Review fixes for the Bridging vs. NAC section (architecture.md) by saroupille · Pull Request #450 · etherlinkcom/docs

saroupille · 2026-08-19T09:22:48Z

Addresses the open review thread on #436 asking for a look at the new "Bridging vs. NAC" section. Targets the import-tx-doc branch.

Fixes:

Michelson runtime label: the first diagram said "Tezos JSON-RPC" — the Tezos RPC is not JSON-RPC (the interface table above correctly says "Tezos RPC").
L1 node & anchoring edges: "Tezos Layer1 (Smart Rollup)" read as if L1 were the rollup — Etherlink is the Smart Rollup, L1 hosts it. The runtimes also don't individually post to L1, so the two per-runtime edges are replaced by a single "anchored as a Smart Rollup" edge from the chain.
Sequence diagram note: "One atomic block" → "One atomic transaction" — the atomicity unit is the transaction, not the block.
Comparison table: native tez is not wrapped by the L1↔L2 bridge (only FA tokens are); withdrawal latency is days (without fast withdrawals), not hours.
Cosmetic: stray colons/spacing in mermaid labels, "a unified execution layer", "economic space".

🤖 Generated with Claude Code

- Michelson runtime exposes the Tezos RPC, not JSON-RPC (matches the interface table above) - L1 node: Etherlink is the Smart Rollup, L1 hosts it; anchor the whole chain to L1 with a single edge instead of per-runtime edges - sequence diagram note: the atomicity unit is the transaction, not the block - comparison table: native tez is not wrapped by the L1<->L2 bridge; withdrawal latency is days, not hours - cosmetic: stray colons/spacing in mermaid labels, 'a unified execution layer', 'economic space' Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

vercel · 2026-08-19T09:22:53Z

@saroupille is attempting to deploy a commit to the Trili Tech Team on Vercel.

A member of the Team first needs to authorize it.

The 13-August/slow-governance line was a placeholder. What actually happened: the slow-governance proposal of 2 August was withdrawn before activation after a security vulnerability was found during continued testing (Mainnet was never exposed), and a patched kernel was resubmitted via fast governance on 19 August, with the upgrade triggered on 20 August. Also: 'Native Atomic Calls' -> 'Native Atomic Composability (NAC)' (terminology used everywhere else in the docs), and a link to the resubmission Agora post. Note: the 20 August date assumes the trigger happens as planned on Thursday morning; re-verify against the actual trigger_kernel_upgrade operation before merging. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

- accounts-and-aliases: 'the the' typo - fees: quantify the EIP-7702 DA-fee term (125 bytes per authorization, matching fees.rs) - terminology: 'Native Atomic Calls' -> 'Native Atomic Composability (NAC)' (bridging.md, network-information.mdx, testnet.mdx) - testnet.mdx: Previewnet is long-running, not a 'temporary testnet' (same fix previewnet.md already received) Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

lthms · 2026-08-19T09:54:09Z

@@ -42,5 +42,5 @@ Examples of using test networks are on other pages in this documentation.

 ## Etherlink{/* TX */} Previewnet


Strictly speaking it's called Tezos X Previewnet and i don't think there is a need to “downcast” it to Etherlink Previewnet? or at least no decision was made about that

saroupille · 2026-08-19T11:58:43Z

Heads-up @lthms: this PR was merged at 09:57 UTC, but a 4th commit (41e4711, Network information Michelson: chain ids, /tezlink prefix, bridge/faucet/indexer) had been pushed to the branch just after — it missed the merge. Since we now have write access, it has been pushed directly to import-tx-doc (fast-forward a29ec13..41e4711). Nothing else pending here.

saroupille · 2026-08-19T09:22:48Z

Addresses the open review thread on #436 asking for a look at the new "Bridging vs. NAC" section. Targets the import-tx-doc branch.

Fixes:

Michelson runtime label: the first diagram said "Tezos JSON-RPC" — the Tezos RPC is not JSON-RPC (the interface table above correctly says "Tezos RPC").
L1 node & anchoring edges: "Tezos Layer1 (Smart Rollup)" read as if L1 were the rollup — Etherlink is the Smart Rollup, L1 hosts it. The runtimes also don't individually post to L1, so the two per-runtime edges are replaced by a single "anchored as a Smart Rollup" edge from the chain.
Sequence diagram note: "One atomic block" → "One atomic transaction" — the atomicity unit is the transaction, not the block.
Comparison table: native tez is not wrapped by the L1↔L2 bridge (only FA tokens are); withdrawal latency is days (without fast withdrawals), not hours.
Cosmetic: stray colons/spacing in mermaid labels, "a unified execution layer", "economic space".

🤖 Generated with Claude Code

- Michelson runtime exposes the Tezos RPC, not JSON-RPC (matches the interface table above) - L1 node: Etherlink is the Smart Rollup, L1 hosts it; anchor the whole chain to L1 with a single edge instead of per-runtime edges - sequence diagram note: the atomicity unit is the transaction, not the block - comparison table: native tez is not wrapped by the L1<->L2 bridge; withdrawal latency is days, not hours - cosmetic: stray colons/spacing in mermaid labels, 'a unified execution layer', 'economic space' Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

vercel · 2026-08-19T09:22:53Z

@saroupille is attempting to deploy a commit to the Trili Tech Team on Vercel.

A member of the Team first needs to authorize it.

The 13-August/slow-governance line was a placeholder. What actually happened: the slow-governance proposal of 2 August was withdrawn before activation after a security vulnerability was found during continued testing (Mainnet was never exposed), and a patched kernel was resubmitted via fast governance on 19 August, with the upgrade triggered on 20 August. Also: 'Native Atomic Calls' -> 'Native Atomic Composability (NAC)' (terminology used everywhere else in the docs), and a link to the resubmission Agora post. Note: the 20 August date assumes the trigger happens as planned on Thursday morning; re-verify against the actual trigger_kernel_upgrade operation before merging. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

- accounts-and-aliases: 'the the' typo - fees: quantify the EIP-7702 DA-fee term (125 bytes per authorization, matching fees.rs) - terminology: 'Native Atomic Calls' -> 'Native Atomic Composability (NAC)' (bridging.md, network-information.mdx, testnet.mdx) - testnet.mdx: Previewnet is long-running, not a 'temporary testnet' (same fix previewnet.md already received) Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

lthms · 2026-08-19T09:54:09Z

@@ -42,5 +42,5 @@ Examples of using test networks are on other pages in this documentation.

 ## Etherlink{/* TX */} Previewnet


Strictly speaking it's called Tezos X Previewnet and i don't think there is a need to “downcast” it to Etherlink Previewnet? or at least no decision was made about that

saroupille · 2026-08-19T11:58:43Z

Heads-up @lthms: this PR was merged at 09:57 UTC, but a 4th commit (41e4711, Network information Michelson: chain ids, /tezlink prefix, bridge/faucet/indexer) had been pushed to the branch just after — it missed the merge. Since we now have write access, it has been pushed directly to import-tx-doc (fast-forward a29ec13..41e4711). Nothing else pending here.

saroupille · 2026-08-19T09:22:48Z

Addresses the open review thread on #436 asking for a look at the new "Bridging vs. NAC" section. Targets the import-tx-doc branch.

Fixes:

Michelson runtime label: the first diagram said "Tezos JSON-RPC" — the Tezos RPC is not JSON-RPC (the interface table above correctly says "Tezos RPC").
L1 node & anchoring edges: "Tezos Layer1 (Smart Rollup)" read as if L1 were the rollup — Etherlink is the Smart Rollup, L1 hosts it. The runtimes also don't individually post to L1, so the two per-runtime edges are replaced by a single "anchored as a Smart Rollup" edge from the chain.
Sequence diagram note: "One atomic block" → "One atomic transaction" — the atomicity unit is the transaction, not the block.
Comparison table: native tez is not wrapped by the L1↔L2 bridge (only FA tokens are); withdrawal latency is days (without fast withdrawals), not hours.
Cosmetic: stray colons/spacing in mermaid labels, "a unified execution layer", "economic space".

🤖 Generated with Claude Code

- Michelson runtime exposes the Tezos RPC, not JSON-RPC (matches the interface table above) - L1 node: Etherlink is the Smart Rollup, L1 hosts it; anchor the whole chain to L1 with a single edge instead of per-runtime edges - sequence diagram note: the atomicity unit is the transaction, not the block - comparison table: native tez is not wrapped by the L1<->L2 bridge; withdrawal latency is days, not hours - cosmetic: stray colons/spacing in mermaid labels, 'a unified execution layer', 'economic space' Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

vercel · 2026-08-19T09:22:53Z

@saroupille is attempting to deploy a commit to the Trili Tech Team on Vercel.

A member of the Team first needs to authorize it.

The 13-August/slow-governance line was a placeholder. What actually happened: the slow-governance proposal of 2 August was withdrawn before activation after a security vulnerability was found during continued testing (Mainnet was never exposed), and a patched kernel was resubmitted via fast governance on 19 August, with the upgrade triggered on 20 August. Also: 'Native Atomic Calls' -> 'Native Atomic Composability (NAC)' (terminology used everywhere else in the docs), and a link to the resubmission Agora post. Note: the 20 August date assumes the trigger happens as planned on Thursday morning; re-verify against the actual trigger_kernel_upgrade operation before merging. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

- accounts-and-aliases: 'the the' typo - fees: quantify the EIP-7702 DA-fee term (125 bytes per authorization, matching fees.rs) - terminology: 'Native Atomic Calls' -> 'Native Atomic Composability (NAC)' (bridging.md, network-information.mdx, testnet.mdx) - testnet.mdx: Previewnet is long-running, not a 'temporary testnet' (same fix previewnet.md already received) Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

lthms · 2026-08-19T09:54:09Z

@@ -42,5 +42,5 @@ Examples of using test networks are on other pages in this documentation.

 ## Etherlink{/* TX */} Previewnet


Strictly speaking it's called Tezos X Previewnet and i don't think there is a need to “downcast” it to Etherlink Previewnet? or at least no decision was made about that

saroupille · 2026-08-19T11:58:43Z

Heads-up @lthms: this PR was merged at 09:57 UTC, but a 4th commit (41e4711, Network information Michelson: chain ids, /tezlink prefix, bridge/faucet/indexer) had been pushed to the branch just after — it missed the merge. Since we now have write access, it has been pushed directly to import-tx-doc (fast-forward a29ec13..41e4711). Nothing else pending here.

saroupille · 2026-08-19T09:22:48Z

Addresses the open review thread on #436 asking for a look at the new "Bridging vs. NAC" section. Targets the import-tx-doc branch.

Fixes:

Michelson runtime label: the first diagram said "Tezos JSON-RPC" — the Tezos RPC is not JSON-RPC (the interface table above correctly says "Tezos RPC").
L1 node & anchoring edges: "Tezos Layer1 (Smart Rollup)" read as if L1 were the rollup — Etherlink is the Smart Rollup, L1 hosts it. The runtimes also don't individually post to L1, so the two per-runtime edges are replaced by a single "anchored as a Smart Rollup" edge from the chain.
Sequence diagram note: "One atomic block" → "One atomic transaction" — the atomicity unit is the transaction, not the block.
Comparison table: native tez is not wrapped by the L1↔L2 bridge (only FA tokens are); withdrawal latency is days (without fast withdrawals), not hours.
Cosmetic: stray colons/spacing in mermaid labels, "a unified execution layer", "economic space".

🤖 Generated with Claude Code

- Michelson runtime exposes the Tezos RPC, not JSON-RPC (matches the interface table above) - L1 node: Etherlink is the Smart Rollup, L1 hosts it; anchor the whole chain to L1 with a single edge instead of per-runtime edges - sequence diagram note: the atomicity unit is the transaction, not the block - comparison table: native tez is not wrapped by the L1<->L2 bridge; withdrawal latency is days, not hours - cosmetic: stray colons/spacing in mermaid labels, 'a unified execution layer', 'economic space' Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

vercel · 2026-08-19T09:22:53Z

@saroupille is attempting to deploy a commit to the Trili Tech Team on Vercel.

A member of the Team first needs to authorize it.

The 13-August/slow-governance line was a placeholder. What actually happened: the slow-governance proposal of 2 August was withdrawn before activation after a security vulnerability was found during continued testing (Mainnet was never exposed), and a patched kernel was resubmitted via fast governance on 19 August, with the upgrade triggered on 20 August. Also: 'Native Atomic Calls' -> 'Native Atomic Composability (NAC)' (terminology used everywhere else in the docs), and a link to the resubmission Agora post. Note: the 20 August date assumes the trigger happens as planned on Thursday morning; re-verify against the actual trigger_kernel_upgrade operation before merging. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

- accounts-and-aliases: 'the the' typo - fees: quantify the EIP-7702 DA-fee term (125 bytes per authorization, matching fees.rs) - terminology: 'Native Atomic Calls' -> 'Native Atomic Composability (NAC)' (bridging.md, network-information.mdx, testnet.mdx) - testnet.mdx: Previewnet is long-running, not a 'temporary testnet' (same fix previewnet.md already received) Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

lthms · 2026-08-19T09:54:09Z

@@ -42,5 +42,5 @@ Examples of using test networks are on other pages in this documentation.

 ## Etherlink{/* TX */} Previewnet


Strictly speaking it's called Tezos X Previewnet and i don't think there is a need to “downcast” it to Etherlink Previewnet? or at least no decision was made about that

saroupille · 2026-08-19T11:58:43Z

Heads-up @lthms: this PR was merged at 09:57 UTC, but a 4th commit (41e4711, Network information Michelson: chain ids, /tezlink prefix, bridge/faucet/indexer) had been pushed to the branch just after — it missed the merge. Since we now have write access, it has been pushed directly to import-tx-doc (fast-forward a29ec13..41e4711). Nothing else pending here.

saroupille · 2026-08-19T09:22:48Z

Addresses the open review thread on #436 asking for a look at the new "Bridging vs. NAC" section. Targets the import-tx-doc branch.

Fixes:

Michelson runtime label: the first diagram said "Tezos JSON-RPC" — the Tezos RPC is not JSON-RPC (the interface table above correctly says "Tezos RPC").
L1 node & anchoring edges: "Tezos Layer1 (Smart Rollup)" read as if L1 were the rollup — Etherlink is the Smart Rollup, L1 hosts it. The runtimes also don't individually post to L1, so the two per-runtime edges are replaced by a single "anchored as a Smart Rollup" edge from the chain.
Sequence diagram note: "One atomic block" → "One atomic transaction" — the atomicity unit is the transaction, not the block.
Comparison table: native tez is not wrapped by the L1↔L2 bridge (only FA tokens are); withdrawal latency is days (without fast withdrawals), not hours.
Cosmetic: stray colons/spacing in mermaid labels, "a unified execution layer", "economic space".

🤖 Generated with Claude Code

- Michelson runtime exposes the Tezos RPC, not JSON-RPC (matches the interface table above) - L1 node: Etherlink is the Smart Rollup, L1 hosts it; anchor the whole chain to L1 with a single edge instead of per-runtime edges - sequence diagram note: the atomicity unit is the transaction, not the block - comparison table: native tez is not wrapped by the L1<->L2 bridge; withdrawal latency is days, not hours - cosmetic: stray colons/spacing in mermaid labels, 'a unified execution layer', 'economic space' Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

vercel · 2026-08-19T09:22:53Z

@saroupille is attempting to deploy a commit to the Trili Tech Team on Vercel.

A member of the Team first needs to authorize it.

The 13-August/slow-governance line was a placeholder. What actually happened: the slow-governance proposal of 2 August was withdrawn before activation after a security vulnerability was found during continued testing (Mainnet was never exposed), and a patched kernel was resubmitted via fast governance on 19 August, with the upgrade triggered on 20 August. Also: 'Native Atomic Calls' -> 'Native Atomic Composability (NAC)' (terminology used everywhere else in the docs), and a link to the resubmission Agora post. Note: the 20 August date assumes the trigger happens as planned on Thursday morning; re-verify against the actual trigger_kernel_upgrade operation before merging. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

- accounts-and-aliases: 'the the' typo - fees: quantify the EIP-7702 DA-fee term (125 bytes per authorization, matching fees.rs) - terminology: 'Native Atomic Calls' -> 'Native Atomic Composability (NAC)' (bridging.md, network-information.mdx, testnet.mdx) - testnet.mdx: Previewnet is long-running, not a 'temporary testnet' (same fix previewnet.md already received) Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

lthms · 2026-08-19T09:54:09Z

@@ -42,5 +42,5 @@ Examples of using test networks are on other pages in this documentation.

 ## Etherlink{/* TX */} Previewnet


Strictly speaking it's called Tezos X Previewnet and i don't think there is a need to “downcast” it to Etherlink Previewnet? or at least no decision was made about that

saroupille · 2026-08-19T11:58:43Z

Heads-up @lthms: this PR was merged at 09:57 UTC, but a 4th commit (41e4711, Network information Michelson: chain ids, /tezlink prefix, bridge/faucet/indexer) had been pushed to the branch just after — it missed the merge. Since we now have write access, it has been pushed directly to import-tx-doc (fast-forward a29ec13..41e4711). Nothing else pending here.

saroupille · 2026-08-19T09:22:48Z

Addresses the open review thread on #436 asking for a look at the new "Bridging vs. NAC" section. Targets the import-tx-doc branch.

Fixes:

Michelson runtime label: the first diagram said "Tezos JSON-RPC" — the Tezos RPC is not JSON-RPC (the interface table above correctly says "Tezos RPC").
L1 node & anchoring edges: "Tezos Layer1 (Smart Rollup)" read as if L1 were the rollup — Etherlink is the Smart Rollup, L1 hosts it. The runtimes also don't individually post to L1, so the two per-runtime edges are replaced by a single "anchored as a Smart Rollup" edge from the chain.
Sequence diagram note: "One atomic block" → "One atomic transaction" — the atomicity unit is the transaction, not the block.
Comparison table: native tez is not wrapped by the L1↔L2 bridge (only FA tokens are); withdrawal latency is days (without fast withdrawals), not hours.
Cosmetic: stray colons/spacing in mermaid labels, "a unified execution layer", "economic space".

🤖 Generated with Claude Code

- Michelson runtime exposes the Tezos RPC, not JSON-RPC (matches the interface table above) - L1 node: Etherlink is the Smart Rollup, L1 hosts it; anchor the whole chain to L1 with a single edge instead of per-runtime edges - sequence diagram note: the atomicity unit is the transaction, not the block - comparison table: native tez is not wrapped by the L1<->L2 bridge; withdrawal latency is days, not hours - cosmetic: stray colons/spacing in mermaid labels, 'a unified execution layer', 'economic space' Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

vercel · 2026-08-19T09:22:53Z

@saroupille is attempting to deploy a commit to the Trili Tech Team on Vercel.

A member of the Team first needs to authorize it.

The 13-August/slow-governance line was a placeholder. What actually happened: the slow-governance proposal of 2 August was withdrawn before activation after a security vulnerability was found during continued testing (Mainnet was never exposed), and a patched kernel was resubmitted via fast governance on 19 August, with the upgrade triggered on 20 August. Also: 'Native Atomic Calls' -> 'Native Atomic Composability (NAC)' (terminology used everywhere else in the docs), and a link to the resubmission Agora post. Note: the 20 August date assumes the trigger happens as planned on Thursday morning; re-verify against the actual trigger_kernel_upgrade operation before merging. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

- accounts-and-aliases: 'the the' typo - fees: quantify the EIP-7702 DA-fee term (125 bytes per authorization, matching fees.rs) - terminology: 'Native Atomic Calls' -> 'Native Atomic Composability (NAC)' (bridging.md, network-information.mdx, testnet.mdx) - testnet.mdx: Previewnet is long-running, not a 'temporary testnet' (same fix previewnet.md already received) Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

lthms · 2026-08-19T09:54:09Z

@@ -42,5 +42,5 @@ Examples of using test networks are on other pages in this documentation.

 ## Etherlink{/* TX */} Previewnet


Strictly speaking it's called Tezos X Previewnet and i don't think there is a need to “downcast” it to Etherlink Previewnet? or at least no decision was made about that

saroupille · 2026-08-19T11:58:43Z

Heads-up @lthms: this PR was merged at 09:57 UTC, but a 4th commit (41e4711, Network information Michelson: chain ids, /tezlink prefix, bridge/faucet/indexer) had been pushed to the branch just after — it missed the merge. Since we now have write access, it has been pushed directly to import-tx-doc (fast-forward a29ec13..41e4711). Nothing else pending here.

saroupille · 2026-08-19T09:22:48Z

Addresses the open review thread on #436 asking for a look at the new "Bridging vs. NAC" section. Targets the import-tx-doc branch.

Fixes:

Michelson runtime label: the first diagram said "Tezos JSON-RPC" — the Tezos RPC is not JSON-RPC (the interface table above correctly says "Tezos RPC").
L1 node & anchoring edges: "Tezos Layer1 (Smart Rollup)" read as if L1 were the rollup — Etherlink is the Smart Rollup, L1 hosts it. The runtimes also don't individually post to L1, so the two per-runtime edges are replaced by a single "anchored as a Smart Rollup" edge from the chain.
Sequence diagram note: "One atomic block" → "One atomic transaction" — the atomicity unit is the transaction, not the block.
Comparison table: native tez is not wrapped by the L1↔L2 bridge (only FA tokens are); withdrawal latency is days (without fast withdrawals), not hours.
Cosmetic: stray colons/spacing in mermaid labels, "a unified execution layer", "economic space".

🤖 Generated with Claude Code

- Michelson runtime exposes the Tezos RPC, not JSON-RPC (matches the interface table above) - L1 node: Etherlink is the Smart Rollup, L1 hosts it; anchor the whole chain to L1 with a single edge instead of per-runtime edges - sequence diagram note: the atomicity unit is the transaction, not the block - comparison table: native tez is not wrapped by the L1<->L2 bridge; withdrawal latency is days, not hours - cosmetic: stray colons/spacing in mermaid labels, 'a unified execution layer', 'economic space' Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

vercel · 2026-08-19T09:22:53Z

@saroupille is attempting to deploy a commit to the Trili Tech Team on Vercel.

A member of the Team first needs to authorize it.

The 13-August/slow-governance line was a placeholder. What actually happened: the slow-governance proposal of 2 August was withdrawn before activation after a security vulnerability was found during continued testing (Mainnet was never exposed), and a patched kernel was resubmitted via fast governance on 19 August, with the upgrade triggered on 20 August. Also: 'Native Atomic Calls' -> 'Native Atomic Composability (NAC)' (terminology used everywhere else in the docs), and a link to the resubmission Agora post. Note: the 20 August date assumes the trigger happens as planned on Thursday morning; re-verify against the actual trigger_kernel_upgrade operation before merging. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

- accounts-and-aliases: 'the the' typo - fees: quantify the EIP-7702 DA-fee term (125 bytes per authorization, matching fees.rs) - terminology: 'Native Atomic Calls' -> 'Native Atomic Composability (NAC)' (bridging.md, network-information.mdx, testnet.mdx) - testnet.mdx: Previewnet is long-running, not a 'temporary testnet' (same fix previewnet.md already received) Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

lthms · 2026-08-19T09:54:09Z

@@ -42,5 +42,5 @@ Examples of using test networks are on other pages in this documentation.

 ## Etherlink{/* TX */} Previewnet


Strictly speaking it's called Tezos X Previewnet and i don't think there is a need to “downcast” it to Etherlink Previewnet? or at least no decision was made about that

saroupille · 2026-08-19T11:58:43Z

Heads-up @lthms: this PR was merged at 09:57 UTC, but a 4th commit (41e4711, Network information Michelson: chain ids, /tezlink prefix, bridge/faucet/indexer) had been pushed to the branch just after — it missed the merge. Since we now have write access, it has been pushed directly to import-tx-doc (fast-forward a29ec13..41e4711). Nothing else pending here.