ1ZQ==
diff --git a/atproto/syntax/testdata/recordkey_syntax_valid.txt b/atproto/syntax/testdata/recordkey_syntax_valid.txt
new file mode 100644
index 00000000..4ba39689
--- /dev/null
+++ b/atproto/syntax/testdata/recordkey_syntax_valid.txt
@@ -0,0 +1,5 @@
+# specs
+self
+example.com
+~1.2-3_
+dHJ1ZQ
-- 
2.51.2