A major cleanup week in rippled with 24 PRs merged to develop — highlighted by ~3,000 lines of dead code removal, a fixCleanup3_6_0 amendment placeholder, several bug fixes (UB in JSON/STData, config validation, RPC load charges, null-SLE guards), and a clean-exit-on-bad-config improvement for node operators. In xrpl.js, 5 PRs merged to main adding hex field validation across 46 transaction fields, fixing dropsToXrp precision loss (now returns BigNumber), and hardening non-UTF-8 and error handling. In xrpl4j, a binary parser decode bug was fixed (closes #805) and TrustLine flag test coverage was expanded. On the developer portal, 6 PRs merged including Python/Go Batch tutorial examples and a batch of error code corrections. In XRPL-Standards, XLS-0096 received a FAQ clarification on confidential balances and TransferFee.
This was a spring-cleaning week for the XRP Ledger's core server software (rippled), with 24 changes merged — the vast majority focused on removing unused code and fixing small but important bugs. Developers systematically deleted roughly 3,000 lines of code that nothing was using anymore: old debugging infrastructure, obsolete data structures, an entire external library that was no longer needed, and test files that had already been migrated to a newer testing framework. While deleting code may sound unexciting, it makes the codebase easier to understand, faster to compile, and reduces the surface area for future bugs. On the bug-fix side, several changes improve the experience for people running XRPL nodes: if you have a typo or duplicate entry in your server configuration file, the server will now print a clear error message and exit cleanly instead of crashing with a cryptic error. Two internal JSON-handling bugs involving "undefined behavior" (a category of subtle programming error that can cause unpredictable results) were also corrected.
For application developers, the JavaScript SDK (xrpl.js) received important improvements that will ship in the next npm release. The most significant is that 46 transaction fields — things like token IDs, vault IDs, and domain names — are now validated to ensure they contain valid hexadecimal strings before being sent to the network. Previously, any string was accepted, which could lead to confusing errors. The dropsToXrp() function, which converts the smallest XRP unit to full XRP, now returns a high-precision number type instead of a regular JavaScript number — this prevents rounding errors when dealing with very large amounts. The Java SDK (xrpl4j) fixed a bug in its binary decoder that was using the wrong formula to calculate field lengths, and exposed several follow-up issues in the same area that the maintainer is now tracking. None of these SDK changes are available to install yet — they're merged into the main development branch and will ship in the next tagged release.
On the documentation side, the developer portal corrected several error code names and numbers that didn't match what the actual server software returns — important for developers writing error-handling logic. Python and Go code examples were also added to the Batch transaction tutorials, so developers working in those languages can now follow along with working code.
For the latest updates, follow @XRPLF and @RippleXDev on X, and explore the XRPLF repos directly for full development activity.
Note: All rippled changes below were merged to the develop branch and are not yet live on the network. A tagged release is required for any change to reach production.
Note: SDK PRs below were merged to each library's main branch. They become available to application developers only after a tagged release is published (npm for xrpl.js, Maven Central for xrpl4j, PyPI for xrpl-py). If no corresponding SDK release appears in the data, treat the feature as merged-but-pending-release and phrase availability accordingly.
fixCleanup3_6_0 amendment placeholder: rippled#8681 (+1/−0) — Adds a catchall amendment intended to gate small bug fixes in the 3.6 release cycle. If no fixes need it, the placeholder can be removed. Validators: no action needed yet — this amendment is not voting-eligible.STData and skip WasmVM death tests under UBSan: rippled#8690 (+18/−3) — STData(SerialIter&) loaded wire values into SerializedTypeID before switching, causing undefined behavior with out-of-range values. Now switches on the raw uint16.Json::Value::asBool() reading wrong union member: rippled#8638 (+50/−2) — The UInt branch read the signed intVal union member instead of uintVal, which is undefined behavior. Closes #7508.getSingleSection now throws at startup when a single-value config section (e.g., [network_id]) appears more than once, with a clear error message. Operators: if your config has duplicate sections, the server will now reject it at startup rather than silently using one.NoRippleCheck and GatewayBalances: rippled#7827 (+18/−0) — Adds defensive null checks to directory-iteration callbacks in doNoRippleCheck and doGatewayBalances, matching the pattern used by all other account handlers. Closes #7510.noripple_check load type: rippled#7729 (+3/−0) — The noripple_check RPC handler now charges feeMediumBurdenRPC instead of the default reference fee, matching its actual workload. Closes #7506.simulate RPC load charge: rippled#6841 (+1/−1) — simulate now charges feeHeavyBurdenRPC to match sign/sign_for, which perform equivalent work. Closes #6801.Quality::ceilOut: rippled#8037 (+1/−1) — Corrects ceil_TAmounts_helper to ceilTAmountsHelper. The template is currently never instantiated, so this is a latent bug fix. Closes #8036.TaggedPointer::getChildIndex: rippled#8661 (+8/−1) — Adds an XRPL_ASSERT to guard both arms of the branch test, matching the existing assertion in SHAMapInnerNode::getChildHash. Release builds are unchanged.A wave of dead-code removal from @mvadari stripped approximately 3,000 lines:
FlowDebugInfo: rippled#8676 (+6/−371) — Every flow() caller passed nullptr, so the debug code never ran.Dir and BookDirs classes: rippled#8674 (+89/−642) — Only tests used these; test usage replaced with a small ownerDirEntries helper.json::Writer and outputJson: rippled#8673 (+9/−987) — Nothing outside tests called this code.AmendmentTableImpl.h and messages.h.extractTarLz4 and libarchive: rippled#8675 (+0/−133) — Eliminates an entire dependency.AccountSet_test: rippled#8678 (+0/−629) — All 14 cases already ported to the gtest suite.ApplyContext: rippled#8637 (+189/−52) — Replaced with a std::reference_wrapper that eliminates the need for suppression annotations.LedgerMaster::gotFetchPack(): rippled#8655 (+3/−11)archive.debian.org in CI: rippled#8693 (+4/−9) — Switches from snapshot.debian.org to a more reliable archive mirror.node: lts: rippled#8672 (+12/−10) — Refreshes hook versions for the pre-commit framework.CLAUDE.md symlinks: rippled#8249 (+3/−6) — Claude Code now reads AGENTS.md directly; the symlinks are obsolete.AGENTS.md: rippled#8198 (+93/−21) — Mines 6 months of merged-PR review comments into directory-scoped AGENTS.md files for AI coding agents.strUnHex test cases: rippled#7626 (+6/−0) — Ensures strUnHex() rejects non-hex input containing non-ASCII byte values.node: lts and refreshed hooks. A nightly development build (nightly-20261106) was also published — this is not recommended for production.not_enabled status now that PermissionDelegationV1_1 is enabled on Mainnet.telNO_DST_PARTIAL formatting (#4050), corrected temBAD_AMM_TOKENS in AMMCreate (#4049), fixed tecUNFUNDED_AMM name (#4047), and corrected six NFT-related tec code numeric values that were off by one (#4046).TransferFee. Non-normative clarification.12-16 to 12-15 in XLS-37 quickstart.pull_request_target workflows, now using the REST API instead.isHexString checks for fields like NFTokenID, VaultID, MPTokenIssuanceID, InvoiceID, Domain, and more. Previously these accepted any string. Merged — pending npm release.dropsToXrp() returns BigNumber for precision: xrpl.js#3341 (+159/−89) — Returns BigNumber instead of JavaScript number to prevent precision loss for large amounts. Client.getXrpBalance() also returns Promise. Closes #3316. Merged — pending npm release. Note: this is a breaking change for code that expects a number return type.jsonInvalid) are received in responses. Merged — pending npm release.validateRequiredField/validateOptionalField helpers. Merged — pending npm release.BinaryParser.readVariableLength used the wrong constant (MAX_SECOND_BYTE_VALUE = 240) instead of the correct MAX_SINGLE_BYTE_LENGTH (192), causing incorrect decoding of variable-length fields. Closes #805. Labeled bug. Merged — pending Maven Central release.AGENTS.md for AI coding agents: xrpl4j#865 (+11/−0) — Adds git workflow guidance so AI agents follow the same conventions as human contributors.rippled — Performance optimizations: Two PRs from @mvadari target transaction processing overhead:
TxQ::tryDirectApply to avoid running preflight twice for every direct-apply transaction — rippled#8683 (+7/−6), approved.ValidVault balance deltas only for vault transactions instead of every transaction — rippled#8682 (+59/−14), approved by 2 reviewers.rippled — Lending Protocol V1.2: Three PRs labeled feature: LP1.2:
LoanSequence check on LoanSet — rippled#8685 (+224/−0), approved.LoanSet data on the Loan ledger entry — rippled#8684 (+141/−0), approved with one non-blocking comment.rippled — Config error messages: rippled#8691 (+166/−56) — Stacked on the merged #8473, this makes the error message report the section name and rejected text instead of a bare std::bad_cast. Approved by 4 reviewers but has merge conflicts.
rippled — Remove date library dependency: rippled#8668 (+377/−154, 21 files) — Replaces Howard Hinnant's date library with C++20 facilities. Has outstanding review comments about potential overflow in clock conversions.
rippled — Validator keys in-tree: rippled#8223 (+6370/−317, 38 files) — Moves validator-keys from a separate repository into the rippled tree with external (HSM) signing and validator-list signing support. Very large PR with active review.
rippled — Batch DeliverMax fix: rippled#8558 (+430/−25) — Ensures the DeliverMax rule applies to inner Batch Payments and to simulate/submit. Approved by 5 reviewers; has merge conflicts.
xrpl4j — Unrecognized PermissionValue handling: xrpl4j#863 (+745/−76) — Deserializes unknown PermissionValue entries as UnknownPermission instead of throwing, which previously made entire transactions unparseable. Labeled breaking. Approved by 3 reviewers.
xrpl4j — Inverted NFTokenOffer buy/sell flags: xrpl4j#868 (+55/−18) — The lsfSellNFToken bit was bound to a constant named BUY_TOKEN, causing every NFTokenOffer read through xrpl4j to report the trade direction backwards. Approved.
xrpl.js — Review policy update: xrpl.js#3573 (+205/−2) — Lowers the review requirement on main from two human approvals to one, with a needs-second-reviewer label safeguard. Approved.
xrpl-dev-portal — Amendments page: xrpl-dev-portal#4044 (+3190/−3, 15 files) — Proposes an xrpl.org/amendments dashboard with per-feature cards, side panels with key dates and audit links, and live vote counts.
xrpl-dev-portal — Ruby SDK and examples: xrpl-dev-portal#3837 (+358/−3) — Adds Ruby tutorials and code samples, approved by @xVet from the Foundation side.
opensource.ripple.com — Lending Protocol 1.2 docs: opensource.ripple.com#243 (+750/−0, 12 files) — Draft documentation for LP1.2 features. Marked "Do not merge — features not finalized."
PermissionValue and NFTokenOffer fixes: #863 (approved, labeled breaking) and #868 (approved) fix real deserialization bugs affecting Java applications reading delegation and NFT data. Watch for a v7.0.0 release that includes these along with the VL decode fix from this week.LendingProtocolV1_2.DeliverMax fix approved by 5 reviewers: rippled#8558 ensures Payments inside Batch transactions follow the DeliverMax field rule — needs conflict resolution before it can merge.Community contributors active across repos this week: @nabe3m contributed four documentation accuracy fixes to the developer portal (error code names and numeric values) plus the permission delegation status update. @hmalinski fixed the Json::Value::asBool() UB in rippled and opened SDK fixes for xrpl.js (OracleSet Scale limit) and xrpl-py (UInt64 decode round-trip). @FawwazLawal's xrpl4j VL decode fix (opened last week) merged with companion tests from maintainer @sappenin. @SAY-5 contributed the null-SLE guards in rippled. @BraedonKlock's noripple_check load-type fix and @shortthefomo's simulate RPC charge fix — both long-standing community PRs — also merged this week. @jarrensj contributed the CLAUDE.md symlink cleanup, and @Krypto-Whitehat fixed the ceilOut template typo.
Clio bug reports: @mvadari filed five bugs (#3289–#3293) after systematically diffing Clio 2.8.0 against rippled 3.4.1 — covering error code mismatches, response field differences, HTTP vs HTTPS in urlgravatar, boolean singleton handling in ledger_entry, and stale data on one altnet node.
xrpl4j follow-up issues: @sappenin filed six issues (#862–#873) exposing pre-existing weaknesses in BinaryParser/BinarySerializer VL prefix handling — including asymmetric encode/decode bounds, truncated input handling, and Error vs IllegalArgumentException throw types. @hmalinski has volunteered to take items 1 and 2.
Standards discussions: @donovanhide proposed a Canonical Transaction Ordering amendment replacing XOR in CanonicalTXSet::accountKey with SipHash-2-4 to prevent prefix-grinding exploits (XRPL-Standards#679). @LeDenti raised a question about XLS-85 (Token Escrow) and whether EscrowFinish by a third party should create the destination trustline (XRPL-Standards#677) — @mvadari responded that this would conflict with the design intent of trustlines requiring active user consent.
New rippled issues: A question was raised about whether the VaultDeposit cap should apply when a vault is already over AssetsMaximum (rippled#8700). A Good First Issue was filed for deferring transaction serialization in ApplyStateTable::apply until it's actually written (rippled#8671).
Compared to last week (October 26 – November 1, 2026):
| Metric | This Week | Last Week | Change |
| Repos with activity | 8 | 8 | flat |
| rippled PRs merged | 24 | 2 | ↑22 |
| rippled PRs opened | 22 | 26 | ↓4 |
| rippled new issues | 4 | 3 | ↑1 |
| rippled closed issues | 5 | 1 | ↑4 |
| rippled releases | 0 | 0 | flat |
| rippled commits | 23 | 2 | ↑21 |
| xrpl.js PRs merged | 5 | 3 | ↑2 |
| xrpl.js PRs opened | 17 | 6 | ↑11 |
| xrpl.js new issues | 4 | 2 | ↑2 |
| xrpl.js closed issues | 1 | 1 | flat |
| xrpl.js releases | 1 | 2 | ↓1 |
| xrpl.js commits | 5 | 3 | ↑2 |
| xrpl-py PRs merged | 1 | 0 | ↑1 |
| xrpl-py PRs opened | 5 | 7 | ↓2 |
| xrpl-py new issues | 0 | 1 | ↓1 |
| xrpl-py releases | 0 | 0 | flat |
| xrpl-py commits | 1 | 0 | ↑1 |
| xrpl-dev-portal PRs merged | 6 | 23 | ↓17 |
| xrpl-dev-portal PRs opened | 13 | 9 | ↑4 |
| xrpl-dev-portal new issues | 2 | 2 | flat |
| xrpl-dev-portal commits | 6 | 25 | ↓19 |
| Clio PRs merged | 2 | 2 | flat |
| Clio PRs opened | 3 | 1 | ↑2 |
| Clio new issues | 7 | 0 | ↑7 |
| Clio releases | 1 | 1 | flat |
| Clio commits | 2 | 2 | flat |
| XRPL-Standards PRs merged | 3 | 2 | ↑1 |
| XRPL-Standards PRs opened | 5 | 4 | ↑1 |
| XRPL-Standards commits | 3 | 2 | ↑1 |
| xrpl4j PRs merged | 4 | 0 | ↑4 |
| xrpl4j PRs opened | 8 | 4 | ↑4 |
| xrpl4j new issues | 8 | 0 | ↑8 |
| xrpl4j closed issues | 3 | 0 | ↑3 |
| xrpl4j releases | 0 | 0 | flat |
| xrpl4j commits | 4 | 0 | ↑4 |
| Total PRs merged (all repos) | 45 | 32 | ↑13 |
| Total releases | 3 | 3 | flat |
Carryovers from last week: Several items noted as "In Progress" last week merged this week: the ApplyContext refactor (rippled#8637), the TaggedPointer assertion (rippled#8661), the dropsToXrp BigNumber fix (xrpl.js#3341), the xrpl4j VL decode fix (xrpl4j#858), and the Batch tutorial expansion (xrpl-dev-portal#4038).
Trends: rippled merges surged from 2 to 24 — last week's wave of 26 opened PRs (robustness and cleanup work) began landing in earnest. The dev portal slowed from 23 merges to 6 after last week's dependency security patch blitz. xrpl4j saw a burst of activity with 4 merges, 8 new PRs opened, and 8 new issues filed — mostly follow-ups from the VL decode bug fix exposing pre-existing weaknesses. Clio saw 7 new bug reports filed (5 from systematic rippled-vs-Clio diffing), signaling a quality alignment effort.