Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

geoindex: produces different result on arm64 #106954

Open
otan opened this issue Jul 17, 2023 · 0 comments
Open

geoindex: produces different result on arm64 #106954

otan opened this issue Jul 17, 2023 · 0 comments
Labels
C-bug Code not up to spec/doc, specs & docs deemed correct. Solution expected to change code/behavior.

Comments

@otan
Copy link
Contributor

otan commented Jul 17, 2023

When running geoindex tests on the M1, you get a different result on GEOS 3.11:

./dev test pkg/geo/geoindex
$ bazel test pkg/geo/geoindex:all --test_env=GOTRACEBACK=all --test_output errors
INFO: Invocation ID: 28a1dbda-ba3c-469a-9108-1a33b359c55d
INFO: Analyzed 4 targets (2 packages loaded, 451 targets configured).
INFO: Found 3 targets and 1 test target...
FAIL: //pkg/geo/geoindex:geoindex_test (see /private/var/tmp/_bazel_otan/d84c8263fef9b9141e9bd680cf5147cb/execroot/com_github_cockroachdb_cockroach/bazel-out/darwin_arm64-fastbuild/testlogs/pkg/geo/geoindex/geoindex_test/test.log)
INFO: From Testing //pkg/geo/geoindex:geoindex_test:
==================== Test output for //pkg/geo/geoindex:geoindex_test:
--- FAIL: TestS2GeometryIndexBasic (0.52s)
    datadriven.go:143:
        /private/var/tmp/_bazel_otan/d84c8263fef9b9141e9bd680cf5147cb/sandbox/darwin-sandbox/23/execroot/com_github_cockroachdb_cockroach/bazel-out/darwin_arm64-fastbuild/bin/pkg/geo/geoindex/geoindex_test_/geoindex_test.runfiles/com_github_cockroachdb_cockroach/pkg/geo/geoindex/testdata/s2_geometry:276:

        expected:
        [F0/L30/000000000000000000000000000000, F0/L30/033333333333333333333333333333],
        F0/L0/, F0/L1/2, F0/L2/22, [F0/L30/222000000000000000000000000000, F0/L30/222333333333333333333333333333],
        spilled

        found:
        [F0/L30/000000000000000000000000000000, F0/L30/333333333333333333333333333333],
        spilled
FAIL
2023/07/16 13:24:41 Test //pkg/geo/geoindex:geoindex_test exited with error code 1
================================================================================

Visualisations

From here

Expected

[F0/L30/000000000000000000000000000000, F0/L30/033333333333333333333333333333],
F0/L0/, F0/L1/2, F0/L2/22, [F0/L30/222000000000000000000000000000, F0/L30/222333333333333333333333333333],
image

Actual

[F0/L30/000000000000000000000000000000, F0/L30/333333333333333333333333333333],
image

Jira issue: CRDB-29775

@otan otan added the C-bug Code not up to spec/doc, specs & docs deemed correct. Solution expected to change code/behavior. label Jul 17, 2023
@otan otan changed the title geoindex: produces different result on M1 macbook geoindex: produces different result on arm64 Jul 18, 2023
craig bot pushed a commit that referenced this issue Jul 18, 2023
103462: cloud: prevent ResumingReader from opening if pos equals size r=dt a=rhu713

Currently, ResumingReader has an issue if the final Read call of its Reader is at the end of the file and returns some other retriable error instead of the expected io.EOF. In this case, on retry, ResumableReader will attempt to open the file at a position equal to the size of the file, which will always result in an invalid range error, causing the read to always fail after exhausting retries. This patch addresses the issue by preventing ResumingReader from calling open on a position equal to the total size of the file.

Fixes #97791

Epic: none

Release note: None

106966: geoindex: skip certain tests for ARM64 r=sumeerbhola a=otan

These tests fail on the ARM64 build as they produce a slightly different result. As such, skip them for now (we have evidence of GEOS doing this for other things like VoronoiPolygons).

Refs: #106954
Resolves #106964

Release note: None

Co-authored-by: Rui Hu <[email protected]>
Co-authored-by: Oliver Tan <[email protected]>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
C-bug Code not up to spec/doc, specs & docs deemed correct. Solution expected to change code/behavior.
Projects
None yet
Development

No branches or pull requests

1 participant