fix config load error when picker is set before the scoerer w/o weight.#1958
Merged
k8s-ci-robot merged 1 commit intokubernetes-sigs:mainfrom Dec 6, 2025
Merged
Conversation
✅ Deploy Preview for gateway-api-inference-extension ready!
To edit notification comments on pull requests, go to your Netlify project configuration. |
k8s-ci-robot
added
size/S
Contributor
|
You may be interested in #1953. I tried to harden this path broadly. Your change will likely be merged first as it is well-scoped. I'll make sure to verify that this is also fixed in my refactoring. |
Contributor
|
/lgtm |
k8s-ci-robot
added
the
lgtm
Contributor
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: ahg-g, zetxqx The full list of commands accepted by this bot can be found here. The pull request process is described here DetailsNeeds approval from an approver in each of these files:
Approvers can indicate their approval by writing |
k8s-ci-robot
added
the
approved
LukeAVanDrie
added a commit
to LukeAVanDrie/gateway-api-inference-extension
that referenced
this pull request
Dec 8, 2025
k8s-ci-robot
pushed a commit
that referenced
this pull request
Dec 8, 2025
) * config: unify loading phases and system defaults Refactors the configuration loader to clarify the two-stage loading process and standardize default injection. This change: - Merges the loading lifecycle into a linear pipeline: Load Raw -> Instantiate -> Apply System Defaults -> Validate -> Build. - Introduces a `registerDefaultPlugin` helper to standardize how mandatory components (like MaxScorePicker and SingleProfileHandler) are injected when omitted by the user. - Improves error wrapping context throughout the loading process. * test: overhaul config loader test suite Rewrites `configloader_test.go` to use table-driven tests and improves overall test hygiene. Key improvements: - Separates tests for "Raw Loading" (Phase 1) and "Instantiation" (Phase 2) to mirror the new production architecture. - Adds comprehensive test data constants in `testdata.go` to prevent drift between YAML and Go struct definitions. - Adds deep validation callbacks to verify post-instantiation state, ensuring defaults are injected correctly. * runner: update entrypoint to use new config loader Updates the EPP runner to use the refactored loader functions. The flow is updated to: 1. `loader.LoadRawConfig` to parse bytes and retrieve feature gates. 2. Initialize the plugin handle. 3. `loader.InstantiateAndConfigure` to build the final configuration. * Address reviewer feedback. * Ensure no regressions from #1958.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What type of PR is this?
/kind bug
What this PR does / why we need it:
W/o this PR, the following EPP scheduler config will get nil pointer
panic: runtime error: invalid memory address or nil pointer dereference [signal SIGSEGV: segmentation violation code=0x1 addr=0x0 pc=0x1d12fc9]We should not assume picker is always the last plugin. Thus, we should not break earlier before setting the defaults for all scorers.
Which issue(s) this PR fixes:
Fixes #
Does this PR introduce a user-facing change?: