feat: implement peer-id spec by GrapeBaBa · Pull Request #3 · blockblaz/peer-id

GrapeBaBa · 2025-08-17T09:41:07Z

This pull request introduces a Zig implementation of the libp2p peer-id specification, including the core data structures, build configuration, and continuous integration setup. The most significant changes are the addition of the main PeerId logic, Zig build system files, documentation, and a GitHub Actions workflow for CI. These changes lay the foundation for a reusable Zig package that can generate, parse, and manipulate peer IDs compatible with libp2p.

Core PeerId implementation:

Added src/id.zig with a full implementation of the PeerId struct, supporting creation from public keys, multihash, bytes, base58 strings, and CIDs, as well as conversion and equality checks. Includes comprehensive unit tests for all major functionality.

Build system and dependencies:

Added build.zig and build.zig.zon to configure Zig package building, dependency management (zmultiformats and gremlin), documentation generation, and unit testing. [1] [2]

Documentation:

Updated README.md with project overview, prerequisites, build/test instructions, documentation generation, and usage guide for integrating the package into other Zig projects.

Protobuf schema for keys:

Added src/keys.proto to define the protobuf schema for public and private keys, supporting multiple key types (RSA, Ed25519, Secp256k1, ECDSA, Curve25519).

Continuous Integration:

Added .github/workflows/ci.yml to set up GitHub Actions for Zig version management, caching, formatting checks, unit testing, and release builds on push and pull request events.

Signed-off-by: Chen Kai <281165273grape@gmail.com>

Copilot

Pull Request Overview

This PR implements a Zig library for libp2p peer-id specification, providing core data structures for creating, parsing, and manipulating peer IDs compatible with libp2p networks.

Core PeerId implementation with support for multihash, CID, and base58 encoding/decoding
Protobuf-based public/private key structures with serialization support
Complete build system with dependencies, testing, and documentation generation

Reviewed Changes

Copilot reviewed 8 out of 9 changed files in this pull request and generated 3 comments.

Show a summary per file

File	Description
src/root.zig	Module entry point exposing key types and PeerId functionality
src/keys.proto.zig	Generated protobuf bindings for public/private key structures
src/keys.proto	Protobuf schema defining key types and message structures
src/id.zig	Core PeerId implementation with conversion, validation, and utility methods
build.zig.zon	Package configuration with dependencies and metadata
build.zig	Build system configuration with protobuf generation and module setup
README.md	Documentation with build instructions and usage examples
.github/workflows/ci.yml	CI pipeline for testing, formatting, and building

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

Copilot · 2025-08-17T09:41:54Z

src/id.zig

+        var peer_id_bytes: [32]u8 = undefined;
+        std.crypto.random.bytes(&peer_id_bytes);
+
+        const mh = try Multihash(64).wrap(MULTIHASH_IDENTITY_CODE, &peer_id_bytes);


The random() function uses MULTIHASH_IDENTITY_CODE with a 32-byte random array, but identity multihash should only be used for keys <= 42 bytes that are actual protobuf-encoded public keys. Random bytes are not valid protobuf-encoded keys and should use SHA2-256 instead.

Suggested change

const mh = try Multihash(64).wrap(MULTIHASH_IDENTITY_CODE, &peer_id_bytes);

var digest: [32]u8 = undefined;

std.crypto.hash.sha2.Sha256.hash(&peer_id_bytes, &digest, .{});

const mh = try Multihash(64).wrap(MULTIHASH_SHA256_CODE, &digest);

i think this makes sense what do you say @GrapeBaBa ?

No, i don't get this comment, how is it related to protobuf?

what it might be saying is that the randomness in the bytes might not be enough so we need to use sha as randomizer oracle. wdyt @gballet ?

this is only for testing, port from other impl, should not use it in real case. I think it is used to create a fake/mock peerid using MULTIHASH_IDENTITY_CODE approach, the real case should always use a public key

Copilot · 2025-08-17T09:41:55Z

src/id.zig

+        if (std.mem.startsWith(u8, s, "Qm") or std.mem.startsWith(u8, s, "1")) {
+            return try Self.fromBase58(allocator, s);
+        }
+


The string parsing logic is incomplete. Base58btc encoded multihashes can start with characters other than 'Qm' or '1' depending on the hash algorithm and length. This will incorrectly reject valid base58btc peer IDs.

Suggested change

// Try to decode as base58btc multihash first

const decoded_len = multibase.MultiBaseCodec.Base58Btc.decodedLen(s);

var decoded_bytes = allocator.alloc(u8, decoded_len) catch null;

if (decoded_bytes) |bytes| {

defer allocator.free(bytes);

if (multibase.MultiBaseCodec.Base58Impl.decodeBtc(bytes, s)) |actual_decoded| {

// Try to parse as multihash

if (Multihash(64).fromBytes(actual_decoded)) |mh| {

return PeerId{ .multihash = mh };

} else |err| {

// Not a valid multihash, fall through to CID parsing

}

} else |err| {

// Not valid base58btc, fall through to CID parsing

}

}

// Try to parse as CID

does this makes sense @GrapeBaBa ?

No, https://github.com/libp2p/specs/blob/master/peer-ids/peer-ids.md#string-representation only two valid prefix, Qm for MULTIHASH_SHA256_CODE and 1 for MULTIHASH_IDENTITY_CODE

Copilot · 2025-08-17T09:41:55Z

README.md

+    .optimize = optimize,
+});
+const peer_id_module = peer_id_dep.module("peer_id");
+root_module.addImport("peer_id", peer_id_module);


The module name 'peer_id' doesn't match the module name 'peer-id' defined in build.zig line 29. This will cause a module not found error.

Suggested change

root_module.addImport("peer_id", peer_id_module);

const peer_id_dep = b.dependency("peer-id", .{

.target = target,

.optimize = optimize,

});

const peer_id_module = peer_id_dep.module("peer-id");

root_module.addImport("peer-id", peer_id_module);

g11tech · 2025-08-22T14:58:23Z

src/id.zig

+    try testing.expect(!peer1.eql(&peer3));
+}
+
+// test "PeerId from public key to CID complete flow" {


we don't need this test? @GrapeBaBa

This test actually works, but it introduced ssl dep. I don't find a easy way to exclude this ssl dep when I import this lib to zig-libp2p. It will cause a conflict because another dep lsquic also depends on ssl. I think i can make this test work without ssl later, so commented out right now.

can't we just use/link the dependency in the test step, but merging the PR for now

g11tech

awesome work @GrapeBaBa !

GrapeBaBa added 5 commits August 14, 2025 18:42

feat:init pee-ids spec

6a631f8

Signed-off-by: Chen Kai <281165273grape@gmail.com>

feat:add ci

ae15796

Signed-off-by: Chen Kai <281165273grape@gmail.com>

feat: update readme

ea429f0

Signed-off-by: Chen Kai <281165273grape@gmail.com>

fix:update ci yml

209b8d7

Signed-off-by: Chen Kai <281165273grape@gmail.com>

fix:remove ssl test dep

3f9be40

Signed-off-by: Chen Kai <281165273grape@gmail.com>

Copilot AI review requested due to automatic review settings August 17, 2025 09:41

Copilot AI reviewed Aug 17, 2025

View reviewed changes

GrapeBaBa requested review from RazorClient, bhaskar1001101, g11tech, gballet, guha-rahul, parithosh and spiral-ladder August 17, 2025 09:43

g11tech reviewed Aug 22, 2025

View reviewed changes

g11tech approved these changes Aug 23, 2025

View reviewed changes

g11tech merged commit b8f98a4 into main Aug 23, 2025
1 check passed

g11tech deleted the peer_id branch August 23, 2025 08:08

-        const mh = try Multihash(64).wrap(MULTIHASH_IDENTITY_CODE, &peer_id_bytes);
+        var digest: [32]u8 = undefined;
+        std.crypto.hash.sha2.Sha256.hash(&peer_id_bytes, &digest, .{});
+        const mh = try Multihash(64).wrap(MULTIHASH_SHA256_CODE, &digest);

+        // Try to decode as base58btc multihash first
+        const decoded_len = multibase.MultiBaseCodec.Base58Btc.decodedLen(s);
+        var decoded_bytes = allocator.alloc(u8, decoded_len) catch null;
+        if (decoded_bytes) |bytes| {
+            defer allocator.free(bytes);
+            if (multibase.MultiBaseCodec.Base58Impl.decodeBtc(bytes, s)) |actual_decoded| {
+                // Try to parse as multihash
+                if (Multihash(64).fromBytes(actual_decoded)) |mh| {
+                    return PeerId{ .multihash = mh };
+                } else |err| {
+                    // Not a valid multihash, fall through to CID parsing
+                }
+            } else |err| {
+                // Not valid base58btc, fall through to CID parsing
+            }
+        }
+        // Try to parse as CID

-root_module.addImport("peer_id", peer_id_module);
+const peer_id_dep = b.dependency("peer-id", .{
+    .target = target,
+    .optimize = optimize,
+});
+const peer_id_module = peer_id_dep.module("peer-id");
+root_module.addImport("peer-id", peer_id_module);

Comments

Conversation

GrapeBaBa commented Aug 17, 2025

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull Request Overview

Reviewed Changes

Uh oh!

Copilot AI Aug 17, 2025

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

GrapeBaBa Aug 23, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Copilot AI Aug 17, 2025

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Copilot AI Aug 17, 2025

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

g11tech left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

GrapeBaBa Aug 23, 2025 •

edited

Loading