Post-merge follow-up: LocalTerra split-fee redeploy + strengthen #533 P6–P8 e2e #549

New issue

Closed

opened 2026-08-17 10:32:46 +00:00 by PlasticDigits · 7 comments

PlasticDigits commented

2026-08-17 10:32:46 +00:00

(Migrated from gitlab.com)

Why

!1075 (#538) and !1072 (#539) are merged to main. Residual work from the merge sanity check is not blocking those issues, but it is still unfinished.

Related: #538, #539, #533.

Current codebase (after !1072 / !1075)

scripts/deploy-dex-local.sh instantiates wrap-mapper with fee_wrap_bps / fee_unwrap_bps (legacy fee_bps fallback) and asserts CreatePair inherit before per-pair SetDiscountRegistry.
getPairDiscountRegistry prefers pair GetDiscountRegistry with LCD raw fallback.
frontend-dapp/e2e/pool-one-sided-533-tx.spec.ts records P4, P5 (native LUNC + cLUNC pair), a combined P6/P7, and a weak P8.
make verify-issue-538 / make verify-issue-539 exist. Inherit SKIP still exits 0 when LocalTerra is down.

Scope

Ops — LocalTerra wrap-mapper split-fee instantiate. Redeploy (make setup-cloud-localterra, --fresh if volumes are stale) so running stacks are not still on legacy fee_bps. Confirm mapper config has split fields. Run make verify-issue-539 with chain up (not only VERIFY_ISSUE_539_SKIP_E2E=1).
Strengthen #533 e2e-tx P6–P8 to match issue wording:
- P6: withdraw as the input CW20 (explicit option, not asOpts.first()).
- P7: withdraw as uluna; assert leftover cLUNC is not drained (A7).
- P8: empty pool — one-sided disabled; Advanced two-sided first deposit.
Manual #538 QA (unchecked on !1075): Swap/Pool fee chrome strikethrough on a wired LocalTerra pair; 1.13.x raw-fallback if any such pairs remain.
Optional: inherit SKIP should not be a silent green when a live chain is expected; CI quota was exceeded on both MRs (ci_quota_exceeded).

Out of scope

Re-running columbus-5 All/Batch (#538 F538-1 already ops-complete).
Changing I10 fail-closed fee behavior.
Mainnet wrap-mapper migrate (#516 / #523).

Verify

make setup-cloud-localterra   # --fresh if instantiate still uses fee_bps
make verify-issue-539         # chain up; do not skip e2e
make verify-issue-538
sg docker -c 'CI=1 make test-e2e'   # or pool-one-sided-533-tx.spec.ts --project=e2e-tx

P6, P7, and P8 must be distinct passing tests matching the bullets above.

## Why !1075 (#538) and !1072 (#539) are merged to `main`. Residual work from the merge sanity check is not blocking those issues, but it is still unfinished. Related: [#538](https://gitlab.com/PlasticDigits/cl8y-dex-terraclassic/-/issues/538), [#539](https://gitlab.com/PlasticDigits/cl8y-dex-terraclassic/-/issues/539), [#533](https://gitlab.com/PlasticDigits/cl8y-dex-terraclassic/-/issues/533). ## Current codebase (after !1072 / !1075) - `scripts/deploy-dex-local.sh` instantiates wrap-mapper with `fee_wrap_bps` / `fee_unwrap_bps` (legacy `fee_bps` fallback) and asserts CreatePair inherit before per-pair `SetDiscountRegistry`. - `getPairDiscountRegistry` prefers pair `GetDiscountRegistry` with LCD raw fallback. - `frontend-dapp/e2e/pool-one-sided-533-tx.spec.ts` records P4, P5 (native LUNC + cLUNC pair), a combined P6/P7, and a weak P8. - `make verify-issue-538` / `make verify-issue-539` exist. Inherit SKIP still exits 0 when LocalTerra is down. ## Scope 1. **Ops — LocalTerra wrap-mapper split-fee instantiate.** Redeploy (`make setup-cloud-localterra`, `--fresh` if volumes are stale) so running stacks are not still on legacy `fee_bps`. Confirm mapper `config` has split fields. Run `make verify-issue-539` with chain up (not only `VERIFY_ISSUE_539_SKIP_E2E=1`). 2. **Strengthen #533 e2e-tx P6–P8** to match issue wording: - P6: withdraw as the input CW20 (explicit option, not `asOpts.first()`). - P7: withdraw as `uluna`; assert leftover cLUNC is not drained (A7). - P8: empty pool — one-sided disabled; Advanced two-sided first deposit. 3. **Manual #538 QA** (unchecked on !1075): Swap/Pool fee chrome strikethrough on a wired LocalTerra pair; 1.13.x raw-fallback if any such pairs remain. 4. **Optional:** inherit SKIP should not be a silent green when a live chain is expected; CI quota was exceeded on both MRs (`ci_quota_exceeded`). ## Out of scope - Re-running columbus-5 All/Batch (#538 F538-1 already ops-complete). - Changing I10 fail-closed fee behavior. - Mainnet wrap-mapper migrate (#516 / #523). ## Verify ```bash make setup-cloud-localterra # --fresh if instantiate still uses fee_bps make verify-issue-539 # chain up; do not skip e2e make verify-issue-538 sg docker -c 'CI=1 make test-e2e' # or pool-one-sided-533-tx.spec.ts --project=e2e-tx ``` P6, P7, and P8 must be distinct passing tests matching the bullets above.

PlasticDigits commented

2026-08-17 10:32:47 +00:00

(Migrated from gitlab.com)

marked as related to #538

PlasticDigits commented

2026-08-17 10:32:48 +00:00

(Migrated from gitlab.com)

marked as related to #539

PlasticDigits commented

2026-08-17 10:32:49 +00:00

(Migrated from gitlab.com)

marked as related to #533

PlasticDigits commented

2026-08-17 10:33:04 +00:00

(Migrated from gitlab.com)

mentioned in issue #538

PlasticDigits commented

2026-08-17 10:33:05 +00:00

(Migrated from gitlab.com)

mentioned in issue #539

PlasticDigits commented

2026-08-17 10:33:06 +00:00

(Migrated from gitlab.com)

Opened after merging !1075 (#538) and !1072 (#539). CI on both MRs failed with ci_quota_exceeded (no job logs); not a product defect. Makefile conflict between the two MRs was resolved by keeping both verify-issue-538 and verify-issue-539.

Opened after merging !1075 (#538) and !1072 (#539). CI on both MRs failed with `ci_quota_exceeded` (no job logs); not a product defect. Makefile conflict between the two MRs was resolved by keeping both `verify-issue-538` and `verify-issue-539`.

PlasticDigits (Migrated from gitlab.com) closed this issue

2026-08-17 10:34:33 +00:00

PlasticDigits commented

2026-08-17 10:34:34 +00:00

(Migrated from gitlab.com)

close due to low priority