Add Gloas consensus type block package #15618

terencechain · 2025-08-21T20:19:42Z

No description provided.

james-prysm · 2025-11-10T21:55:04Z

proto/engine/v1/gloas.pb.go

+	_ = protoimpl.EnforceVersion(protoimpl.MaxVersion - 20)
+)
+
+type ExecutionPayloadEnvelope struct {


hmm why was this file included without the protobuf file?

nice catch, the file was a left over and i checked it in by accident
in the latest commit i have that removed and addressed your other feedback

james-prysm · 2025-11-10T21:56:58Z

proto/prysm/v1alpha1/cloners.go

+		copied.SyncAggregate = body.SyncAggregate.Copy()
+	}
+
+	copied.SignedExecutionPayloadBid = copySignedExecutionPayloadBid(body.SignedExecutionPayloadBid)


nit can we move this to the bottom since it's new?

james-prysm

LGTM

terencechain · 2025-11-18T23:27:14Z

beacon-chain/blockchain/process_block_test.go

+		if testVersion == version.Gloas {
+			// TODO(16027): Unskip light client tests for Gloas
+			continue
+		}


note we are going to redo this, we are experimenting a different approach in: #16030
but it shouldn't block anyone reviewing this PR

potuz

A bunch of getters should now error on Gloas, at least Execution, KZG commitments, Execution Requests.

First serious pass, will do another in a couple of hours.

potuz · 2025-11-20T14:17:30Z

consensus-types/blocks/factory.go

@@ -82,13 +82,15 @@ func NewSignedBeaconBlock(i any) (interfaces.SignedBeaconBlock, error) {
 		return initBlindedSignedBlockFromProtoFulu(b)
 	case *eth.GenericSignedBeaconBlock_BlindedFulu:
 		return initBlindedSignedBlockFromProtoFulu(b.BlindedFulu)
+	case *eth.SignedBeaconBlockGloas:


Do we need a "Generic" version like for previous forks, something like b.Gloas.Block?

potuz · 2025-11-20T14:18:07Z

consensus-types/blocks/factory.go

@@ -138,13 +140,15 @@ func NewBeaconBlock(i any) (interfaces.ReadOnlyBeaconBlock, error) {
 		return initBlindedBlockFromProtoFulu(b)
 	case *eth.GenericBeaconBlock_BlindedFulu:
 		return initBlindedBlockFromProtoFulu(b.BlindedFulu)
+	case *eth.BeaconBlockGloas:


Do we need a generic version like previous forks? b.Gloas.Block

potuz · 2025-11-20T14:21:56Z

consensus-types/blocks/factory.go

@@ -629,6 +641,8 @@ func BuildSignedBeaconBlockFromExecutionPayload(blk interfaces.ReadOnlySignedBea
 			},
 			Signature: sig[:],
 		}
+	case version.Gloas:


This case should fail much earlier when trying to get the Execution from the body. It's better to return early from this function with an error. Also, the fact that this function passed tests shows that this couldn't have been covered. So a regression test is needed.

potuz · 2025-11-20T14:24:53Z

consensus-types/blocks/getters.go

@@ -157,14 +159,17 @@ func (b *SignedBeaconBlock) PbGenericBlock() (*eth.GenericSignedBeaconBlock, err
 		return &eth.GenericSignedBeaconBlock{
 			Block: &eth.GenericSignedBeaconBlock_Fulu{Fulu: bc},
 		}, nil
+	case version.Gloas:
+		// Gloas doesn't support GenericSignedBeaconBlock yet


This answers my pervious questions.... Probably better to discuss this offline

potuz · 2025-11-20T14:40:18Z

consensus-types/blocks/getters.go

+		if err := pb.UnmarshalSSZ(buf); err != nil {
+			return err
+		}
+		var err error


I know this is the design on all previous cases, but is there any reason to not have done:

Suggested change

if err := pb.UnmarshalSSZ(buf); err != nil {

return err

}

var err error

err := pb.UnmarshalSSZ(buf);

if err != nil {

return err

}

potuz · 2025-11-20T14:45:04Z

consensus-types/blocks/proto.go

+		signedExecutionPayloadBid: pb.SignedExecutionPayloadBid,
+		blsToExecutionChanges:     pb.BlsToExecutionChanges,


Why the ordering?

potuz · 2025-11-20T14:47:01Z

consensus-types/blocks/setters.go

@@ -182,3 +182,21 @@ func (b *SignedBeaconBlock) SetExecutionRequests(req *enginev1.ExecutionRequests
 	b.block.body.executionRequests = req
 	return nil
 }
+


SetExecutionRequests should also fail if the version is Gloas or more, same for the Execution Payload setter.

potuz · 2025-11-20T14:50:09Z

consensus-types/blocks/getters.go

@@ -80,6 +80,8 @@ func (b *SignedBeaconBlock) Copy() (interfaces.SignedBeaconBlock, error) {
 			return initBlindedSignedBlockFromProtoFulu(pb.(*eth.SignedBlindedBeaconBlockFulu).Copy())
 		}
 		return initSignedBlockFromProtoFulu(pb.(*eth.SignedBeaconBlockFulu).Copy())
+	case version.Gloas:


Not related to this line. But the getters file needs to be futher modified. Some getters do not make sense for Gloas and instead of returning nil and risking a panic they should return an error. For example the current path for Execution() will return nil instead of erroring out. The same for the ExecutionRequests(). We should add errors for all getters that would fail.

potuz · 2025-11-20T14:51:56Z

consensus-types/blocks/types.go

+	payloadAttestations       []*eth.PayloadAttestation
+	signedExecutionPayloadBid *eth.SignedExecutionPayloadBid


Reverse order here.

**Review after #16033 is merged** **What type of PR is this?** Other **What does this PR do? Why is it needed?** This PR refactors the code to find the corresponding slot of a given block root using state summary or the block itself, into its own function `SlotByBlockRoot(ctx, blockroot)`. Note that there exists a function `slotByBlockRoot(ctx context.Context, tx *bolt.Tx, blockRoot []byte)` immediately below the new function. Also note that this function has two drawbacks, which led to creation of the new function: - the old function requires a boltdb tx, which is not necessarily available to the caller. - the old function does NOT make use of the state summary cache. edit: - the old function also uses the state bucket to retrieve the state and it's slot. this is not something we want in the state diff feature, since there is no state bucket.

terencechain added the gloas label Aug 21, 2025

terencechain force-pushed the add-gloas-block branch 2 times, most recently from da955a3 to 8bbcf1f Compare August 31, 2025 03:16

terencechain force-pushed the add-gloas-block branch from 8bbcf1f to 8a4a318 Compare September 20, 2025 02:55

terencechain force-pushed the add-gloas-block branch from 8a4a318 to 5df3604 Compare October 9, 2025 20:48

terencechain force-pushed the add-gloas-block branch from 5df3604 to 1869386 Compare November 5, 2025 16:22

terencechain marked this pull request as ready for review November 5, 2025 16:23

terencechain force-pushed the add-gloas-block branch 3 times, most recently from 8ecec5d to 4370e34 Compare November 5, 2025 21:10

Add Gloas protobuf definitions with spec tests

5c286ed

terencechain force-pushed the add-gloas-block branch from 4370e34 to 6333253 Compare November 10, 2025 16:19

james-prysm reviewed Nov 10, 2025

View reviewed changes

terencechain force-pushed the add-gloas-block branch from 6333253 to 693b303 Compare November 11, 2025 00:13

terencechain force-pushed the add-gloas-block branch from 693b303 to 8bd5e43 Compare November 18, 2025 03:57

Implement Gloas fork support in consensus-types/blocks

09ec764

terencechain force-pushed the add-gloas-block branch from 8bd5e43 to 09ec764 Compare November 18, 2025 04:11

james-prysm previously approved these changes Nov 18, 2025

View reviewed changes

terencechain commented Nov 18, 2025

View reviewed changes

potuz reviewed Nov 20, 2025

View reviewed changes

Potuz's feedback

f798338

terencechain dismissed james-prysm’s stale review via f798338 November 20, 2025 16:03

Go fmt

9646238

potuz previously approved these changes Nov 24, 2025

View reviewed changes

Better conditions for Execution()

dca25e2

terencechain dismissed potuz’s stale review via dca25e2 November 24, 2025 18:07

IsExecutionEnabled for post capella state

52b444a

potuz previously approved these changes Nov 24, 2025

View reviewed changes

terencechain enabled auto-merge November 24, 2025 19:11

terencechain and others added 3 commits November 24, 2025 14:42

Merge branch 'develop' into add-gloas-block

3bba46e

Check exec enabled condition to capella

5692778

terencechain dismissed potuz’s stale review via 5692778 November 24, 2025 23:01

Revert test want to false for empty bellatrix case

39dcd63

potuz approved these changes Nov 25, 2025

View reviewed changes

terencechain added this pull request to the merge queue Nov 25, 2025

Merged via the queue into develop with commit 5bbdebe Nov 25, 2025
18 checks passed

terencechain deleted the add-gloas-block branch November 25, 2025 09:29

		signedExecutionPayloadBid: pb.SignedExecutionPayloadBid,
		blsToExecutionChanges: pb.BlsToExecutionChanges,

		payloadAttestations []*eth.PayloadAttestation
		signedExecutionPayloadBid *eth.SignedExecutionPayloadBid

Add Gloas consensus type block package #15618

Add Gloas consensus type block package #15618

Uh oh!

Conversation

terencechain commented Aug 21, 2025

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!

james-prysm left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

potuz left a comment

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!

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!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants