Demonstrate that arrays of geo_shape can no longer be ingested
#14196
Conversation
This was first broken in commit ebe6f2c Demonstrates opensearch-project#14193 Signed-off-by: Kurt McKee <[email protected]>
kurtmckee
requested review from
anasalkouz,
andrross,
Bukhtawar,
CEHENKLE,
dblock,
dbwiddis,
dreamer-89,
gbbafna,
kotwanikunal,
mch2,
msfroh,
nknize,
owaiskazi19,
reta,
Rishikesh1159,
sachinpkale,
saratvemulapalli,
shwetathareja,
sohami,
tlfeng and
VachaShah
as code owners
|
❌ Gradle check result for ce13959: FAILURE Please examine the workflow log, locate, and copy-paste the failure(s) below, then iterate to green. Is the failure a flaky test unrelated to your change? |
Signed-off-by: Kurt McKee <[email protected]>
kurtmckee
force-pushed
the
fix-arrays-of-geo_shape-issue-14193
branch
from
acdfeaf to
674dd0d
Compare
|
❌ Gradle check result for acdfeaf: FAILURE Please examine the workflow log, locate, and copy-paste the failure(s) below, then iterate to green. Is the failure a flaky test unrelated to your change? |
|
❌ Gradle check result for 674dd0d: FAILURE Please examine the workflow log, locate, and copy-paste the failure(s) below, then iterate to green. Is the failure a flaky test unrelated to your change? |
|
This PR is stalled because it has been open for 30 days with no activity. |
|
It is incorrect to mark this as stalled; this PR exists to ensure that #14193 can be easily tested. Gradle checks are failing because this PR offers the full day of research I put in to make it trivial to demonstrate the "break" part of a "break/fix" PR. I would like OpenSearch maintainers to add the "fix" commits to this PR to resolve #14193. |
|
@kurtmckee thank you for opening a PR that demonstrates the problem! Do you know how to fix it? If so please don't wait for someone else to fix it, cc: @navneet1v - are you working on a fix or do you know anyone working on one? |
|
Unfortunately I'm not familiar with Java at all, and the commit that introduced the issue (ebe6f2c) was too large for me to get a toe-hold into the code. It took almost a full day to produce the "break" commits in this PR, and this is as far as I can go. 😅 |
|
@dblock I will try to take a look at this. I know where the changes are needed. It might not be a straightforward change as multiple shapes cannot be part of 1 single doc value. There was other ways to do it which was using a GEOMETRY_COLLECTION to achieve this. but give me some time I will try to see the feasibility here. |
|
This PR is stalled because it has been open for 30 days with no activity. |
|
This PR is not stalled; it is awaiting a OpenSearch developer to push the "fix bug" portion of this "break CI / fix bug" PR. The day I spent to hunt this down and craft this handoff PR should preferably not be discarded by closing this PR. |
Yeah the bot doesn't know that the PR is reproducing a problem, but it's ok even if it gets auto-closed. The issue will remain open and someone who is actually fixing the bug will use it. |
|
This PR is stalled because it has been open for 30 days with no activity. |
|
This PR isn't stalled. I don't want it to be auto-closed. @navneet1v Have you been able to review this regression? |
|
@kurtmckee Maybe you want to try and bisect the regression to a specific commit? Probably will become clear how it broke and hopefully easy to fix. |
Thanks @kurtmckee. I'll try to ping some folks offline. |
Description
Note
This PR is not intended for merging as-is. It demonstrates issue #14193, which reports a regression in OpenSearch introduced in #4266 and #8301, and released in v2.9.0 and higher.
It is expected that OpenSearch maintainers can fetch this branch, add fix commits to the branch to resolve the underlying bug, and push directly to this PR. Thank you!
Related Issues
Resolves #14193
Check List
By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.
For more information on following Developer Certificate of Origin and signing off your commits, please check here.