Visitar URL original
Tripwires fired: #9014, #9017, #9018, #9019, #9027 are fixed upstream by tae898 · Pull Request #143 · humemai/arcadedb-embedded-python · GitHub
Skip to content

Tripwires fired: #9014, #9017, #9018, #9019, #9027 are fixed upstream - #143

Merged
tae898 merged 1 commit into
mainfrom
tripwires-fired-9014-9017-9018-9019-9027
Oct 4, 2026
Merged

tae898 merged 1 commit into
mainfrom
tripwires-fired-9014-9017-9018-9019-9027

Conversation

@tae898

@tae898 tae898 commented Oct 4, 2026

Copy link
Copy Markdown

Fixes #142.

CI tests against the 26.10.1-SNAPSHOT image, which now carries upstream's fixes, so the strict xfail tripwires for ArcadeData#9014, ArcadeData#9027, ArcadeData#9017, ArcadeData#9018, and ArcadeData#9019 XPASS and fail every run (8 failures on every platform on PR #141). The tripwires become tests of the fixed behavior; the ArcadeData#9017 tests assert that CREATE PROPERTY over records lacking the property is refused and keep only the fill-first workaround (the relax-afterwards path is now refused at the declaration); the known-issues entries and the GraphBatch API note say Fixed in 26.10.1 with the PR that fixed each. ArcadeData#9021 stays a tripwire (upstream PR ArcadeData#9126 is open).

Checked locally: pre-commit (black reformatted, then clean), bandit as CI runs it (rc 0), the tests compile. The behavior itself is tested by this PR's CI on the snapshot engine, as the user asked (heavy runs on GitHub, not the laptop).

🤖 Generated with Claude Code

https://claude.ai/code/session_01DeKEhHXJTy1H1GgbfBZ8mL

…eData#9018, ArcadeData#9019, ArcadeData#9027; known-issues entries say Fixed in 26.10.1 (#142)

The 26.10.1-SNAPSHOT engine CI tests against carries upstream's fixes
(PRs ArcadeData#9121, ArcadeData#9116, ArcadeData#9107, ArcadeData#9108), so the strict xfail tests now assert the
fixed behavior: a value the declared type cannot hold is refused, CREATE
PROPERTY over records that violate it is refused, and a GraphBatch edge keeps
a null as null and does not wrap an out-of-range SHORT. The ArcadeData#9017 workaround
test keeps only the fill-first path, the other one is now refused. ArcadeData#9021
stays a tripwire.

Fixes #142

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DeKEhHXJTy1H1GgbfBZ8mL
@tae898
tae898 merged commit fea27c5 into main Oct 4, 2026
48 checks passed
@tae898
tae898 deleted the tripwires-fired-9014-9017-9018-9019-9027 branch October 4, 2026 07:41
tae898 added a commit that referenced this pull request Oct 4, 2026
… entry says Fixed in 26.10.1 (PR ArcadeData#9126) (#166)

Upstream PR ArcadeData#9126 (merge 0ef1877) makes an index on an integral key
answer for the exact bound. The strict xfail for ArcadeData#9021 now asserts the
fixed rows (i = 12.5 matches nothing, i >= 12.5 returns 13, i < 12.5
returns 11 and 12) on the indexed and on the unindexed type, and checks
that the plan of the indexed query reads the index, so the test cannot
pass by comparing a scan with a scan.

Since #143 the ArcadeData#9021 entry in known-issues.md carried the Fixed paragraph
of ArcadeData#9014 and ArcadeData#9027 (PR ArcadeData#9121). That paragraph moves back to its own
entry, and ArcadeData#9021 gets a Fixed paragraph for PR ArcadeData#9126. The rounding
workaround is kept for engines before 26.10.1.

Checked with the 2026-10-03 wheel: on its engine (b22b5e9) the three
new cases fail with the answers the issue reports; with the wheel's
engine jars replaced by builds of upstream main 2e09816 and
422a5f6, all 11 tests in the file pass.


Claude-Session: https://claude.ai/code/session_01Ubz3m6QXHESerCkehC1qgi

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Tripwires fired: upstream fixed #9014, #9017, #9018, #9019, #9027 (known-issues entries and tests)

1 participant