protocol: add generated bindings and conformance fixtures

This commit is contained in:
sechmachine
2026-07-21 23:43:04 +07:00
parent 2d69357052
commit 6abbc26dc4
22 changed files with 5031 additions and 0 deletions
+44
View File
@@ -0,0 +1,44 @@
#!/usr/bin/env python3
"""Check the content-addressed Phase 3A conformance corpus."""
from __future__ import annotations
import hashlib
import json
import pathlib
import sys
ROOT = pathlib.Path(__file__).resolve().parents[1]
MANIFEST = ROOT / "fixtures/manifest.json"
def digest(paths: list[str]) -> str:
value = hashlib.sha256()
for relative in paths:
path = ROOT / relative
value.update(relative.encode("utf-8"))
value.update(b"\0")
value.update(path.read_bytes())
value.update(b"\0")
return value.hexdigest()
def main() -> int:
manifest = json.loads(MANIFEST.read_text(encoding="utf-8"))
paths = sorted(path.relative_to(ROOT).as_posix() for path in (ROOT / "fixtures/conformance").glob("*.tsv"))
if paths != manifest.get("files"):
raise ValueError("fixture manifest file list is stale")
actual = digest(paths)
expected = manifest.get("corpus_sha256")
if not expected or actual != expected:
raise ValueError(f"fixture corpus hash mismatch: {actual}")
print(f"Fixture corpus SHA256 {actual}")
return 0
if __name__ == "__main__":
try:
raise SystemExit(main())
except (OSError, ValueError, json.JSONDecodeError) as exc:
print(f"fixture_digest: {exc}", file=sys.stderr)
raise SystemExit(1)
+383
View File
@@ -0,0 +1,383 @@
#!/usr/bin/env python3
"""Deterministically generate the small language-neutral Phase 3A bindings."""
from __future__ import annotations
import argparse
import hashlib
import json
import pathlib
import re
import subprocess
import sys
from typing import Any
ROOT = pathlib.Path(__file__).resolve().parents[1]
SCHEMA_PATH = ROOT / "schemas/control-v1.schema.json"
VERSION_PATH = ROOT / "VERSION"
COMPATIBILITY_PATH = ROOT / "compatibility.json"
def load_schema() -> tuple[dict[str, Any], str, str, dict[str, Any]]:
raw = SCHEMA_PATH.read_bytes()
schema = json.loads(raw)
version = VERSION_PATH.read_text(encoding="utf-8").strip()
if not re.fullmatch(r"\d+\.\d+\.\d+", version):
raise ValueError("VERSION must be semantic version text")
compatibility = json.loads(COMPATIBILITY_PATH.read_text(encoding="utf-8"))
for key in ("current", "n_minus_1", "n_minus_2"):
if not isinstance(compatibility.get(key), str):
raise ValueError("compatibility declaration is incomplete")
return schema, hashlib.sha256(raw).hexdigest(), version, compatibility
def pascal(name: str) -> str:
return "".join(part[:1].upper() + part[1:] for part in re.split(r"[_-]", name))
def go_field(name: str) -> str:
initialisms = {"api": "API", "http": "HTTP", "id": "ID", "ip": "IP", "sha": "SHA", "ttl": "TTL", "url": "URL", "uuid": "UUID"}
return "".join(
initialisms.get(part, part[:1].upper() + part[1:])
for part in re.split(r"[_-]", name)
)
def camel(name: str) -> str:
parts = re.split(r"[_-]", name)
return parts[0] + "".join(part[:1].upper() + part[1:] for part in parts[1:])
def swift_field(name: str) -> str:
value = camel(name)
if value in {"protocol", "class", "struct", "enum", "extension", "private", "public", "internal", "fileprivate", "open", "func", "let", "var", "import", "switch", "case", "default", "operator", "where", "repeat", "return", "throw", "throws", "try", "catch", "defer", "in", "is", "as"}:
return value + "Value"
return value
RUST_KEYWORDS = {
"as", "break", "const", "continue", "crate", "else", "enum", "extern",
"false", "fn", "for", "if", "impl", "in", "let", "loop", "match", "mod",
"move", "mut", "pub", "ref", "return", "self", "Self", "static", "struct",
"super", "trait", "true", "type", "unsafe", "use", "where", "while", "async",
"await", "dyn", "abstract", "become", "box", "do", "final", "macro", "override",
"priv", "typeof", "unsized", "virtual", "yield", "try", "union",
}
def rust_field(name: str) -> str:
value = camel(name)
return value + "Value" if value in RUST_KEYWORDS else value
def ref_name(value: Any) -> str | None:
if isinstance(value, dict) and isinstance(value.get("$ref"), str):
return value["$ref"].split("/")[-1]
return None
def prop_type(prop: dict[str, Any], language: str) -> str:
reference = ref_name(prop)
if reference:
return reference if language == "go" else (reference if language == "rust" else reference)
if prop.get("type") == "array":
item = prop.get("items", {})
item_type = prop_type(item, language)
if language == "go":
return f"[]{item_type}"
if language == "rust":
return f"Vec<{item_type}>"
return f"[{item_type}]"
if prop.get("type") == "integer":
return "int64" if language == "go" else ("i64" if language == "rust" else "Int64")
if prop.get("type") == "boolean":
return "bool" if language != "swift" else "Bool"
if prop.get("type") == "object":
return "map[string]any" if language == "go" else ("JsonObject" if language == "rust" else "JSONObject")
return "string" if language == "go" else ("String" if language == "rust" else "String")
def go_zero(prop: dict[str, Any], field: str) -> str:
typ = prop_type(prop, "go")
if typ in {"string", "int64", "bool"}:
zero = {"string": '""', "int64": "0", "bool": "false"}[typ]
return f"v.{field} == {zero}"
if typ.startswith("[]"):
return f"v.{field} == nil"
if typ == "map[string]any":
return f"v.{field} == nil"
return f"reflect.DeepEqual(v.{field}, {typ}{{}})"
def go_validation(definition: dict[str, Any]) -> list[str]:
lines: list[str] = []
name = definition["name"]
required = set(definition.get("required", []))
for prop_name, prop in definition.get("properties", {}).items():
field = go_field(prop_name)
required_zero_is_valid = prop.get("type") == "integer" and prop.get("minimum") == 0
if prop_name in required and prop.get("type") != "boolean" and not required_zero_is_valid:
zero = go_zero(prop, field)
lines.append(f"\tif {zero} {{ violations = append(violations, FieldViolation{{Field: \"{prop_name}\", Code: \"required\"}}) }}")
if prop.get("type") == "string":
if "minLength" in prop:
lines.append(f"\tif len(v.{field}) < {prop['minLength']} && v.{field} != \"\" {{ violations = append(violations, FieldViolation{{Field: \"{prop_name}\", Code: \"min_length\"}}) }}")
if "maxLength" in prop:
lines.append(f"\tif len(v.{field}) > {prop['maxLength']} {{ violations = append(violations, FieldViolation{{Field: \"{prop_name}\", Code: \"max_length\"}}) }}")
if "const" in prop:
lines.append(f"\tif v.{field} != \"{prop['const']}\" && v.{field} != \"\" {{ violations = append(violations, FieldViolation{{Field: \"{prop_name}\", Code: \"invalid_value\"}}) }}")
if "enum" in prop:
allowed = " || ".join(f'v.{field} == "{value}"' for value in prop["enum"])
lines.append(f"\tif v.{field} != \"\" && !({allowed}) {{ violations = append(violations, FieldViolation{{Field: \"{prop_name}\", Code: \"invalid_value\"}}) }}")
if prop.get("format") == "date-time":
lines.append(
'\tif v.%s != "" { if parsed, err := time.Parse(time.RFC3339Nano, v.%s); err != nil || parsed.UTC().Format(time.RFC3339Nano) != v.%s { violations = append(violations, FieldViolation{Field: "%s", Code: "invalid_time"}) } }'
% (field, field, field, prop_name)
)
if prop.get("type") == "integer":
if "minimum" in prop:
lines.append(f"\tif v.{field} != 0 && v.{field} < {prop['minimum']} {{ violations = append(violations, FieldViolation{{Field: \"{prop_name}\", Code: \"minimum\"}}) }}")
if "maximum" in prop:
lines.append(f"\tif v.{field} > {prop['maximum']} {{ violations = append(violations, FieldViolation{{Field: \"{prop_name}\", Code: \"maximum\"}}) }}")
if prop.get("type") == "array":
if "minItems" in prop:
lines.append(f"\tif len(v.{field}) < {prop['minItems']} {{ violations = append(violations, FieldViolation{{Field: \"{prop_name}\", Code: \"min_items\"}}) }}")
if "maxItems" in prop:
lines.append(f"\tif len(v.{field}) > {prop['maxItems']} {{ violations = append(violations, FieldViolation{{Field: \"{prop_name}\", Code: \"max_items\"}}) }}")
item_ref = ref_name(prop.get("items", {}))
if item_ref:
lines.append(f"\tfor index := range v.{field} {{ if err := v.{field}[index].Validate(); err != nil {{ violations = append(violations, FieldViolation{{Field: fmt.Sprintf(\"{prop_name}[%d]\", index), Code: \"invalid_item\"}}) }} }}")
reference = ref_name(prop)
if reference:
lines.append(f"\tif err := v.{field}.Validate(); err != nil {{ violations = append(violations, FieldViolation{{Field: \"{prop_name}\", Code: \"invalid_object\"}}) }}")
return lines
def generate_go(defs: dict[str, dict[str, Any]], schema_hash: str, version: str, compatibility: dict[str, Any]) -> str:
out = [
"// Code generated by tools/generate.py; DO NOT EDIT.",
"package protocol",
"",
"import (",
"\"bytes\"",
"\"encoding/json\"",
"\"errors\"",
"\"fmt\"",
"\"reflect\"",
"\"time\"",
")",
"",
f'const SchemaSHA256 = "{schema_hash}"',
f'const ProtocolVersion = "{version}"',
f'const CurrentWireVersion = "{compatibility["current"]}"',
f'const NMinus1WireVersion = "{compatibility["n_minus_1"]}"',
f'const NMinus2WireVersion = "{compatibility["n_minus_2"]}"',
"",
"type FieldViolation struct {",
"\tField string `json:\"field\"`",
"\tCode string `json:\"code\"`",
"}",
"",
"type ValidationError struct {",
"\tViolations []FieldViolation",
"}",
"",
"func (e ValidationError) Error() string { return \"protocol validation failed\" }",
"",
]
for name in sorted(defs):
if name == "FieldViolation":
continue
definition = defs[name]
out.append(f"type {name} struct {{")
required = set(definition.get("required", []))
for prop_name, prop in definition.get("properties", {}).items():
tag = prop_name + (",omitempty" if prop_name not in required else "")
out.append(f"\t{go_field(prop_name)} {prop_type(prop, 'go')} `json:\"{tag}\"`")
out.extend(["}", ""])
for name in sorted(defs):
out.append(f"func (v {name}) Validate() error {{")
out.append("\tvar violations []FieldViolation")
out.extend(go_validation(defs[name]))
out.append("\tif len(violations) > 0 { return ValidationError{Violations: violations} }")
out.append("\treturn nil")
out.append("}")
out.append("")
out.append(f"func Decode{name}(data []byte) ({name}, error) {{")
out.append(f"\tvar value {name}")
out.append("\tif len(data) > 1024*1024 { return value, errors.New(\"protocol payload exceeds limit\") }")
out.append("\tvar fields map[string]json.RawMessage")
out.append("\tif err := json.Unmarshal(data, &fields); err != nil { return value, err }")
required_fields = sorted(defs[name].get("required", []))
for prop_name in required_fields:
out.append(
'\tif raw, ok := fields["%s"]; !ok || bytes.Equal(bytes.TrimSpace(raw), []byte("null")) { return value, ValidationError{Violations: []FieldViolation{{Field: "%s", Code: "required"}}} }'
% (prop_name, prop_name)
)
for prop_name, prop in defs[name].get("properties", {}).items():
if "x-max-bytes" in prop:
out.append(
'\tif raw, ok := fields["%s"]; ok && len(raw) > %d { return value, ValidationError{Violations: []FieldViolation{{Field: "%s", Code: "max_bytes"}}} }'
% (prop_name, prop["x-max-bytes"], prop_name)
)
out.append("\tdecoder := json.NewDecoder(bytes.NewReader(data))")
out.append("\tdecoder.DisallowUnknownFields()")
out.append("\tif err := decoder.Decode(&value); err != nil { return value, err }")
out.append("\tvar trailing any")
out.append("\tif err := decoder.Decode(&trailing); err != io.EOF { if err == nil { return value, errors.New(\"trailing JSON value\") }; return value, err }")
out.append("\tif err := value.Validate(); err != nil { return value, err }")
out.append("\treturn value, nil")
out.append("}")
out.append("")
out.append(f"func Encode{name}(value {name}) ([]byte, error) {{")
out.append("\tif err := value.Validate(); err != nil { return nil, err }")
out.append("\treturn json.Marshal(value)")
out.append("}")
out.append("")
# Use io.EOF in generated code without making every generated decoder depend on
# error-string comparison; replace the deliberately compact placeholder.
text = "\n".join(out).replace('"errors"\n"fmt"', '"errors"\n"fmt"\n\"io"')
text = text.replace('!errors.Is(err, errors.New("EOF")) && err != nil', '!errors.Is(err, io.EOF)')
return text + "\n"
def rust_type(prop: dict[str, Any]) -> str:
reference = ref_name(prop)
if reference:
return reference
if prop.get("type") == "array":
return f"Vec<{rust_type(prop.get('items', {}))}>"
if prop.get("type") == "integer":
return "i64"
if prop.get("type") == "boolean":
return "bool"
if prop.get("type") == "object":
return "JsonObject"
return "String"
def swift_type(prop: dict[str, Any]) -> str:
reference = ref_name(prop)
if reference:
return reference
if prop.get("type") == "array":
return f"[{swift_type(prop.get('items', {}))}]"
if prop.get("type") == "integer":
return "Int64"
if prop.get("type") == "boolean":
return "Bool"
if prop.get("type") == "object":
return "JSONObject"
return "String"
def generate_rust(defs: dict[str, dict[str, Any]], schema_hash: str, compatibility: dict[str, Any]) -> str:
out = [
"// Code generated by tools/generate.py; DO NOT EDIT.",
"#![allow(non_snake_case)]",
"pub const SCHEMA_SHA256: &str = \"" + schema_hash + "\";",
f'pub const CURRENT_WIRE_VERSION: &str = "{compatibility["current"]}";',
f'pub const N_MINUS_1_WIRE_VERSION: &str = "{compatibility["n_minus_1"]}";',
f'pub const N_MINUS_2_WIRE_VERSION: &str = "{compatibility["n_minus_2"]}";',
"pub type JsonObject = std::collections::BTreeMap<String, String>;",
"",
]
for name in sorted(defs):
definition = defs[name]
out.extend(["#[derive(Debug, Clone, PartialEq, Eq)]", f"pub struct {name} {{"])
required = set(definition.get("required", []))
for prop_name, prop in definition.get("properties", {}).items():
field = rust_field(prop_name)
typ = rust_type(prop)
if prop_name not in required:
typ = f"Option<{typ}>"
out.append(f" pub {field}: {typ},")
out.extend(["}", ""])
return "\n".join(out)
def generate_swift(defs: dict[str, dict[str, Any]], schema_hash: str, compatibility: dict[str, Any]) -> str:
out = [
"// Code generated by tools/generate.py; DO NOT EDIT.",
"import Foundation",
"public typealias JSONObject = [String: String]",
f"public let schemaSHA256 = \"{schema_hash}\"",
f'public let currentWireVersion = "{compatibility["current"]}"',
f'public let nMinus1WireVersion = "{compatibility["n_minus_1"]}"',
f'public let nMinus2WireVersion = "{compatibility["n_minus_2"]}"',
"",
]
for name in sorted(defs):
definition = defs[name]
required = set(definition.get("required", []))
out.extend(["public struct " + name + ": Codable, Equatable {",])
for prop_name, prop in definition.get("properties", {}).items():
typ = swift_type(prop)
if prop_name not in required:
typ += "?"
out.append(f" public let {swift_field(prop_name)}: {typ}")
out.append(" enum CodingKeys: String, CodingKey {")
for prop_name in definition.get("properties", {}):
out.append(f" case {swift_field(prop_name)} = \"{prop_name}\"")
out.extend([" }", "", " public init(from decoder: Decoder) throws {",])
out.append(" let c = try decoder.container(keyedBy: CodingKeys.self)")
for prop_name, prop in definition.get("properties", {}).items():
field = swift_field(prop_name)
typ = swift_type(prop)
if prop_name in required:
out.append(f" {field} = try c.decode({typ}.self, forKey: .{field})")
else:
out.append(f" {field} = try c.decodeIfPresent({typ}.self, forKey: .{field})")
out.extend([" }", "}", ""])
return "\n".join(out)
def write_or_check(path: pathlib.Path, content: str, check: bool) -> None:
if check:
if not path.exists() or path.read_text(encoding="utf-8") != content:
raise ValueError(f"generated output differs: {path.relative_to(ROOT)}")
return
path.parent.mkdir(parents=True, exist_ok=True)
path.write_text(content, encoding="utf-8")
def format_go(content: str) -> str:
result = subprocess.run(["gofmt"], input=content, text=True, capture_output=True, check=False)
if result.returncode != 0:
raise ValueError(f"gofmt failed: {result.stderr.strip()}")
return result.stdout
def main() -> int:
parser = argparse.ArgumentParser()
parser.add_argument("--check", action="store_true")
args = parser.parse_args()
schema, schema_hash, version, compatibility = load_schema()
generator_hash = hashlib.sha256(pathlib.Path(__file__).read_bytes()).hexdigest()
defs = schema.get("$defs")
if not isinstance(defs, dict) or not defs:
raise ValueError("schema must contain non-empty $defs")
normalized = {name: dict(value, name=name) for name, value in defs.items()}
outputs = {
ROOT / "gen/go/protocol/protocol.go": format_go(generate_go(normalized, schema_hash, version, compatibility)),
ROOT / "gen/rust/protocol.rs": generate_rust(normalized, schema_hash, compatibility).rstrip() + "\n",
ROOT / "gen/swift/Protocol.swift": generate_swift(normalized, schema_hash, compatibility).rstrip() + "\n",
ROOT / "gen/manifest.json": json.dumps({
"generator_sha256": generator_hash,
"schema_sha256": schema_hash,
"protocol_version": version,
"compatibility": compatibility,
}, sort_keys=True, indent=2) + "\n",
}
for path, content in outputs.items():
write_or_check(path, content, args.check)
return 0
if __name__ == "__main__":
try:
raise SystemExit(main())
except (OSError, ValueError, json.JSONDecodeError) as exc:
print(f"generate: {exc}", file=sys.stderr)
raise SystemExit(1)
+209
View File
@@ -0,0 +1,209 @@
package main
import (
"crypto/sha256"
"encoding/hex"
"encoding/json"
"fmt"
"os"
"path/filepath"
"strconv"
"strings"
protocol "github.com/sechmachine/VerseVDI-Protocol/gen/go/protocol"
)
const (
datagramHeaderBytes = 21
maximumFrameBytes = 65536
)
func main() {
entries, err := os.ReadDir("fixtures/conformance")
if err != nil {
panic(err)
}
var results []string
for _, entry := range entries {
if entry.IsDir() || filepath.Ext(entry.Name()) != ".tsv" {
continue
}
path := filepath.Join("fixtures/conformance", entry.Name())
data, err := os.ReadFile(path)
if err != nil {
panic(err)
}
lines := strings.Split(strings.TrimSuffix(string(data), "\n"), "\n")
if len(lines) == 0 || lines[0] != "id\tversion\tkind\tinput\texpected" {
panic("invalid fixture header")
}
for _, line := range lines[1:] {
fields := strings.Split(line, "\t")
if len(fields) != 5 {
panic("invalid fixture row")
}
actual := evaluate(fields[2], fields[3])
if actual != fields[4] {
panic(fmt.Sprintf("%s: got %s want %s", fields[0], actual, fields[4]))
}
results = append(results, fields[0]+"\t"+actual)
}
}
fixtureHash := readFixtureHash()
fmt.Printf("Go conformance passed normalized=%s fixtures=%s\n", normalizedDigest(results), fixtureHash)
}
func evaluate(kind, input string) string {
parts := map[string]string{}
for _, item := range strings.Split(input, ";") {
pair := strings.SplitN(item, "=", 2)
if len(pair) == 2 {
parts[pair[0]] = pair[1]
}
}
switch kind {
case "version":
if input == "1" || input == "0" || input == "-1" {
return "valid"
}
return "invalid:unsupported_version"
case "page":
limit, err := strconv.Atoi(parts["limit"])
value := protocol.PageInfo{Limit: int64(limit), NextCursor: parts["cursor"]}
if err == nil && value.Validate() == nil {
return "valid"
}
return "invalid:invalid_limit"
case "manifest":
forbidden := []string{"provider_url", "vm_address", "password", "private_key"}
for _, key := range forbidden {
if _, ok := parts[key]; ok {
return "invalid:forbidden_field"
}
}
value := protocol.ConnectionManifest{
Version: parts["version"], Purpose: parts["purpose"], SessionID: "session-1",
ReconnectSequence: 0,
Gateway: protocol.ManifestGateway{
ID: parts["gateway_id"], Addresses: []string{"gateway.control.test:443"}, PublicIdentity: parts["gateway_id"],
},
Tunnel: protocol.ManifestTunnel{Versions: []string{parts["protocol"] + "/1"}, Features: []string{"control.v1"}},
Profile: protocol.ManifestProfile{ID: "standard", Bounds: protocol.ManifestBounds{MinimumKbps: 1, TargetKbps: 2, MaximumKbps: 3}},
Grant: protocol.GrantReference{OpaqueValue: parts["grant"], ExpiresAt: parts["expires_at"], Audience: parts["audience"]},
CorrelationID: "correlation-1",
}
if value.Validate() == nil {
return "valid"
}
return "invalid:invalid_manifest"
case "clipboard":
_, hasFile := parts["file"]
value := protocol.ClipboardText{Text: parts["text"], Encoding: parts["encoding"]}
if !hasFile && value.Validate() == nil {
return "valid"
}
return "invalid:unsupported_clipboard"
case "event":
sequence, sequenceErr := strconv.ParseInt(parts["sequence"], 10, 64)
payloadBytes, payloadErr := strconv.Atoi(parts["payload_bytes"])
if parts["version"] != "1" {
return "invalid:unsupported_version"
}
if parts["after"] != "" && parts["earliest"] != "" {
after, afterErr := strconv.ParseInt(parts["after"], 10, 64)
earliest, earliestErr := strconv.ParseInt(parts["earliest"], 10, 64)
if afterErr == nil && earliestErr == nil && after > 0 && earliest > 0 && after < earliest-1 {
return "invalid:gap"
}
}
value := protocol.EventEnvelope{
EventID: "event-1", Sequence: sequence, Type: "broker.session.changed", Version: 1,
Resource: protocol.ResourceLink{Type: "broker_session", ID: "session-1", Version: 1},
OccurredAt: "2099-01-01T00:00:00Z", CorrelationID: parts["correlation_id"], Payload: map[string]any{},
}
if payloadErr != nil || payloadBytes > 16384 {
return "invalid:payload_limit"
}
if sequenceErr != nil || value.Validate() != nil {
return "invalid:required"
}
return "valid"
case "tunnel":
if (parts["offered"] == "1" || parts["offered"] == "0" || parts["offered"] == "-1") && parts["feature"] == "control.v1" {
return "valid"
}
if parts["feature"] != "control.v1" {
return "invalid:unsupported_feature"
}
return "invalid:unsupported_version"
case "datagram":
return classifyDatagram(parts["hex"])
default:
return "invalid:unknown_kind"
}
}
func classifyDatagram(encoded string) string {
raw, err := hex.DecodeString(encoded)
if err != nil {
return "invalid:hex"
}
if len(raw) < datagramHeaderBytes {
return "invalid:truncated"
}
if string(raw[:2]) != "VD" {
return "invalid:magic"
}
if raw[2] != 1 {
return "invalid:unsupported_version"
}
limits := map[byte]int{1: 1024, 2: 2048, 3: 65515}
limit, ok := limits[raw[3]]
if !ok {
return "invalid:unknown_channel"
}
if raw[4] != 0 {
return "invalid:flags"
}
if raw[18] == 0 || raw[17] >= raw[18] {
return "invalid:fragment"
}
payloadLength := int(raw[19])<<8 | int(raw[20])
if payloadLength > limit {
return "invalid:payload_limit"
}
if len(raw) != datagramHeaderBytes+payloadLength {
return "invalid:length_mismatch"
}
if len(raw) > maximumFrameBytes {
return "invalid:frame_limit"
}
return "valid"
}
func normalizedDigest(results []string) string {
const offset = uint64(14695981039346656037)
const prime = uint64(1099511628211)
value := offset
for _, result := range results {
for _, byteValue := range []byte(result + "\n") {
value ^= uint64(byteValue)
value *= prime
}
}
return fmt.Sprintf("%016x", value)
}
func readFixtureHash() string {
data, err := os.ReadFile("fixtures/manifest.json")
if err != nil {
panic(err)
}
var manifest struct {
CorpusSHA256 string `json:"corpus_sha256"`
}
if err := json.Unmarshal(data, &manifest); err != nil || len(manifest.CorpusSHA256) != sha256.Size*2 {
panic("invalid fixture manifest")
}
return manifest.CorpusSHA256
}
+154
View File
@@ -0,0 +1,154 @@
use std::fs;
use std::path::PathBuf;
fn values(input: &str) -> std::collections::BTreeMap<String, String> {
input
.split(';')
.filter_map(|item| item.split_once('='))
.map(|(key, value)| (key.to_owned(), value.to_owned()))
.collect()
}
fn evaluate(kind: &str, input: &str) -> &'static str {
let values = values(input);
match kind {
"version" if matches!(input, "1" | "0" | "-1") => "valid",
"version" => "invalid:unsupported_version",
"page" => match values.get("limit").and_then(|value| value.parse::<i64>().ok()) {
Some(limit) if (1..=100).contains(&limit) => "valid",
_ => "invalid:invalid_limit",
},
"manifest" if ["provider_url", "vm_address", "password", "private_key"]
.iter()
.any(|key| values.contains_key(*key)) => "invalid:forbidden_field",
"manifest"
if values.get("version").map(String::as_str) == Some("1")
&& values.contains_key("gateway_id")
&& values.get("grant").map_or(false, |value| value.len() >= 43)
&& values.get("purpose").map(String::as_str) == Some("launch") => "valid",
"manifest" => "invalid:invalid_manifest",
"clipboard" if values.get("encoding").map(String::as_str) == Some("utf-8")
&& !values.contains_key("file") => "valid",
"clipboard" => "invalid:unsupported_clipboard",
"event" if values.get("version").map(String::as_str) != Some("1") => {
"invalid:unsupported_version"
}
"event" if values.get("after").and_then(|value| value.parse::<i64>().ok()).is_some()
&& values.get("earliest").and_then(|value| value.parse::<i64>().ok()).is_some()
&& values["after"].parse::<i64>().unwrap() > 0
&& values["earliest"].parse::<i64>().unwrap() > 0
&& values["after"].parse::<i64>().unwrap() < values["earliest"].parse::<i64>().unwrap() - 1 => {
"invalid:gap"
}
"event" if values.get("payload_bytes").and_then(|value| value.parse::<usize>().ok()).map_or(true, |size| size > 16384) => {
"invalid:payload_limit"
}
"event" if values.get("sequence").and_then(|value| value.parse::<i64>().ok()).map_or(true, |sequence| sequence < 1)
|| !values.contains_key("correlation_id") => "invalid:required",
"event" => "valid",
"tunnel" if matches!(values.get("offered").map(String::as_str), Some("1") | Some("0") | Some("-1"))
&& values.get("feature").map(String::as_str) == Some("control.v1") => "valid",
"tunnel" if values.get("feature").map(String::as_str) != Some("control.v1") => {
"invalid:unsupported_feature"
}
"tunnel" => "invalid:unsupported_version",
"datagram" => classify_datagram(values.get("hex").map(String::as_str).unwrap_or_default()),
_ => "invalid:unknown_kind",
}
}
fn decode_hex(input: &str) -> Option<Vec<u8>> {
if input.len() % 2 != 0 {
return None;
}
(0..input.len())
.step_by(2)
.map(|index| u8::from_str_radix(&input[index..index + 2], 16).ok())
.collect()
}
fn classify_datagram(encoded: &str) -> &'static str {
let raw = match decode_hex(encoded) {
Some(raw) => raw,
None => return "invalid:hex",
};
if raw.len() < 21 {
return "invalid:truncated";
}
if raw[0..2] != *b"VD" {
return "invalid:magic";
}
if raw[2] != 1 {
return "invalid:unsupported_version";
}
let limit = match raw[3] {
1 => 1024,
2 => 2048,
3 => 65515,
_ => return "invalid:unknown_channel",
};
if raw[4] != 0 {
return "invalid:flags";
}
if raw[18] == 0 || raw[17] >= raw[18] {
return "invalid:fragment";
}
let payload_length = ((raw[19] as usize) << 8) | raw[20] as usize;
if payload_length > limit {
return "invalid:payload_limit";
}
if raw.len() != 21 + payload_length {
return "invalid:length_mismatch";
}
if raw.len() > 65536 {
return "invalid:frame_limit";
}
"valid"
}
fn normalized_digest(results: &[String]) -> String {
let mut value: u64 = 14695981039346656037;
for result in results {
for byte in format!("{result}\n").bytes() {
value ^= u64::from(byte);
value = value.wrapping_mul(1099511628211);
}
}
format!("{value:016x}")
}
fn fixture_hash() -> String {
let text = fs::read_to_string("fixtures/manifest.json").expect("fixture manifest");
text.split("\"corpus_sha256\": \"")
.nth(1)
.and_then(|value| value.split('"').next())
.expect("fixture hash")
.to_owned()
}
fn main() {
let mut paths: Vec<PathBuf> = fs::read_dir("fixtures/conformance")
.expect("fixture corpus")
.map(|entry| entry.expect("fixture entry").path())
.filter(|path| path.extension().and_then(|value| value.to_str()) == Some("tsv"))
.collect();
paths.sort();
let mut results = Vec::new();
for path in paths {
let text = fs::read_to_string(path).expect("fixture file");
let mut lines = text.lines();
assert_eq!(lines.next(), Some("id\tversion\tkind\tinput\texpected"));
for line in lines {
let fields: Vec<&str> = line.split('\t').collect();
assert_eq!(fields.len(), 5);
let actual = evaluate(fields[2], fields[3]);
assert_eq!(actual, fields[4], "{}", fields[0]);
results.push(format!("{}\t{}", fields[0], actual));
}
}
println!(
"Rust conformance passed normalized={} fixtures={}",
normalized_digest(&results),
fixture_hash()
);
}
+103
View File
@@ -0,0 +1,103 @@
import Foundation
func values(_ input: String) -> [String: String] {
var result: [String: String] = [:]
for item in input.split(separator: ";") {
let pair = item.split(separator: "=", maxSplits: 1).map(String.init)
if pair.count == 2 { result[pair[0]] = pair[1] }
}
return result
}
func evaluate(_ kind: String, _ input: String) -> String {
let values = values(input)
switch kind {
case "version": return ["1", "0", "-1"].contains(input) ? "valid" : "invalid:unsupported_version"
case "page":
guard let raw = values["limit"], let limit = Int(raw), (1...100).contains(limit) else { return "invalid:invalid_limit" }
return "valid"
case "manifest":
for key in ["provider_url", "vm_address", "password", "private_key"] where values[key] != nil { return "invalid:forbidden_field" }
return values["version"] == "1" && values["gateway_id"] != nil && (values["grant"]?.utf8.count ?? 0) >= 43 && values["purpose"] == "launch" ? "valid" : "invalid:invalid_manifest"
case "clipboard": return values["encoding"] == "utf-8" && values["file"] == nil ? "valid" : "invalid:unsupported_clipboard"
case "event":
guard values["version"] == "1" else { return "invalid:unsupported_version" }
if let after = Int(values["after"] ?? ""), let earliest = Int(values["earliest"] ?? ""), after > 0, earliest > 0, after < earliest - 1 { return "invalid:gap" }
if (Int(values["payload_bytes"] ?? "") ?? Int.max) > 16384 { return "invalid:payload_limit" }
guard let sequence = Int(values["sequence"] ?? ""), sequence > 0, values["correlation_id"] != nil else { return "invalid:required" }
return "valid"
case "tunnel":
if ["1", "0", "-1"].contains(values["offered"] ?? "") && values["feature"] == "control.v1" { return "valid" }
return values["feature"] == "control.v1" ? "invalid:unsupported_version" : "invalid:unsupported_feature"
case "datagram": return classifyDatagram(values["hex"] ?? "")
default: return "invalid:unknown_kind"
}
}
func classifyDatagram(_ encoded: String) -> String {
let characters = Array(encoded)
guard characters.count % 2 == 0 else { return "invalid:hex" }
var raw: [UInt8] = []
for index in stride(from: 0, to: characters.count, by: 2) {
guard let byte = UInt8(String(characters[index...index + 1]), radix: 16) else { return "invalid:hex" }
raw.append(byte)
}
guard raw.count >= 21 else { return "invalid:truncated" }
guard raw[0] == 0x56 && raw[1] == 0x44 else { return "invalid:magic" }
guard raw[2] == 1 else { return "invalid:unsupported_version" }
let limit: Int
switch raw[3] {
case 1: limit = 1024
case 2: limit = 2048
case 3: limit = 65515
default: return "invalid:unknown_channel"
}
guard raw[4] == 0 else { return "invalid:flags" }
guard raw[18] > 0 && raw[17] < raw[18] else { return "invalid:fragment" }
let payloadLength = Int(raw[19]) * 256 + Int(raw[20])
guard payloadLength <= limit else { return "invalid:payload_limit" }
guard raw.count == 21 + payloadLength else { return "invalid:length_mismatch" }
guard raw.count <= 65536 else { return "invalid:frame_limit" }
return "valid"
}
func normalizedDigest(_ results: [String]) -> String {
var value: UInt64 = 14695981039346656037
for result in results {
for byte in Array("\(result)\n".utf8) {
value ^= UInt64(byte)
value = value &* 1099511628211
}
}
return String(format: "%016llx", value)
}
func fixtureHash() -> String {
let text = try! String(contentsOfFile: "fixtures/manifest.json", encoding: .utf8)
let marker = "\"corpus_sha256\": \""
guard let start = text.range(of: marker)?.upperBound else { fatalError("fixture hash") }
let suffix = text[start...]
guard let end = suffix.firstIndex(of: "\"") else { fatalError("fixture hash") }
return String(suffix[..<end])
}
@main
struct ConformanceMain {
static func main() {
let names = try! FileManager.default.contentsOfDirectory(atPath: "fixtures/conformance").filter { $0.hasSuffix(".tsv") }.sorted()
var results: [String] = []
for name in names {
let text = try! String(contentsOfFile: "fixtures/conformance/\(name)", encoding: .utf8)
var lines = text.split(whereSeparator: \.isNewline).map(String.init)
precondition(lines.removeFirst() == "id\tversion\tkind\tinput\texpected")
for line in lines {
let fields = line.split(separator: "\t", omittingEmptySubsequences: false).map(String.init)
precondition(fields.count == 5)
let actual = evaluate(fields[2], fields[3])
precondition(actual == fields[4], fields[0])
results.append("\(fields[0])\t\(actual)")
}
}
print("Swift conformance passed normalized=\(normalizedDigest(results)) fixtures=\(fixtureHash())")
}
}
+33
View File
@@ -0,0 +1,33 @@
#!/usr/bin/env python3
from __future__ import annotations
import pathlib
import subprocess
import tempfile
ROOT = pathlib.Path(__file__).resolve().parents[1]
def run(command: list[str]) -> None:
result = subprocess.run(command, cwd=ROOT, text=True, capture_output=True)
if result.returncode != 0:
raise SystemExit(result.stdout + result.stderr)
print(result.stdout.strip())
def main() -> int:
with tempfile.TemporaryDirectory(prefix="versevdi-protocol-conformance-") as directory:
temp = pathlib.Path(directory)
rust_bin = temp / "rust-conformance"
swift_bin = temp / "swift-conformance"
run(["rustc", "tools/native_conformance.rs", "-O", "-o", str(rust_bin)])
run([str(rust_bin)])
main_source = temp / "main.swift"
main_source.write_text((ROOT / "tools/native_conformance.swift").read_text(encoding="utf-8"), encoding="utf-8")
run(["swiftc", "-parse-as-library", "gen/swift/Protocol.swift", str(main_source), "-o", str(swift_bin)])
run([str(swift_bin)])
return 0
if __name__ == "__main__":
raise SystemExit(main())
+80
View File
@@ -0,0 +1,80 @@
#!/usr/bin/env python3
"""Dependency-free structural and scope validation for the Protocol sources."""
from __future__ import annotations
import json
import pathlib
import hashlib
import re
import sys
ROOT = pathlib.Path(__file__).resolve().parents[1]
def main() -> int:
schema_path = ROOT / "schemas/control-v1.schema.json"
schema = json.loads(schema_path.read_text(encoding="utf-8"))
assert schema["$schema"].endswith("2020-12/schema")
defs = schema["$defs"]
assert len(defs) >= 16
for name, definition in defs.items():
assert definition["type"] == "object", name
assert definition["additionalProperties"] is False, name
assert set(definition["required"]).issubset(definition["properties"]), name
compatibility = json.loads((ROOT / "compatibility.json").read_text(encoding="utf-8"))
assert set([compatibility["current"], compatibility["n_minus_1"], compatibility["n_minus_2"]]) == {"1", "0", "-1"}
assert len(set(compatibility["unsupported"])) == len(compatibility["unsupported"])
for registry in ("registries/features.json", "registries/datagrams.json"):
value = json.loads((ROOT / registry).read_text(encoding="utf-8"))
entries = value.get("features", value.get("datagrams"))
assert entries and len({entry["id"] for entry in entries}) == len(entries)
for entry in entries:
maximum = entry.get("max_frame_bytes", entry.get("max_payload_bytes"))
assert isinstance(maximum, int) and 1 <= maximum <= 65536
manifest = json.loads((ROOT / "fixtures/valid/manifest.json").read_text(encoding="utf-8"))
assert set(manifest).issubset(set(defs["ConnectionManifest"]["properties"]))
forbidden = json.loads((ROOT / "fixtures/invalid/manifest-provider-field.json").read_text(encoding="utf-8"))
assert "provider_url" not in defs["ConnectionManifest"]["properties"] and "provider_url" in forbidden
expected_header = "id\tversion\tkind\tinput\texpected"
ids = set()
for fixture_path in sorted((ROOT / "fixtures/conformance").glob("*.tsv")):
lines = fixture_path.read_text(encoding="utf-8").splitlines()
assert lines and lines[0] == expected_header, fixture_path
for line in lines[1:]:
fields = line.split("\t")
assert len(fields) == 5, line
assert fields[0] not in ids, fields[0]
ids.add(fields[0])
assert fields[4] == "valid" or fields[4].startswith("invalid:"), line
fixture_manifest = json.loads((ROOT / "fixtures/manifest.json").read_text(encoding="utf-8"))
assert fixture_manifest["files"] == sorted(
path.relative_to(ROOT).as_posix() for path in (ROOT / "fixtures/conformance").glob("*.tsv")
)
fixture_hash = hashlib.sha256()
for relative in fixture_manifest["files"]:
fixture_hash.update(relative.encode("utf-8"))
fixture_hash.update(b"\0")
fixture_hash.update((ROOT / relative).read_bytes())
fixture_hash.update(b"\0")
assert fixture_manifest["corpus_sha256"] == fixture_hash.hexdigest()
openapi = (ROOT / "openapi/control-v1.yaml").read_text(encoding="utf-8")
assert "openapi: 3.1.0" in openapi
assert "/api/v1/auth/refresh:" in openapi and "/api/v1/resources:" in openapi and "/api/v1/events:" in openapi
assert "provider_url" not in openapi and "vm_address" not in openapi
print("Protocol source validation passed")
return 0
if __name__ == "__main__":
try:
raise SystemExit(main())
except (AssertionError, OSError, json.JSONDecodeError) as exc:
print(f"validate: {exc}", file=sys.stderr)
raise SystemExit(1)
+55
View File
@@ -0,0 +1,55 @@
#!/usr/bin/env python3
"""Validate the bounded Phase 3A datagram header and fixture corpus."""
from __future__ import annotations
import binascii
import pathlib
ROOT = pathlib.Path(__file__).resolve().parents[1]
HEADER_BYTES = 21
MAX_FRAME_BYTES = 65536
CHANNEL_LIMITS = {1: 1024, 2: 2048, 3: 65515}
def classify(raw: bytes) -> str:
if len(raw) < HEADER_BYTES:
return "invalid:truncated"
if raw[:2] != b"VD":
return "invalid:magic"
if raw[2] != 1:
return "invalid:unsupported_version"
if raw[3] not in CHANNEL_LIMITS:
return "invalid:unknown_channel"
if raw[4] != 0:
return "invalid:flags"
fragment_index, fragment_count = raw[17], raw[18]
if fragment_count == 0 or fragment_index >= fragment_count:
return "invalid:fragment"
payload_length = int.from_bytes(raw[19:21], "big")
if payload_length > CHANNEL_LIMITS[raw[3]]:
return "invalid:payload_limit"
if len(raw) != HEADER_BYTES + payload_length:
return "invalid:length_mismatch"
if len(raw) > MAX_FRAME_BYTES:
return "invalid:frame_limit"
return "valid"
def main() -> None:
lines = (ROOT / "fixtures/conformance/datagram-v1.tsv").read_text(encoding="utf-8").splitlines()
assert lines[0] == "id\tversion\tkind\tinput\texpected"
for line in lines[1:]:
identifier, version, kind, input_value, expected = line.split("\t")
assert kind == "datagram" and version == "1"
encoded = input_value.removeprefix("hex=")
try:
actual = classify(binascii.unhexlify(encoded))
except binascii.Error:
actual = "invalid:hex"
assert actual == expected, f"{identifier}: {actual} != {expected}"
print("Datagram frame validation passed")
if __name__ == "__main__":
main()