feat!: enable parsing of array flag configurations for flagd by aepfli · Pull Request #1797 · open-feature/flagd

aepfli · 2025-09-08T14:45:29Z

This pull request introduces a significant internal refactoring of the flag storage mechanism, transitioning from a map-based structure to a slice-based one. This change aims to enhance the flexibility and robustness of flag management, particularly in scenarios involving multiple flag sources and flag sets. The update also simplifies the evaluator's state management interface, contributing to a cleaner and more focused API for flag updates.

Highlights

Store Interface Refactoring: The core store interface (IStore) and its implementations have been refactored to manage flags as a slice of model.Flag instead of a map. This change impacts how flags are stored, retrieved, and updated internally, allowing for more flexible handling of flags, especially those associated with different flag sets.
IEvaluator Interface Simplification: The IEvaluator interface has been simplified by removing the GetState() method and modifying the SetState() method signature. The SetState() method no longer returns detailed event information or a resync boolean, streamlining its contract to simply indicate success or failure via an error.
Flag Key Handling in Store: To accommodate flags from different sources and flag sets, the internal storage now uses a compound key (FlagSetId|Key) to uniquely identify flags. This ensures correct handling of flags that might have the same Key but originate from different flag sets or sources.
Test Suite Updates: Extensive updates have been made across the test suite, including fractional_test.go, json_test.go, semver_test.go, string_comparison_test.go, and store_test.go, to align with the new []model.Flag structure and the simplified IEvaluator.SetState() signature. This includes changes to test data structures and assertion logic.

netlify · 2025-09-08T14:45:34Z

✅ Deploy Preview for polite-licorice-3db33c canceled.

Name	Link
🔨 Latest commit	`b646be0`
🔍 Latest deploy log	https://app.netlify.com/projects/polite-licorice-3db33c/deploys/68e0261f9e693c00083b745a

gemini-code-assist

Summary of Changes

Hello @aepfli, I'm Gemini Code Assist¹! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!

This pull request introduces a significant internal refactoring of the flag storage mechanism, transitioning from a map-based structure to a slice-based one. This change aims to enhance the flexibility and robustness of flag management, particularly in scenarios involving multiple flag sources and flag sets. The update also simplifies the evaluator's state management interface, contributing to a cleaner and more focused API for flag updates.

Highlights

Store Interface Refactoring: The core store interface (IStore) and its implementations have been refactored to manage flags as a slice of model.Flag instead of a map. This change impacts how flags are stored, retrieved, and updated internally, allowing for more flexible handling of flags, especially those associated with different flag sets.
IEvaluator Interface Simplification: The IEvaluator interface has been simplified by removing the GetState() method and modifying the SetState() method signature. The SetState() method no longer returns detailed event information or a resync boolean, streamlining its contract to simply indicate success or failure via an error.
Flag Key Handling in Store: To accommodate flags from different sources and flag sets, the internal storage now uses a compound key (FlagSetId|Key) to uniquely identify flags. This ensures correct handling of flags that might have the same Key but originate from different flag sets or sources.
Test Suite Updates: Extensive updates have been made across the test suite, including fractional_test.go, json_test.go, semver_test.go, string_comparison_test.go, and store_test.go, to align with the new []model.Flag structure and the simplified IEvaluator.SetState() signature. This includes changes to test data structures and assertion logic.

Using Gemini Code Assist

The full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips.

Invoking Gemini

You can request assistance from Gemini at any point in your pull request via creating an issue comment (i.e. comment on the pull request page) using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands.

Feature	Command	Description
Code Review	`/gemini review`	Performs a code review for the current pull request in its current state.
Pull Request Summary	`/gemini summary`	Provides a summary of the current pull request in its current state.
Comment	@gemini-code-assist	Responds in comments when explicitly tagged, both in issue comments and review comments.
Help	`/gemini help`	Displays a list of available commands.

Customization

To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a .gemini/ folder in the base of the repository. Detailed instructions can be found here.

Limitations & Feedback

Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for Github and other Google products, sign up here.

You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension.

Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩

gemini-code-assist

Code Review

This pull request refactors the flag storage from a map to a slice, which is a positive change for performance and deterministic ordering. The associated test improvements, especially in TestState_Evaluator, are also commendable.

My review focuses on a few areas for improvement:

Restoring a lost observability metric.
Performance optimizations in the store implementation.
A potential bug in the flag update logic.

Overall, this is a solid refactoring with significant improvements. Addressing the points in my review will make it even better.

aepfli · 2025-09-08T15:33:55Z

/gemini review

gemini-code-assist

Code Review

This pull request refactors the internal flag storage from a map[string]model.Flag to a []model.Flag. This is a significant change that touches many parts of the codebase, especially the store, evaluator, and tests.

The changes are well-implemented for the most part. The store logic has been updated to handle the new data structure, including de-duplication and sorting, which improves determinism. The tests have been diligently updated to reflect the new structures.

I've found one potential high-severity bug in the eventing logic where converting the flag slice to a map could lead to data loss due to duplicate keys. I've also pointed out a minor issue with an outdated comment in a test file that could cause confusion.

Overall, this is a good refactoring. Please address the identified issues.

toddbaert

LGTM:

tests make sense
indexes make sense
I did some manual testing and selection/notification seem to work as expected

tangenti

Could you also update benchmark.txt to reflect the impact of this change?

tangenti · 2025-09-10T13:52:36Z

 							Indexes: []memdb.Indexer{
 								&memdb.StringFieldIndex{Field: model.FlagSetId, Lowercase: false},
 								&memdb.StringFieldIndex{Field: model.Key, Lowercase: false},
+								&memdb.StringFieldIndex{Field: model.Source, Lowercase: false},


Why we need to index the source?

so here comes my assumption, lets say we do have two sources, with the same flagset "awesomeFlagSet" and the same flag, call them sourceA and sourceB. if i select based on flagset i want to get the flag from sourceB. but if i select based on source and request sourceA, i want to get the flag from sourceA - but without adding this, to the primary key, we ensure that both versions stay in the database

I think this is a valid-enough use case that we can justify this.

We can of course NOT implement this if we want to be strict about the composite key being SOLELY the flagSet/Key.

I think before release, we should document this more thoroughly though, along with deciding on this and documenting that as well. We need documentation around how our new selector works in general. I'm happy to write these docs once we are all in agreement and have a release candidate.

It sounds a new feature instead of a refactor? Before having the selectors for flag set IDs and memdb change, flags with the same key from different source are overriding based on the priority IIUC.

i am not sure if this is a new feature or a undefined side-effect of the maps configuration. the maps configuration and the selector allow for different possible configuration and retrieval of flags. We never specified what happens if a flag is defined in multiple sources for the same flagset, and what is the impact for the user. -> my interpretation, we are using a database and a selector, i allow selection per flagset and per source, if i select by source, it is unexpected for me, that the flag is missing, although it is defined in the source. Maybe we should clarify what we expect in this case. Do we need to improve the ADR or create a new one? Is it worth it?

Sorry, I was sick last week. Feel free to ping me on Slack if I won't respond in time.

Regarding the cases you mentioned, yes it's a problem of today and there are more cases we should handle more gracefully, such as the granularity of the override, whether a source could revert a version etc. It's an important improvement that I'd like to have, however I don't think this PR is a good place to solve that problem.

I'd appreciate if you or someone could start an ADR to discuss the problems we aim to solve more thoroughly, and check if there's API changes required for the changes. Once people agree on the scope and the solution, we could then go back to the implementation.

fyi: @open-feature/flagd-approvers - any other input?

I think this is not something we should add now, and at the moment I'm not convinced it should be added at all.

My reasoning:

though this helps in the hypothetical situation that a flagdSetId+key are duplicated in multiple sources, it doesn't help with the case that it's duplicated in the same source (so we aren't really fully removing the edge case)

it's much easier to explain to someone "the PK for flags in flagd is a compound of flagSetId and flagKey" than to also explain sources; without sources it's a simpler model

the problem this is solving is not a use-case I think will really be encountered in often in real usage; simply don't create source with flags with duplicate PKs (I want to document this once we're done all the associated work).

there's no ADR discussing this, it's a scope creep on top of the original goal of supporting arrays of flags

If somebody requests this feature, I'm happy to consider adding it, but I truly don't think it's worth the complexity before that point, and, so far, this functionality hasn't been agreed on.

I agree that this doesn't seem like an edge case we need to support now. When we document this feature, we can explicitly mention that duplicate flag key + flag set ID combos are not supported. That was the uniqueness constraint we agreed on in the ADR.

i still think that as a library/tool implementor, it is my responsibility to provide a solution for all the possible usecases which the configuration allows, although it is not within my usecase. - i reverted the change, i commented out the tests, in which i am loading the sources in different kind of orders (might still be the case that his is happening due to latency of different sources). so they could be easier activated if needed

toddbaert · 2025-09-10T14:34:11Z

@aepfli @tangenti

I re-ran the benchmark: 3e99ad4 (no significant deviation, looks like 15% faster/slower across various tests).

I hate that we don't have automation for this; I may have mentioned before that we once did, but github actions are far too vulnerable to noisy-neighbor effects to be useful in this sort of benchmarking.

aepfli · 2025-09-29T18:16:11Z

/gemini review

gemini-code-assist

Code Review

This pull request introduces a major internal refactoring to transition the flag storage from a map-based to a slice-based structure, enabling support for array-based flag configurations and multiple flag sets. The changes are extensive, touching core components like the store and evaluator, and simplifying their interfaces. While this is a valuable improvement, I've identified some critical issues in the new store implementation related to memdb usage that could lead to data loss and incorrect priority handling. Additionally, some utility functions that convert flag slices to maps do not correctly handle duplicate keys, which could cause data loss for clients of the sync and eventing services.

toddbaert

@aepfli Thanks!

I made a few small changes in this commit: a47ddd6

removed branch from submodule
fixed a race condition in a test (only the test itself had the race, not the prod code)
removed an unused type

Note this is only a change in the core lib, not flagd itself, so I will modify the flagd release accordingly when we release this.

Now that the index changes have been reverted, I will merge this next week unless I hear objections. cc @chrfwow @tangenti @beeme1mr

Signed-off-by: Simon Schrottner <simon.schrottner@dynatrace.com> diff --git c/core/pkg/evaluator/fractional_test.go i/core/pkg/evaluator/fractional_test.go index e933e86..c1dfb9a 100644 --- c/core/pkg/evaluator/fractional_test.go +++ i/core/pkg/evaluator/fractional_test.go @@ -15,11 +15,12 @@ func TestFractionalEvaluation(t *testing.T) { var sources = []string{source} ctx := context.Background() - commonFlags := map[string]model.Flag{ - "headerColor": { + commonFlags := []model.Flag{ + { + Key: "headerColor", State: "ENABLED", DefaultVariant: "red", - Variants: colorVariants, + Variants: colorVariants, Targeting: []byte(`{ "if": [ { @@ -51,10 +52,11 @@ func TestFractionalEvaluation(t *testing.T) { ] }`), }, - "customSeededHeaderColor": { + { + Key: "customSeededHeaderColor", State: "ENABLED", DefaultVariant: "red", - Variants: colorVariants, + Variants: colorVariants, Targeting: []byte(`{ "if": [ { @@ -77,7 +79,7 @@ func TestFractionalEvaluation(t *testing.T) { } tests := map[string]struct { - flags map[string]model.Flag + flags []model.Flag flagKey string context map[string]any expectedValue string @@ -166,12 +168,12 @@ func TestFractionalEvaluation(t *testing.T) { expectedReason: model.TargetingMatchReason, }, "ross@faas.com with different flag key": { - flags: map[string]model.Flag{ - "footerColor": { - State: "ENABLED", - DefaultVariant: "red", - Variants: colorVariants, - Targeting: []byte(`{ + flags: []model.Flag{{ + Key: "footerColor", + State: "ENABLED", + DefaultVariant: "red", + Variants: colorVariants, + Targeting: []byte(`{ "if": [ { "in": ["@faas.com", { @@ -201,7 +203,7 @@ func TestFractionalEvaluation(t *testing.T) { }, null ] }`), - }, + }, }, flagKey: "footerColor", context: map[string]any{ @@ -212,12 +214,12 @@ func TestFractionalEvaluation(t *testing.T) { expectedReason: model.TargetingMatchReason, }, "non even split": { - flags: map[string]model.Flag{ - "headerColor": { - State: "ENABLED", - DefaultVariant: "red", - Variants: colorVariants, - Targeting: []byte(`{ + flags: []model.Flag{{ + Key: "headerColor", + State: "ENABLED", + DefaultVariant: "red", + Variants: colorVariants, + Targeting: []byte(`{ "if": [ { "in": ["@faas.com", { @@ -243,7 +245,7 @@ func TestFractionalEvaluation(t *testing.T) { }, null ] }`), - }, + }, }, flagKey: "headerColor", context: map[string]any{ @@ -254,12 +256,12 @@ func TestFractionalEvaluation(t *testing.T) { expectedReason: model.TargetingMatchReason, }, "fallback to default variant if no email provided": { - flags: map[string]model.Flag{ - "headerColor": { - State: "ENABLED", - DefaultVariant: "red", - Variants: colorVariants, - Targeting: []byte(`{ + flags: []model.Flag{{ + Key: "headerColor", + State: "ENABLED", + DefaultVariant: "red", + Variants: colorVariants, + Targeting: []byte(`{ "fractional": [ {"var": "email"}, [ @@ -280,7 +282,7 @@ func TestFractionalEvaluation(t *testing.T) { ] ] }`), - }, + }, }, flagKey: "headerColor", context: map[string]any{}, @@ -289,12 +291,12 @@ func TestFractionalEvaluation(t *testing.T) { expectedReason: model.DefaultReason, }, "get variant for non-percentage weight values": { - flags: map[string]model.Flag{ - "headerColor": { - State: "ENABLED", - DefaultVariant: "red", - Variants: colorVariants, - Targeting: []byte(`{ + flags: []model.Flag{{ + Key: "headerColor", + State: "ENABLED", + DefaultVariant: "red", + Variants: colorVariants, + Targeting: []byte(`{ "fractional": [ {"var": "email"}, [ @@ -307,7 +309,7 @@ func TestFractionalEvaluation(t *testing.T) { ] ] }`), - }, + }, }, flagKey: "headerColor", context: map[string]any{ @@ -318,12 +320,12 @@ func TestFractionalEvaluation(t *testing.T) { expectedReason: model.TargetingMatchReason, }, "get variant for non-specified weight values": { - flags: map[string]model.Flag{ - "headerColor": { - State: "ENABLED", - DefaultVariant: "red", - Variants: colorVariants, - Targeting: []byte(`{ + flags: []model.Flag{{ + Key: "headerColor", + State: "ENABLED", + DefaultVariant: "red", + Variants: colorVariants, + Targeting: []byte(`{ "fractional": [ {"var": "email"}, [ @@ -334,7 +336,7 @@ func TestFractionalEvaluation(t *testing.T) { ] ] }`), - }, + }, }, flagKey: "headerColor", context: map[string]any{ @@ -345,12 +347,12 @@ func TestFractionalEvaluation(t *testing.T) { expectedReason: model.TargetingMatchReason, }, "default to targetingKey if no bucket key provided": { - flags: map[string]model.Flag{ - "headerColor": { - State: "ENABLED", - DefaultVariant: "red", - Variants: colorVariants, - Targeting: []byte(`{ + flags: []model.Flag{{ + Key: "headerColor", + State: "ENABLED", + DefaultVariant: "red", + Variants: colorVariants, + Targeting: []byte(`{ "fractional": [ [ "blue", @@ -362,7 +364,7 @@ func TestFractionalEvaluation(t *testing.T) { ] ] }`), - }, + }, }, flagKey: "headerColor", context: map[string]any{ @@ -373,20 +375,20 @@ func TestFractionalEvaluation(t *testing.T) { expectedReason: model.TargetingMatchReason, }, "missing email - parser should ignore nil/missing custom variables and continue": { - flags: map[string]model.Flag{ - "headerColor": { - State: "ENABLED", - DefaultVariant: "red", - Variants: colorVariants, - Targeting: []byte( - `{ + flags: []model.Flag{{ + Key: "headerColor", + State: "ENABLED", + DefaultVariant: "red", + Variants: colorVariants, + Targeting: []byte( + `{ "fractional": [ {"var": "email"}, ["red",50], ["blue",50] ] }`), - }, + }, }, flagKey: "headerColor", context: map[string]any{ @@ -438,12 +440,12 @@ func BenchmarkFractionalEvaluation(b *testing.B) { var sources = []string{source} ctx := context.Background() - flags := map[string]model.Flag{ - "headerColor": { - State: "ENABLED", - DefaultVariant: "red", - Variants: colorVariants, - Targeting: []byte(`{ + flags := []model.Flag{{ + Key: "headerColor", + State: "ENABLED", + DefaultVariant: "red", + Variants: colorVariants, + Targeting: []byte(`{ "if": [ { "in": ["@faas.com", { @@ -473,11 +475,11 @@ func BenchmarkFractionalEvaluation(b *testing.B) { }, null ] }`), - }, + }, } tests := map[string]struct { - flags map[string]model.Flag + flags []model.Flag flagKey string context map[string]any expectedValue string diff --git c/core/pkg/evaluator/json.go i/core/pkg/evaluator/json.go index 12b862b..5f8fce1 100644 --- c/core/pkg/evaluator/json.go +++ i/core/pkg/evaluator/json.go @@ -6,6 +6,7 @@ import ( "encoding/json" "errors" "fmt" + "github.com/xeipuuv/gojsonschema" "regexp" "strings" "time" @@ -177,7 +178,7 @@ func (je *Resolver) ResolveAllValues(ctx context.Context, reqID string, context var reason string var metadata map[string]interface{} - for flagKey, flag := range allFlags { + for _, flag := range allFlags { if flag.State == Disabled { // ignore evaluation of disabled flag continue @@ -186,18 +187,18 @@ func (je *Resolver) ResolveAllValues(ctx context.Context, reqID string, context defaultValue := flag.Variants[flag.DefaultVariant] switch defaultValue.(type) { case bool: - value, variant, reason, metadata, err = resolve[bool](ctx, reqID, flagKey, context, je.evaluateVariant) + value, variant, reason, metadata, err = resolve[bool](ctx, reqID, flag.Key, context, je.evaluateVariant) case string: - value, variant, reason, metadata, err = resolve[string](ctx, reqID, flagKey, context, je.evaluateVariant) + value, variant, reason, metadata, err = resolve[string](ctx, reqID, flag.Key, context, je.evaluateVariant) case float64: - value, variant, reason, metadata, err = resolve[float64](ctx, reqID, flagKey, context, je.evaluateVariant) + value, variant, reason, metadata, err = resolve[float64](ctx, reqID, flag.Key, context, je.evaluateVariant) case map[string]any: - value, variant, reason, metadata, err = resolve[map[string]any](ctx, reqID, flagKey, context, je.evaluateVariant) + value, variant, reason, metadata, err = resolve[map[string]any](ctx, reqID, flag.Key, context, je.evaluateVariant) } if err != nil { - je.Logger.ErrorWithID(reqID, fmt.Sprintf("bulk evaluation: key: %s returned error: %s", flagKey, err.Error())) + je.Logger.ErrorWithID(reqID, fmt.Sprintf("bulk evaluation: key: %s returned error: %s", flag.Key, err.Error())) } - values = append(values, NewAnyValue(value, variant, reason, flagKey, metadata, err)) + values = append(values, NewAnyValue(value, variant, reason, flag.Key, metadata, err)) } return values, flagSetMetadata, nil @@ -453,23 +454,32 @@ func (je *JSON) configToFlagDefinition(config string, definition *Definition) er "flag definition does not conform to the schema; validation errors: %s", err), ) } - + type JsonRawDef struct { + Flags map[string]model.Flag `json:"flags"` + Metadata map[string]interface{} `json:"metadata"` + } + // Transpose evaluators and unmarshal directly into JsonDef transposedConfig, err := transposeEvaluators(config) if err != nil { return fmt.Errorf("transposing evaluators: %w", err) } - err = json.Unmarshal([]byte(transposedConfig), &definition) + var rawDef JsonRawDef + err = json.Unmarshal([]byte(transposedConfig), &rawDef) if err != nil { return fmt.Errorf("unmarshalling provided configurations: %w", err) } - + definition.Metadata = rawDef.Metadata + for s, flag := range rawDef.Flags { + flag.Key = s + definition.Flags = append(definition.Flags, flag) + } return validateDefaultVariants(definition) } // validateDefaultVariants returns an error if any of the default variants aren't valid func validateDefaultVariants(flags *Definition) error { - for name, flag := range flags.Flags { + for _, flag := range flags.Flags { // Default Variant is not provided in the config if flag.DefaultVariant == "" { continue @@ -477,7 +487,7 @@ func validateDefaultVariants(flags *Definition) error { if _, ok := flag.Variants[flag.DefaultVariant]; !ok { return fmt.Errorf( - "default variant: '%s' isn't a valid variant of flag: '%s'", flag.DefaultVariant, name, + "default variant: '%s' isn't a valid variant of flag: '%s'", flag.DefaultVariant, flag.Key, ) } } diff --git c/core/pkg/evaluator/json_model.go i/core/pkg/evaluator/json_model.go index 0f09eec..826f390 100644 --- c/core/pkg/evaluator/json_model.go +++ i/core/pkg/evaluator/json_model.go @@ -11,7 +11,7 @@ type Evaluators struct { } type Definition struct { - Flags map[string]model.Flag `json:"flags"` + Flags []model.Flag `json:"flags"` Metadata map[string]interface{} `json:"metadata"` } diff --git c/core/pkg/evaluator/semver_test.go i/core/pkg/evaluator/semver_test.go index fbc6582..52f59a9 100644 --- c/core/pkg/evaluator/semver_test.go +++ i/core/pkg/evaluator/semver_test.go @@ -321,7 +321,7 @@ func TestJSONEvaluator_semVerEvaluation(t *testing.T) { ctx := context.Background() tests := map[string]struct { - flags map[string]model.Flag + flags []model.Flag flagKey string context map[string]any expectedValue string @@ -330,12 +330,12 @@ func TestJSONEvaluator_semVerEvaluation(t *testing.T) { expectedError error }{ "versions and operator provided - match": { - flags: map[string]model.Flag{ - "headerColor": { - State: "ENABLED", - DefaultVariant: "red", - Variants: colorVariants, - Targeting: []byte(`{ + flags: []model.Flag{{ + Key: "headerColor", + State: "ENABLED", + DefaultVariant: "red", + Variants: colorVariants, + Targeting: []byte(`{ "if": [ { "sem_ver": ["1.0.0", ">", "0.1.0"] @@ -343,7 +343,7 @@ func TestJSONEvaluator_semVerEvaluation(t *testing.T) { "red", null ] }`), - }, + }, }, flagKey: "headerColor", context: map[string]any{ @@ -354,12 +354,12 @@ func TestJSONEvaluator_semVerEvaluation(t *testing.T) { expectedReason: model.TargetingMatchReason, }, "resolve target property using nested operation - match": { - flags: map[string]model.Flag{ - "headerColor": { - State: "ENABLED", - DefaultVariant: "red", - Variants: colorVariants, - Targeting: []byte(`{ + flags: []model.Flag{{ + Key: "headerColor", + State: "ENABLED", + DefaultVariant: "red", + Variants: colorVariants, + Targeting: []byte(`{ "if": [ { "sem_ver": [{"var": "version"}, ">", "1.0.0"] @@ -367,7 +367,7 @@ func TestJSONEvaluator_semVerEvaluation(t *testing.T) { "red", null ] }`), - }, + }, }, flagKey: "headerColor", context: map[string]any{ @@ -378,12 +378,12 @@ func TestJSONEvaluator_semVerEvaluation(t *testing.T) { expectedReason: model.TargetingMatchReason, }, "versions and operator provided - no match": { - flags: map[string]model.Flag{ - "headerColor": { - State: "ENABLED", - DefaultVariant: "red", - Variants: colorVariants, - Targeting: []byte(`{ + flags: []model.Flag{{ + Key: "headerColor", + State: "ENABLED", + DefaultVariant: "red", + Variants: colorVariants, + Targeting: []byte(`{ "if": [ { "sem_ver": ["1.0.0", ">", "1.0.0"] @@ -391,7 +391,7 @@ func TestJSONEvaluator_semVerEvaluation(t *testing.T) { "red", "green" ] }`), - }, + }, }, flagKey: "headerColor", context: map[string]any{ @@ -402,12 +402,12 @@ func TestJSONEvaluator_semVerEvaluation(t *testing.T) { expectedReason: model.TargetingMatchReason, }, "versions and major-version operator provided - match": { - flags: map[string]model.Flag{ - "headerColor": { - State: "ENABLED", - DefaultVariant: "red", - Variants: colorVariants, - Targeting: []byte(`{ + flags: []model.Flag{{ + Key: "headerColor", + State: "ENABLED", + DefaultVariant: "red", + Variants: colorVariants, + Targeting: []byte(`{ "if": [ { "sem_ver": ["1.2.3", "^", "1.5.6"] @@ -415,7 +415,7 @@ func TestJSONEvaluator_semVerEvaluation(t *testing.T) { "red", "green" ] }`), - }, + }, }, flagKey: "headerColor", context: map[string]any{ @@ -426,12 +426,12 @@ func TestJSONEvaluator_semVerEvaluation(t *testing.T) { expectedReason: model.TargetingMatchReason, }, "versions and minor-version operator provided - match": { - flags: map[string]model.Flag{ - "headerColor": { - State: "ENABLED", - DefaultVariant: "red", - Variants: colorVariants, - Targeting: []byte(`{ + flags: []model.Flag{{ + Key: "headerColor", + State: "ENABLED", + DefaultVariant: "red", + Variants: colorVariants, + Targeting: []byte(`{ "if": [ { "sem_ver": ["1.2.3", "~", "1.2.6"] @@ -439,7 +439,7 @@ func TestJSONEvaluator_semVerEvaluation(t *testing.T) { "red", "green" ] }`), - }, + }, }, flagKey: "headerColor", context: map[string]any{ @@ -450,12 +450,12 @@ func TestJSONEvaluator_semVerEvaluation(t *testing.T) { expectedReason: model.TargetingMatchReason, }, "versions given as double - match": { - flags: map[string]model.Flag{ - "headerColor": { - State: "ENABLED", - DefaultVariant: "red", - Variants: colorVariants, - Targeting: []byte(`{ + flags: []model.Flag{{ + Key: "headerColor", + State: "ENABLED", + DefaultVariant: "red", + Variants: colorVariants, + Targeting: []byte(`{ "if": [ { "sem_ver": [1.2, "=", "1.2"] @@ -463,7 +463,7 @@ func TestJSONEvaluator_semVerEvaluation(t *testing.T) { "red", "green" ] }`), - }, + }, }, flagKey: "headerColor", context: map[string]any{ @@ -474,12 +474,12 @@ func TestJSONEvaluator_semVerEvaluation(t *testing.T) { expectedReason: model.TargetingMatchReason, }, "versions given as int - match": { - flags: map[string]model.Flag{ - "headerColor": { - State: "ENABLED", - DefaultVariant: "red", - Variants: colorVariants, - Targeting: []byte(`{ + flags: []model.Flag{{ + Key: "headerColor", + State: "ENABLED", + DefaultVariant: "red", + Variants: colorVariants, + Targeting: []byte(`{ "if": [ { "sem_ver": [1, "=", "v1.0.0"] @@ -487,7 +487,7 @@ func TestJSONEvaluator_semVerEvaluation(t *testing.T) { "red", "green" ] }`), - }, + }, }, flagKey: "headerColor", context: map[string]any{ @@ -498,12 +498,12 @@ func TestJSONEvaluator_semVerEvaluation(t *testing.T) { expectedReason: model.TargetingMatchReason, }, "versions and minor-version without patch version operator provided - match": { - flags: map[string]model.Flag{ - "headerColor": { - State: "ENABLED", - DefaultVariant: "red", - Variants: colorVariants, - Targeting: []byte(`{ + flags: []model.Flag{{ + Key: "headerColor", + State: "ENABLED", + DefaultVariant: "red", + Variants: colorVariants, + Targeting: []byte(`{ "if": [ { "sem_ver": [1.2, "=", "1.2"] @@ -511,7 +511,7 @@ func TestJSONEvaluator_semVerEvaluation(t *testing.T) { "red", "green" ] }`), - }, + }, }, flagKey: "headerColor", context: map[string]any{ @@ -522,12 +522,12 @@ func TestJSONEvaluator_semVerEvaluation(t *testing.T) { expectedReason: model.TargetingMatchReason, }, "versions with prefixed v operator provided - match": { - flags: map[string]model.Flag{ - "headerColor": { - State: "ENABLED", - DefaultVariant: "red", - Variants: colorVariants, - Targeting: []byte(`{ + flags: []model.Flag{{ + Key: "headerColor", + State: "ENABLED", + DefaultVariant: "red", + Variants: colorVariants, + Targeting: []byte(`{ "if": [ { "sem_ver": [{"var": "version"}, "<", "v1.2"] @@ -535,7 +535,7 @@ func TestJSONEvaluator_semVerEvaluation(t *testing.T) { "red", "green" ] }`), - }, + }, }, flagKey: "headerColor", context: map[string]any{ @@ -546,12 +546,12 @@ func TestJSONEvaluator_semVerEvaluation(t *testing.T) { expectedReason: model.TargetingMatchReason, }, "versions and major-version operator provided - no match": { - flags: map[string]model.Flag{ - "headerColor": { - State: "ENABLED", - DefaultVariant: "red", - Variants: colorVariants, - Targeting: []byte(`{ + flags: []model.Flag{{ + Key: "headerColor", + State: "ENABLED", + DefaultVariant: "red", + Variants: colorVariants, + Targeting: []byte(`{ "if": [ { "sem_ver": ["2.2.3", "^", "1.2.3"] @@ -559,7 +559,7 @@ func TestJSONEvaluator_semVerEvaluation(t *testing.T) { "red", "green" ] }`), - }, + }, }, flagKey: "headerColor", context: map[string]any{ @@ -570,12 +570,12 @@ func TestJSONEvaluator_semVerEvaluation(t *testing.T) { expectedReason: model.TargetingMatchReason, }, "versions and minor-version operator provided - no match": { - flags: map[string]model.Flag{ - "headerColor": { - State: "ENABLED", - DefaultVariant: "red", - Variants: colorVariants, - Targeting: []byte(`{ + flags: []model.Flag{{ + Key: "headerColor", + State: "ENABLED", + DefaultVariant: "red", + Variants: colorVariants, + Targeting: []byte(`{ "if": [ { "sem_ver": ["1.3.3", "~", "1.2.6"] @@ -583,7 +583,7 @@ func TestJSONEvaluator_semVerEvaluation(t *testing.T) { "red", "green" ] }`), - }, + }, }, flagKey: "headerColor", context: map[string]any{ @@ -594,12 +594,12 @@ func TestJSONEvaluator_semVerEvaluation(t *testing.T) { expectedReason: model.TargetingMatchReason, }, "resolve target property using nested operation - no match": { - flags: map[string]model.Flag{ - "headerColor": { - State: "ENABLED", - DefaultVariant: "red", - Variants: colorVariants, - Targeting: []byte(`{ + flags: []model.Flag{{ + Key: "headerColor", + State: "ENABLED", + DefaultVariant: "red", + Variants: colorVariants, + Targeting: []byte(`{ "if": [ { "sem_ver": [{"var": "version"}, ">", "1.0.0"] @@ -607,7 +607,7 @@ func TestJSONEvaluator_semVerEvaluation(t *testing.T) { "red", "green" ] }`), - }, + }, }, flagKey: "headerColor", context: map[string]any{ @@ -618,12 +618,12 @@ func TestJSONEvaluator_semVerEvaluation(t *testing.T) { expectedReason: model.TargetingMatchReason, }, "error during parsing (not an array) - return default": { - flags: map[string]model.Flag{ - "headerColor": { - State: "ENABLED", - DefaultVariant: "red", - Variants: colorVariants, - Targeting: []byte(`{ + flags: []model.Flag{{ + Key: "headerColor", + State: "ENABLED", + DefaultVariant: "red", + Variants: colorVariants, + Targeting: []byte(`{ "if": [ { "sem_ver": "not an array" @@ -631,7 +631,7 @@ func TestJSONEvaluator_semVerEvaluation(t *testing.T) { "red", "green" ] }`), - }, + }, }, flagKey: "headerColor", context: map[string]any{ @@ -642,12 +642,12 @@ func TestJSONEvaluator_semVerEvaluation(t *testing.T) { expectedReason: model.TargetingMatchReason, }, "error during parsing (wrong number of items in array) - return default": { - flags: map[string]model.Flag{ - "headerColor": { - State: "ENABLED", - DefaultVariant: "red", - Variants: colorVariants, - Targeting: []byte(`{ + flags: []model.Flag{{ + Key: "headerColor", + State: "ENABLED", + DefaultVariant: "red", + Variants: colorVariants, + Targeting: []byte(`{ "if": [ { "sem_ver": ["not", "enough"] @@ -655,7 +655,7 @@ func TestJSONEvaluator_semVerEvaluation(t *testing.T) { "red", "green" ] }`), - }, + }, }, flagKey: "headerColor", context: map[string]any{ @@ -666,12 +666,12 @@ func TestJSONEvaluator_semVerEvaluation(t *testing.T) { expectedReason: model.TargetingMatchReason, }, "error during parsing (invalid property value) - return default": { - flags: map[string]model.Flag{ - "headerColor": { - State: "ENABLED", - DefaultVariant: "red", - Variants: colorVariants, - Targeting: []byte(`{ + flags: []model.Flag{{ + Key: "headerColor", + State: "ENABLED", + DefaultVariant: "red", + Variants: colorVariants, + Targeting: []byte(`{ "if": [ { "sem_ver": ["invalid", ">", "1.0.0"] @@ -679,8 +679,9 @@ func TestJSONEvaluator_semVerEvaluation(t *testing.T) { "red", "green" ] }`), - }, }, + }, + flagKey: "headerColor", context: map[string]any{ "email": "user@faas.com", @@ -690,12 +691,12 @@ func TestJSONEvaluator_semVerEvaluation(t *testing.T) { expectedReason: model.TargetingMatchReason, }, "error during parsing (invalid property type) - return default": { - flags: map[string]model.Flag{ - "headerColor": { - State: "ENABLED", - DefaultVariant: "red", - Variants: colorVariants, - Targeting: []byte(`{ + flags: []model.Flag{{ + Key: "headerColor", + State: "ENABLED", + DefaultVariant: "red", + Variants: colorVariants, + Targeting: []byte(`{ "if": [ { "sem_ver": [1.0, ">", "1.0.0"] @@ -703,8 +704,9 @@ func TestJSONEvaluator_semVerEvaluation(t *testing.T) { "red", "green" ] }`), - }, }, + }, + flagKey: "headerColor", context: map[string]any{ "email": "user@faas.com", @@ -714,12 +716,12 @@ func TestJSONEvaluator_semVerEvaluation(t *testing.T) { expectedReason: model.TargetingMatchReason, }, "error during parsing (invalid operator) - return default": { - flags: map[string]model.Flag{ - "headerColor": { - State: "ENABLED", - DefaultVariant: "red", - Variants: colorVariants, - Targeting: []byte(`{ + flags: []model.Flag{{ + Key: "headerColor", + State: "ENABLED", + DefaultVariant: "red", + Variants: colorVariants, + Targeting: []byte(`{ "if": [ { "sem_ver": ["1.0.0", "invalid", "1.0.0"] @@ -727,7 +729,7 @@ func TestJSONEvaluator_semVerEvaluation(t *testing.T) { "red", "green" ] }`), - }, + }, }, flagKey: "headerColor", context: map[string]any{ @@ -738,12 +740,12 @@ func TestJSONEvaluator_semVerEvaluation(t *testing.T) { expectedReason: model.TargetingMatchReason, }, "error during parsing (invalid operator type) - return default": { - flags: map[string]model.Flag{ - "headerColor": { - State: "ENABLED", - DefaultVariant: "red", - Variants: colorVariants, - Targeting: []byte(`{ + flags: []model.Flag{{ + Key: "headerColor", + State: "ENABLED", + DefaultVariant: "red", + Variants: colorVariants, + Targeting: []byte(`{ "if": [ { "sem_ver": ["1.0.0", 1, "1.0.0"] @@ -751,7 +753,7 @@ func TestJSONEvaluator_semVerEvaluation(t *testing.T) { "red", "green" ] }`), - }, + }, }, flagKey: "headerColor", context: map[string]any{ @@ -762,12 +764,12 @@ func TestJSONEvaluator_semVerEvaluation(t *testing.T) { expectedReason: model.TargetingMatchReason, }, "error during parsing (invalid target version) - return default": { - flags: map[string]model.Flag{ - "headerColor": { - State: "ENABLED", - DefaultVariant: "red", - Variants: colorVariants, - Targeting: []byte(`{ + flags: []model.Flag{{ + Key: "headerColor", + State: "ENABLED", + DefaultVariant: "red", + Variants: colorVariants, + Targeting: []byte(`{ "if": [ { "sem_ver": ["1.0.0", ">", "invalid"] @@ -775,7 +777,7 @@ func TestJSONEvaluator_semVerEvaluation(t *testing.T) { "red", "green" ] }`), - }, + }, }, flagKey: "headerColor", context: map[string]any{ diff --git c/core/pkg/evaluator/string_comparison_test.go i/core/pkg/evaluator/string_comparison_test.go index 3e6163c..f22466f 100644 --- c/core/pkg/evaluator/string_comparison_test.go +++ i/core/pkg/evaluator/string_comparison_test.go @@ -18,7 +18,7 @@ func TestJSONEvaluator_startsWithEvaluation(t *testing.T) { ctx := context.Background() tests := map[string]struct { - flags map[string]model.Flag + flags []model.Flag flagKey string context map[string]any expectedValue string @@ -27,12 +27,12 @@ func TestJSONEvaluator_startsWithEvaluation(t *testing.T) { expectedError error }{ "two strings provided - match": { - flags: map[string]model.Flag{ - "headerColor": { - State: "ENABLED", - DefaultVariant: "red", - Variants: colorVariants, - Targeting: []byte(`{ + flags: []model.Flag{{ + Key: "headerColor", + State: "ENABLED", + DefaultVariant: "red", + Variants: colorVariants, + Targeting: []byte(`{ "if": [ { "starts_with": ["user@faas.com", "user@faas"] @@ -40,7 +40,7 @@ func TestJSONEvaluator_startsWithEvaluation(t *testing.T) { "red", null ] }`), - }, + }, }, flagKey: "headerColor", context: map[string]any{ @@ -51,12 +51,12 @@ func TestJSONEvaluator_startsWithEvaluation(t *testing.T) { expectedReason: model.TargetingMatchReason, }, "resolve target property using nested operation - match": { - flags: map[string]model.Flag{ - "headerColor": { - State: "ENABLED", - DefaultVariant: "red", - Variants: colorVariants, - Targeting: []byte(`{ + flags: []model.Flag{{ + Key: "headerColor", + State: "ENABLED", + DefaultVariant: "red", + Variants: colorVariants, + Targeting: []byte(`{ "if": [ { "starts_with": [{"var": "email"}, "user@faas"] @@ -64,7 +64,7 @@ func TestJSONEvaluator_startsWithEvaluation(t *testing.T) { "red", null ] }`), - }, + }, }, flagKey: "headerColor", context: map[string]any{ @@ -75,12 +75,12 @@ func TestJSONEvaluator_startsWithEvaluation(t *testing.T) { expectedReason: model.TargetingMatchReason, }, "two strings provided - no match": { - flags: map[string]model.Flag{ - "headerColor": { - State: "ENABLED", - DefaultVariant: "red", - Variants: colorVariants, - Targeting: []byte(`{ + flags: []model.Flag{{ + Key: "headerColor", + State: "ENABLED", + DefaultVariant: "red", + Variants: colorVariants, + Targeting: []byte(`{ "if": [ { "starts_with": ["user@faas.com", "nope"] @@ -88,7 +88,7 @@ func TestJSONEvaluator_startsWithEvaluation(t *testing.T) { "red", "green" ] }`), - }, + }, }, flagKey: "headerColor", context: map[string]any{ @@ -99,12 +99,12 @@ func TestJSONEvaluator_startsWithEvaluation(t *testing.T) { expectedReason: model.TargetingMatchReason, }, "resolve target property using nested operation - no match": { - flags: map[string]model.Flag{ - "headerColor": { - State: "ENABLED", - DefaultVariant: "red", - Variants: colorVariants, - Targeting: []byte(`{ + flags: []model.Flag{{ + Key: "headerColor", + State: "ENABLED", + DefaultVariant: "red", + Variants: colorVariants, + Targeting: []byte(`{ "if": [ { "starts_with": [{"var": "email"}, "nope"] @@ -112,7 +112,7 @@ func TestJSONEvaluator_startsWithEvaluation(t *testing.T) { "red", "green" ] }`), - }, + }, }, flagKey: "headerColor", context: map[string]any{ @@ -123,12 +123,12 @@ func TestJSONEvaluator_startsWithEvaluation(t *testing.T) { expectedReason: model.TargetingMatchReason, }, "error during parsing - return default": { - flags: map[string]model.Flag{ - "headerColor": { - State: "ENABLED", - DefaultVariant: "red", - Variants: colorVariants, - Targeting: []byte(`{ + flags: []model.Flag{{ + Key: "headerColor", + State: "ENABLED", + DefaultVariant: "red", + Variants: colorVariants, + Targeting: []byte(`{ "if": [ { "starts_with": "no-array" @@ -136,7 +136,7 @@ func TestJSONEvaluator_startsWithEvaluation(t *testing.T) { "red", "green" ] }`), - }, + }, }, flagKey: "headerColor", context: map[string]any{ @@ -186,7 +186,7 @@ func TestJSONEvaluator_endsWithEvaluation(t *testing.T) { ctx := context.Background() tests := map[string]struct { - flags map[string]model.Flag + flags []model.Flag flagKey string context map[string]any expectedValue string @@ -195,12 +195,12 @@ func TestJSONEvaluator_endsWithEvaluation(t *testing.T) { expectedError error }{ "two strings provided - match": { - flags: map[string]model.Flag{ - "headerColor": { - State: "ENABLED", - DefaultVariant: "red", - Variants: colorVariants, - Targeting: []byte(`{ + flags: []model.Flag{{ + Key: "headerColor", + State: "ENABLED", + DefaultVariant: "red", + Variants: colorVariants, + Targeting: []byte(`{ "if": [ { "ends_with": ["user@faas.com", "faas.com"] @@ -208,7 +208,7 @@ func TestJSONEvaluator_endsWithEvaluation(t *testing.T) { "red", null ] }`), - }, + }, }, flagKey: "headerColor", context: map[string]any{ @@ -219,12 +219,12 @@ func TestJSONEvaluator_endsWithEvaluation(t *testing.T) { expectedReason: model.TargetingMatchReason, }, "resolve target property using nested operation - match": { - flags: map[string]model.Flag{ - "headerColor": { - State: "ENABLED", - DefaultVariant: "red", - Variants: colorVariants, - Targeting: []byte(`{ + flags: []model.Flag{{ + Key: "headerColor", + State: "ENABLED", + DefaultVariant: "red", + Variants: colorVariants, + Targeting: []byte(`{ "if": [ { "ends_with": [{"var": "email"}, "faas.com"] @@ -232,7 +232,7 @@ func TestJSONEvaluator_endsWithEvaluation(t *testing.T) { "red", null ] }`), - }, + }, }, flagKey: "headerColor", context: map[string]any{ @@ -243,12 +243,12 @@ func TestJSONEvaluator_endsWithEvaluation(t *testing.T) { expectedReason: model.TargetingMatchReason, }, "two strings provided - no match": { - flags: map[string]model.Flag{ - "headerColor": { - State: "ENABLED", - DefaultVariant: "red", - Variants: colorVariants, - Targeting: []byte(`{ + flags: []model.Flag{{ + Key: "headerColor", + State: "ENABLED", + DefaultVariant: "red", + Variants: colorVariants, + Targeting: []byte(`{ "if": [ { "ends_with": ["user@faas.com", "nope"] @@ -256,7 +256,7 @@ func TestJSONEvaluator_endsWithEvaluation(t *testing.T) { "red", "green" ] }`), - }, + }, }, flagKey: "headerColor", context: map[string]any{ @@ -267,12 +267,12 @@ func TestJSONEvaluator_endsWithEvaluation(t *testing.T) { expectedReason: model.TargetingMatchReason, }, "resolve target property using nested operation - no match": { - flags: map[string]model.Flag{ - "headerColor": { - State: "ENABLED", - DefaultVariant: "red", - Variants: colorVariants, - Targeting: []byte(`{ + flags: []model.Flag{{ + Key: "headerColor", + State: "ENABLED", + DefaultVariant: "red", + Variants: colorVariants, + Targeting: []byte(`{ "if": [ { "ends_with": [{"var": "email"}, "nope"] @@ -280,7 +280,7 @@ func TestJSONEvaluator_endsWithEvaluation(t *testing.T) { "red", "green" ] }`), - }, + }, }, flagKey: "headerColor", context: map[string]any{ @@ -291,12 +291,12 @@ func TestJSONEvaluator_endsWithEvaluation(t *testing.T) { expectedReason: model.TargetingMatchReason, }, "error during parsing - return default": { - flags: map[string]model.Flag{ - "headerColor": { - State: "ENABLED", - DefaultVariant: "red", - Variants: colorVariants, - Targeting: []byte(`{ + flags: []model.Flag{{ + Key: "headerColor", + State: "ENABLED", + DefaultVariant: "red", + Variants: colorVariants, + Targeting: []byte(`{ "if": [ { "ends_with": "no-array" @@ -304,7 +304,7 @@ func TestJSONEvaluator_endsWithEvaluation(t *testing.T) { "red", "green" ] }`), - }, + }, }, flagKey: "headerColor", context: map[string]any{ diff --git c/core/pkg/store/store.go i/core/pkg/store/store.go index 0c6bc16..404604a 100644 --- c/core/pkg/store/store.go +++ i/core/pkg/store/store.go @@ -4,6 +4,7 @@ import ( "context" "fmt" "slices" + "sort" "github.com/hashicorp/go-memdb" "github.com/open-feature/flagd/core/pkg/logger" @@ -15,14 +16,14 @@ var noValidatedSources = []string{} type SelectorContextKey struct{} type FlagQueryResult struct { - Flags map[string]model.Flag + Flags []model.Flag } type IStore interface { Get(ctx context.Context, key string, selector *Selector) (model.Flag, model.Metadata, error) - GetAll(ctx context.Context, selector *Selector) (map[string]model.Flag, model.Metadata, error) + GetAll(ctx context.Context, selector *Selector) ([]model.Flag, model.Metadata, error) Watch(ctx context.Context, selector *Selector, watcher chan<- FlagQueryResult) - Update(source string, flags map[string]model.Flag, metadata model.Metadata) + Update(source string, flags []model.Flag, metadata model.Metadata) } var _ IStore = (*Store)(nil) @@ -192,8 +193,8 @@ func (s *Store) Get(_ context.Context, key string, selector *Selector) (model.Fl } // GetAll returns a copy of the store's state (copy in order to be concurrency safe) -func (s *Store) GetAll(ctx context.Context, selector *Selector) (map[string]model.Flag, model.Metadata, error) { - flags := make(map[string]model.Flag) +func (s *Store) GetAll(ctx context.Context, selector *Selector) ([]model.Flag, model.Metadata, error) { + var flags []model.Flag queryMeta := selector.ToMetadata() it, err := s.selectOrAll(selector) @@ -208,7 +209,7 @@ func (s *Store) GetAll(ctx context.Context, selector *Selector) (map[string]mode // Update the flag state with the provided flags. func (s *Store) Update( source string, - flags map[string]model.Flag, + flags []model.Flag, metadata model.Metadata, ) { if source == "" { @@ -225,32 +226,10 @@ func (s *Store) Update( priority = 0 } - txn := s.db.Txn(true) - defer txn.Abort() - - // get all flags for the source we are updating - selector := NewSelector(sourceIndex + "=" + source) - oldFlags, _, _ := s.GetAll(context.Background(), &selector) - - for key := range oldFlags { - if _, ok := flags[key]; !ok { - // flag has been deleted - s.logger.Debug(fmt.Sprintf("flag %s has been deleted from source %s", key, source)) - - count, err := txn.DeleteAll(flagsTable, keySourceCompoundIndex, key, source) - s.logger.Debug(fmt.Sprintf("deleted %d flags with key %s from source %s", count, key, source)) - - if err != nil { - s.logger.Error(fmt.Sprintf("error deleting flag: %s, %v", key, err)) - } - continue - } - } - - for key, newFlag := range flags { + newFlags := make(map[string]model.Flag) + for _, newFlag := range flags { s.logger.Debug(fmt.Sprintf("got metadata %v", metadata)) - newFlag.Key = key newFlag.Source = source newFlag.Priority = priority newFlag.Metadata = patchMetadata(metadata, newFlag.Metadata) @@ -263,10 +242,42 @@ func (s *Store) Update( flagSetId = setFlagSetId } newFlag.FlagSetId = flagSetId + newFlags[newFlag.FlagSetId+"|"+newFlag.Key] = newFlag + } - raw, err := txn.First(flagsTable, keySourceCompoundIndex, key, source) + txn := s.db.Txn(true) + defer txn.Abort() + + // get all flags for the source we are updating + selector := NewSelector(sourceIndex + "=" + source) + oldFlags, _, _ := s.GetAll(context.Background(), &selector) + + for _, oldFlag := range oldFlags { + if _, ok := newFlags[oldFlag.FlagSetId+"|"+oldFlag.Key]; !ok { + // flag has been deleted + s.logger.Debug(fmt.Sprintf("flag '%s' and flagSetId '%s' has been deleted from source '%s'", oldFlag.Key, oldFlag.FlagSetId, source)) + + count, err := txn.DeleteAll(flagsTable, flagSetIdKeySourceCompoundIndex, oldFlag.FlagSetId, oldFlag.Key, source) + s.logger.Debug(fmt.Sprintf( + "deleted %d flags with key '%s' and flagSetId '%s' from source '%s'", + count, + oldFlag.Key, + oldFlag.FlagSetId, + source, + )) + + if err != nil { + s.logger.Error(fmt.Sprintf("error deleting flag: %s, %v", oldFlag.Key, err)) + } + continue + } + } + + for _, newFlag := range newFlags { + + raw, err := txn.First(flagsTable, keySourceCompoundIndex, newFlag.Key, source) if err != nil { - s.logger.Error(fmt.Sprintf("unable to get flag %s from source %s: %v", key, source, err)) + s.logger.Error(fmt.Sprintf("unable to get flag %s from source %s: %v", newFlag.Key, source, err)) continue } oldFlag, ok := raw.(model.Flag) @@ -275,9 +286,9 @@ func (s *Store) Update( if oldFlag.FlagSetId != newFlag.FlagSetId { // If the flagSetId is different, we need to delete the entry, since flagSetId+key represents the primary index, and it's now been changed. // This is important especially for clients listening to flagSetId changes, as they expect the flag to be removed from the set in this case. - _, err = txn.DeleteAll(flagsTable, idIndex, oldFlag.FlagSetId, key) + _, err = txn.DeleteAll(flagsTable, idIndex, oldFlag.FlagSetId, newFlag.Key) if err != nil { - s.logger.Error(fmt.Sprintf("unable to delete flags with key %s and flagSetId %s: %v", key, oldFlag.FlagSetId, err)) + s.logger.Error(fmt.Sprintf("unable to delete flags with key %s and flagSetId %s: %v", newFlag.Key, oldFlag.FlagSetId, err)) continue } } @@ -286,7 +297,7 @@ func (s *Store) Update( s.logger.Debug(fmt.Sprintf("storing flag: %v", newFlag)) err = txn.Insert(flagsTable, newFlag) if err != nil { - s.logger.Error(fmt.Sprintf("unable to insert flag %s: %v", key, err)) + s.logger.Error(fmt.Sprintf("unable to insert flag %s: %v", newFlag.Key, err)) continue } } @@ -335,20 +346,32 @@ func (s *Store) selectOrAll(selector *Selector) (it memdb.ResultIterator, err er } // collects flags from an iterator, ensuring that only the highest priority flag is kept when there are duplicates -func (s *Store) collect(it memdb.ResultIterator) map[string]model.Flag { +func (s *Store) collect(it memdb.ResultIterator) []model.Flag { flags := make(map[string]model.Flag) for raw := it.Next(); raw != nil; raw = it.Next() { flag := raw.(model.Flag) - if existing, ok := flags[flag.Key]; ok { + + // checking for multiple flags with the same key, as they can be defined multiple times in different sources + if existing, ok := flags[flag.FlagSetId+"|"+flag.Key]; ok { if flag.Priority < existing.Priority { - s.logger.Debug(fmt.Sprintf("discarding duplicate flag %s from lower priority source %s in favor of flag from source %s", flag.Key, s.sources[flag.Priority], s.sources[existing.Priority])) + s.logger.Debug(fmt.Sprintf("discarding duplicate flag with key '%s' and flagSetId '%s' from lower priority source '%s' in favor of flag from source '%s'", flag.Key, flag.FlagSetId, s.sources[flag.Priority], s.sources[existing.Priority])) continue // we already have a higher priority flag } - s.logger.Debug(fmt.Sprintf("overwriting duplicate flag %s from lower priority source %s in favor of flag from source %s", flag.Key, s.sources[existing.Priority], s.sources[flag.Priority])) + s.logger.Debug(fmt.Sprintf("overwriting duplicate flag with key '%s' and flagSetId '%s' from lower priority source '%s' in favor of flag from source '%s'", flag.Key, flag.FlagSetId, s.sources[existing.Priority], s.sources[flag.Priority])) } - flags[flag.Key] = flag + + flags[flag.FlagSetId+"|"+flag.Key] = flag } - return flags + + flattenedFlags := make([]model.Flag, 0, len(flags)) + for _, value := range flags { + flattenedFlags = append(flattenedFlags, value) + } + // we should order to keep the same order all the time in our response + sort.Slice(flattenedFlags, func(i, j int) bool { + return fmt.Sprintf("%s|%s", flattenedFlags[i].FlagSetId, flattenedFlags[i].Key) < fmt.Sprintf("%s|%s", flattenedFlags[j].FlagSetId, flattenedFlags[j].Key) + }) + return flattenedFlags } func patchMetadata(original, patch model.Metadata) model.Metadata { diff --git c/core/pkg/store/store_test.go i/core/pkg/store/store_test.go index c6cf2dd..f482e6a 100644 --- c/core/pkg/store/store_test.go +++ i/core/pkg/store/store_test.go @@ -2,6 +2,7 @@ package store import ( "context" + "sort" "testing" "time" @@ -21,9 +22,9 @@ func TestUpdateFlags(t *testing.T) { tests := []struct { name string setup func(t *testing.T) IStore - newFlags map[string]model.Flag + newFlags []model.Flag source string - wantFlags map[string]model.Flag + wantFlags []model.Flag setMetadata model.Metadata }{ { @@ -37,7 +38,7 @@ func TestUpdateFlags(t *testing.T) { }, source: source1, newFlags: nil, - wantFlags: map[string]model.Flag{}, + wantFlags: []model.Flag{}, }, { name: "both empty flags", @@ -49,8 +50,8 @@ func TestUpdateFlags(t *testing.T) { return s }, source: source1, - newFlags: map[string]model.Flag{}, - wantFlags: map[string]model.Flag{}, + newFlags: []model.Flag{}, + wantFlags: []model.Flag{}, }, { name: "empty new", @@ -63,7 +64,7 @@ func TestUpdateFlags(t *testing.T) { }, source: source1, newFlags: nil, - wantFlags: map[string]model.Flag{}, + wantFlags: []model.Flag{}, }, { name: "update from source 1 (old flag removed)", @@ -72,17 +73,17 @@ func TestUpdateFlags(t *testing.T) { if err != nil { t.Fatalf("NewStore failed: %v", err) } - s.Update(source1, map[string]model.Flag{ - "waka": {DefaultVariant: "off"}, + s.Update(source1, []model.Flag{ + {Key: "waka", DefaultVariant: "off"}, }, nil) return s }, - newFlags: map[string]model.Flag{ - "paka": {DefaultVariant: "on"}, + newFlags: []model.Flag{ + {Key: "paka", DefaultVariant: "on"}, }, source: source1, - wantFlags: map[string]model.Flag{ - "paka": {Key: "paka", DefaultVariant: "on", Source: source1, FlagSetId: nilFlagSetId, Priority: 0}, + wantFlags: []model.Flag{ + {Key: "paka", DefaultVariant: "on", Source: source1, FlagSetId: nilFlagSetId, Priority: 0}, }, }, { @@ -92,18 +93,18 @@ func TestUpdateFlags(t *testing.T) { if err != nil { t.Fatalf("NewStore failed: %v", err) } - s.Update(source1, map[string]model.Flag{ - "waka": {DefaultVariant: "off"}, + s.Update(source1, []model.Flag{ + {Key: "waka", DefaultVariant: "off"}, }, nil) return s }, - newFlags: map[string]model.Flag{ - "paka": {DefaultVariant: "on"}, + newFlags: []model.Flag{ + {Key: "paka", DefaultVariant: "on"}, }, source: source2, - wantFlags: map[string]model.Flag{ - "waka": {Key: "waka", DefaultVariant: "off", Source: source1, FlagSetId: nilFlagSetId, Priority: 0}, - "paka": {Key: "paka", DefaultVariant: "on", Source: source2, FlagSetId: nilFlagSetId, Priority: 1}, + wantFlags: []model.Flag{ + {Key: "waka", DefaultVariant: "off", Source: source1, FlagSetId: nilFlagSetId, Priority: 0}, + {Key: "paka", DefaultVariant: "on", Source: source2, FlagSetId: nilFlagSetId, Priority: 1}, }, }, { @@ -113,20 +114,20 @@ func TestUpdateFlags(t *testing.T) { if err != nil { t.Fatalf("NewStore failed: %v", err) } - s.Update(source1, map[string]model.Flag{}, model.Metadata{}) + s.Update(source1, []model.Flag{}, model.Metadata{}) return s }, setMetadata: model.Metadata{ "flagSetId": "topLevelSet", // top level set metadata, including flagSetId }, - newFlags: map[string]model.Flag{ - "waka": {DefaultVariant: "on"}, - "paka": {DefaultVariant: "on", Metadata: model.Metadata{"flagSetId": "flagLevelSet"}}, // overrides set level flagSetId + newFlags: []model.Flag{ + {Key: "waka", DefaultVariant: "on"}, + {Key: "paka", DefaultVariant: "on", Metadata: model.Metadata{"flagSetId": "flagLevelSet"}}, // overrides set level flagSetId }, source: source1, - wantFlags: map[string]model.Flag{ - "waka": {Key: "waka", DefaultVariant: "on", Source: source1, FlagSetId: "topLevelSet", Priority: 0, Metadata: model.Metadata{"flagSetId": "topLevelSet"}}, - "paka": {Key: "paka", DefaultVariant: "on", Source: source1, FlagSetId: "flagLevelSet", Priority: 0, Metadata: model.Metadata{"flagSetId": "flagLevelSet"}}, + wantFlags: []model.Flag{ + {Key: "waka", DefaultVariant: "on", Source: source1, FlagSetId: "topLevelSet", Priority: 0, Metadata: model.Metadata{"flagSetId": "topLevelSet"}}, + {Key: "paka", DefaultVariant: "on", Source: source1, FlagSetId: "flagLevelSet", Priority: 0, Metadata: model.Metadata{"flagSetId": "flagLevelSet"}}, }, }, } @@ -138,8 +139,13 @@ func TestUpdateFlags(t *testing.T) { store := tt.setup(t) store.Update(tt.source, tt.newFlags, tt.setMetadata) gotFlags, _, _ := store.GetAll(context.Background(), nil) - - require.Equal(t, tt.wantFlags, gotFlags) + sort.Slice(tt.wantFlags, func(i, j int) bool { + return tt.wantFlags[i].FlagSetId+"|"+tt.wantFlags[i].Key > tt.wantFlags[j].FlagSetId+"|"+tt.wantFlags[j].Key + }) + sort.Slice(gotFlags, func(i, j int) bool { + return gotFlags[i].FlagSetId+"|"+gotFlags[i].Key > gotFlags[j].FlagSetId+"|"+gotFlags[j].Key + }) + require.EqualValues(t, tt.wantFlags, gotFlags) }) } } @@ -206,16 +212,16 @@ func TestGet(t *testing.T) { t.Run(tt.name, func(t *testing.T) { t.Parallel() - sourceAFlags := map[string]model.Flag{ - "flagA": {Key: "flagA", DefaultVariant: "off"}, - "dupe": {Key: "dupe", DefaultVariant: "on"}, + sourceAFlags := []model.Flag{ + {Key: "flagA", DefaultVariant: "off"}, + {Key: "dupe", DefaultVariant: "on"}, } - sourceBFlags := map[string]model.Flag{ - "flagB": {Key: "flagB", DefaultVariant: "off", Metadata: model.Metadata{"flagSetId": flagSetIdB}}, + sourceBFlags := []model.Flag{ + {Key: "flagB", DefaultVariant: "off", Metadata: model.Metadata{"flagSetId": flagSetIdB}}, } - sourceCFlags := map[string]model.Flag{ - "flagC": {Key: "flagC", DefaultVariant: "off", Metadata: model.Metadata{"flagSetId": flagSetIdC}}, - "dupe": {Key: "dupe", DefaultVariant: "off", Metadata: model.Metadata{"flagSetId": flagSetIdC}}, + sourceCFlags := []model.Flag{ + {Key: "flagC", DefaultVariant: "off", Metadata: model.Metadata{"flagSetId": flagSetIdC}}, + {Key: "dupe", DefaultVariant: "off", Metadata: model.Metadata{"flagSetId": flagSetIdC}}, } store, err := NewStore(logger.NewLogger(nil, false), sources) @@ -253,35 +259,36 @@ func TestGetAllNoWatcher(t *testing.T) { tests := []struct { name string selector *Selector - wantFlags map[string]model.Flag + wantFlags []model.Flag }{ { name: "nil selector", selector: nil, - wantFlags: map[string]model.Flag{ + wantFlags: []model.Flag{ // "dupe" should be overwritten by higher priority flag - "flagA": {Key: "flagA", DefaultVariant: "off", Source: sourceA, FlagSetId: nilFlagSetId, Priority: 0}, - "flagB": {Key: "flagB", DefaultVariant: "off", Source: sourceB, FlagSetId: flagSetIdB, Priority: 1, Metadata: model.Metadata{"flagSetId": flagSetIdB}}, - "flagC": {Key: "flagC", DefaultVariant: "off", Source: sourceC, FlagSetId: flagSetIdC, Priority: 2, Metadata: model.Metadata{"flagSetId": flagSetIdC}}, - "dupe": {Key: "dupe", DefaultVariant: "off", Source: sourceC, FlagSetId: flagSetIdC, Priority: 2, Metadata: model.Metadata{"flagSetId": flagSetIdC}}, + {Key: "dupe", DefaultVariant: "on", Source: sourceA, FlagSetId: nilFlagSetId, Priority: 0}, + {Key: "flagA", DefaultVariant: "off", Source: sourceA, FlagSetId: nilFlagSetId, Priority: 0}, + {Key: "flagB", DefaultVariant: "off", Source: sourceB, FlagSetId: flagSetIdB, Priority: 1, Metadata: model.Metadata{"flagSetId": flagSetIdB}}, + {Key: "flagC", DefaultVariant: "off", Source: sourceC, FlagSetId: flagSetIdC, Priority: 2, Metadata: model.Metadata{"flagSetId": flagSetIdC}}, + {Key: "dupe", DefaultVariant: "off", Source: sourceC, FlagSetId: flagSetIdC, Priority: 2, Metadata: model.Metadata{"flagSetId": flagSetIdC}}, }, }, { name: "source selector", selector: &sourceASelector, - wantFlags: map[string]model.Flag{ + wantFlags: []model.Flag{ // we should get the "dupe" from sourceA - "flagA": {Key: "flagA", DefaultVariant: "off", Source: sourceA, FlagSetId: nilFlagSetId, Priority: 0}, - "dupe": {Key: "dupe", DefaultVariant: "on", Source: sourceA, FlagSetId: nilFlagSetId, Priority: 0}, + {Key: "flagA", DefaultVariant: "off", Source: sourceA, FlagSetId: nilFlagSetId, Priority: 0}, + {Key: "dupe", DefaultVariant: "on", Source: sourceA, FlagSetId: nilFlagSetId, Priority: 0}, }, }, { name: "flagSetId selector", selector: &flagSetIdCSelector, - wantFlags: map[string]model.Flag{ + wantFlags: []model.Flag{ // we should get the "dupe" from flagSetIdC - "flagC": {Key: "flagC", DefaultVariant: "off", Source: sourceC, FlagSetId: flagSetIdC, Priority: 2, Metadata: model.Metadata{"flagSetId": flagSetIdC}}, - "dupe": {Key: "dupe", DefaultVariant: "off", Source: sourceC, FlagSetId: flagSetIdC, Priority: 2, Metadata: model.Metadata{"flagSetId": flagSetIdC}}, + {Key: "flagC", DefaultVariant: "off", Source: sourceC, FlagSetId: flagSetIdC, Priority: 2, Metadata: model.Metadata{"flagSetId": flagSetIdC}}, + {Key: "dupe", DefaultVariant: "off", Source: sourceC, FlagSetId: flagSetIdC, Priority: 2, Metadata: model.Metadata{"flagSetId": flagSetIdC}}, }, }, } @@ -291,16 +298,16 @@ func TestGetAllNoWatcher(t *testing.T) { t.Run(tt.name, func(t *testing.T) { t.Parallel() - sourceAFlags := map[string]model.Flag{ - "flagA": {Key: "flagA", DefaultVariant: "off"}, - "dupe": {Key: "dupe", DefaultVariant: "on"}, + sourceAFlags := []model.Flag{ + {Key: "flagA", DefaultVariant: "off"}, + {Key: "dupe", DefaultVariant: "on"}, } - sourceBFlags := map[string]model.Flag{ - "flagB": {Key: "flagB", DefaultVariant: "off", Metadata: model.Metadata{"flagSetId": flagSetIdB}}, + sourceBFlags := []model.Flag{ + {Key: "flagB", DefaultVariant: "off", Metadata: model.Metadata{"flagSetId": flagSetIdB}}, } - sourceCFlags := map[string]model.Flag{ - "flagC": {Key: "flagC", DefaultVariant: "off", Metadata: model.Metadata{"flagSetId": flagSetIdC}}, - "dupe": {Key: "dupe", DefaultVariant: "off", Metadata: model.Metadata{"flagSetId": flagSetIdC}}, + sourceCFlags := []model.Flag{ + {Key: "flagC", DefaultVariant: "off", Metadata: model.Metadata{"flagSetId": flagSetIdC}}, + {Key: "dupe", DefaultVariant: "off", Metadata: model.Metadata{"flagSetId": flagSetIdC}}, } store, err := NewStore(logger.NewLogger(nil, false), sources) @@ -314,6 +321,12 @@ func TestGetAllNoWatcher(t *testing.T) { gotFlags, _, _ := store.GetAll(context.Background(), tt.selector) require.Equal(t, len(tt.wantFlags), len(gotFlags)) + sort.Slice(tt.wantFlags, func(i, j int) bool { + return tt.wantFlags[i].FlagSetId+"|"+tt.wantFlags[i].Key > tt.wantFlags[j].FlagSetId+"|"+tt.wantFlags[j].Key + }) + sort.Slice(gotFlags, func(i, j int) bool { + return gotFlags[i].FlagSetId+"|"+gotFlags[i].Key > gotFlags[j].FlagSetId+"|"+gotFlags[j].Key + }) require.Equal(t, tt.wantFlags, gotFlags) }) } @@ -365,14 +378,12 @@ func TestWatch(t *testing.T) { t.Run(tt.name, func(t *testing.T) { t.Parallel() - sourceAFlags := map[string]model.Flag{ - "flagA": {Key: "flagA", DefaultVariant: "off"}, + sourceAFlags := []model.Flag{ + {Key: "flagA", DefaultVariant: "off"}, } - sourceBFlags := map[string]model.Flag{ - "flagB": {Key: "flagB", DefaultVariant: "off", Metadata: model.Metadata{"flagSetId": myFlagSetId}}, - } - sourceCFlags := map[string]model.Flag{ - "flagC": {Key: "flagC", DefaultVariant: "off"}, + sourceBFlags := []model.Flag{{Key: "flagB", DefaultVariant: "off", Metadata: model.Metadata{"flagSetId": myFlagSetId}}} + sourceCFlags := []model.Flag{ + {Key: "flagC", DefaultVariant: "off"}, } store, err := NewStore(logger.NewLogger(nil, false), sources) @@ -396,29 +407,30 @@ func TestWatch(t *testing.T) { time.Sleep(pauseTime) // changing a flag default variant should trigger an update - store.Update(sourceA, map[string]model.Flag{ - "flagA": {Key: "flagA", DefaultVariant: "on"}, + store.Update(sourceA, []model.Flag{ + {Key: "flagA", DefaultVariant: "on"}, }, model.Metadata{}) time.Sleep(pauseTime) // changing a flag default variant should trigger an update - store.Update(sourceB, map[string]model.Flag{ - "flagB": {Key: "flagB", DefaultVariant: "on", Metadata: model.Metadata{"flagSetId": myFlagSetId}}, + store.Update(sourceB, []model.Flag{ + {Key: "flagB", DefaultVariant: "on", Metadata: model.Metadata{"flagSetId": myFlagSetId}}, }, model.Metadata{}) time.Sleep(pauseTime) // removing a flag set id should trigger an update (even for flag set id selectors; it should remove the flag from the set) - store.Update(sourceB, map[string]model.Flag{ - "flagB": {Key: "flagB", DefaultVariant: "on"}, + // TODO: challenge this test and behaviour + store.Update(sourceB, []model.Flag{ + {Key: "flagB", DefaultVariant: "on"}, }, model.Metadata{}) time.Sleep(pauseTime) // adding a flag set id should trigger an update - store.Update(sourceB, map[string]model.Flag{ - "flagB": {Key: "flagB", DefaultVariant: "on", Metadata: model.Metadata{"flagSetId": myFlagSetId}}, + store.Update(sourceB, []model.Flag{ + {Key: "flagB", DefaultVariant: "on", Metadata: model.Metadata{"flagSetId": myFlagSetId}}, }, model.Metadata{}) }() @@ -448,9 +460,9 @@ func TestQueryMetadata(t *testing.T) { otherSource := "otherSource" nonExistingFlagSetId := "nonExistingFlagSetId" var sources = []string{sourceA} - sourceAFlags := map[string]model.Flag{ - "flagA": {Key: "flagA", DefaultVariant: "off"}, - "flagB": {Key: "flagB", DefaultVariant: "on"}, + sourceAFlags := []model.Flag{ + {Key: "flagA", DefaultVariant: "off"}, + {Key: "flagB", DefaultVariant: "on"}, } store, err := NewStore(logger.NewLogger(nil, false), sources)

Signed-off-by: Todd Baert <todd.baert@dynatrace.com>

toddbaert · 2025-10-08T15:45:35Z

Note this is only a change in the core lib, not flagd itself, so I will modify the flagd release accordingly when we release this.

Tested this manually as well, merging now.

@toddbaert

…ite (#1827) with #1797 we introduced this bug, that the format of the response is not correct. current state: ```javacript { flagConfiguration: {/* flag object */} } ``` should be: ```javacript { flagConfiguration: { flags: {/* flag object */} } } ``` Clarification from @toddbaert - this is an UNRELEASED bug so far. --------- Signed-off-by: Simon Schrottner <simon.schrottner@dynatrace.com> Signed-off-by: Todd Baert <todd.baert@dynatrace.com> Co-authored-by: Todd Baert <todd.baert@dynatrace.com> Co-authored-by: alexandraoberaigner <82218944+alexandraoberaigner@users.noreply.github.com>

🤖 I have created a release *beep* *boop* --- <details><summary>flagd: 0.13.0</summary> ## [0.13.0](flagd/v0.12.9...flagd/v0.13.0) (2025-12-23) ### 🐛 Bug Fixes * fixing sync return format missing flag layer, adding full e2e suite ([#1827](#1827)) ([570693d](570693d)) * **security:** update module github.com/go-viper/mapstructure/v2 to v2.4.0 [security] ([#1784](#1784)) ([037e30b](037e30b)) * **security:** update module golang.org/x/crypto to v0.45.0 [security] ([#1826](#1826)) ([7e0762b](7e0762b)) ### ✨ New Features * add support for http-based ofrep metrics ([#1803](#1803)) ([fcd19b3](fcd19b3)) * cleanup evaluator interface ([#1793](#1793)) ([aa504f7](aa504f7)) * enable parsing of array flag configurations for flagd ([#1797](#1797)) ([97c6ffa](97c6ffa)) * multi-project support via selectors and flagSetId namespacing ([#1702](#1702)) ([f9ce46f](f9ce46f)) * normalize selector in sync (use header as in OFREP and RPC) ([#1815](#1815)) ([c1f06cb](c1f06cb)) ### 🧹 Chore * **refactor:** use memdb for flag storage ([#1697](#1697)) ([5c5c1cf](5c5c1cf)) ### 🔄 Refactoring * store cleanup ([#1705](#1705)) ([bcff8d7](bcff8d7)) </details> <details><summary>flagd-proxy: 0.8.1</summary> ## [0.8.1](flagd-proxy/v0.8.0...flagd-proxy/v0.8.1) (2025-12-23) ### 🐛 Bug Fixes * **security:** update module github.com/go-viper/mapstructure/v2 to v2.4.0 [security] ([#1784](#1784)) ([037e30b](037e30b)) * **security:** update module golang.org/x/crypto to v0.45.0 [security] ([#1826](#1826)) ([7e0762b](7e0762b)) </details> <details><summary>core: 0.13.0</summary> ## [0.13.0](core/v0.12.1...core/v0.13.0) (2025-12-23) ### ⚠ BREAKING CHANGES * enable parsing of array flag configurations for flagd ([#1797](#1797)) * cleanup evaluator interface ([#1793](#1793)) * removes the `fractionalEvaluation` operator since it has been replaced with `fractional`. ([#1704](#1704)) ### 🐛 Bug Fixes * **security:** update module github.com/go-viper/mapstructure/v2 to v2.4.0 [security] ([#1784](#1784)) ([037e30b](037e30b)) * **security:** update module golang.org/x/crypto to v0.45.0 [security] ([#1825](#1825)) ([44edcc9](44edcc9)) * **security:** update module golang.org/x/crypto to v0.45.0 [security] ([#1826](#1826)) ([7e0762b](7e0762b)) ### ✨ New Features * Add OAuth support for HTTP Sync ([#1791](#1791)) ([268fd75](268fd75)) * Add OTEL default variables ([#1812](#1812)) ([c2e3fc6](c2e3fc6)) * allow null flagSetId Selector, restrict Selector to single key-value-pairs ([#1708](#1708)) ([#1811](#1811)) ([c12a0ae](c12a0ae)) * change jsonschema parser ([#1794](#1794)) ([bf3f722](bf3f722)) * cleanup evaluator interface ([#1793](#1793)) ([aa504f7](aa504f7)) * enable parsing of array flag configurations for flagd ([#1797](#1797)) ([97c6ffa](97c6ffa)) * multi-project support via selectors and flagSetId namespacing ([#1702](#1702)) ([f9ce46f](f9ce46f)) ### 🧹 Chore * **refactor:** use memdb for flag storage ([#1697](#1697)) ([5c5c1cf](5c5c1cf)) * removes the `fractionalEvaluation` operator since it has been replaced with `fractional`. ([#1704](#1704)) ([3228ad8](3228ad8)) ### 🔄 Refactoring * remove deprecated bearerToken option ([#1816](#1816)) ([efda06a](efda06a)) * removed unused Selector from Flag and Store. ([#1747](#1747)) ([1083005](1083005)) * store cleanup ([#1705](#1705)) ([bcff8d7](bcff8d7)) </details> --- This PR was generated with [Release Please](https://github.com/googleapis/release-please). See [documentation](https://github.com/googleapis/release-please#release-please). --------- Signed-off-by: OpenFeature Bot <109696520+openfeaturebot@users.noreply.github.com> Signed-off-by: Todd Baert <todd.baert@dynatrace.com> Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> Co-authored-by: Todd Baert <todd.baert@dynatrace.com>

Removes the State Resync Events section from syncs.md as it describes behavior that flagd does not actually support. Issue open-feature#1797 clarified that flagd does not handle cases where the same flag key is defined across multiple sources. The removed section incorrectly implied that a resync event would restore a flag definition from a lower-priority source when a higher-priority source deleted its version of the flag. Closes open-feature#1809

aepfli requested review from a team as code owners September 8, 2025 14:45

dosubot Bot added the size:XXL This PR changes 1000+ lines, ignoring generated files. label Sep 8, 2025

gemini-code-assist Bot reviewed Sep 8, 2025

View reviewed changes

aepfli force-pushed the refactor/change_store_interface branch from 6049dad to d805690 Compare September 8, 2025 14:47

dosubot Bot added size:XL This PR changes 500-999 lines, ignoring generated files. and removed size:XXL This PR changes 1000+ lines, ignoring generated files. labels Sep 8, 2025

gemini-code-assist Bot reviewed Sep 8, 2025

View reviewed changes

Comment thread core/pkg/store/store.go Outdated

Comment thread core/pkg/store/store.go

aepfli force-pushed the refactor/change_store_interface branch 2 times, most recently from f754958 to b87ddac Compare September 8, 2025 15:19

gemini-code-assist Bot reviewed Sep 8, 2025

View reviewed changes

Comment thread flagd/pkg/service/flag-evaluation/eventing.go

Comment thread core/pkg/store/store_test.go

aepfli changed the title ~~feat: refactor store interface to use a slice instead of a map~~ feat!: refactor store interface to use a slice instead of a map Sep 8, 2025

aepfli changed the title ~~feat!: refactor store interface to use a slice instead of a map~~ feat!: Enable parsing of array flag configurations for flagd Sep 9, 2025

dosubot Bot added size:XXL This PR changes 1000+ lines, ignoring generated files. and removed size:XL This PR changes 500-999 lines, ignoring generated files. labels Sep 9, 2025

aepfli mentioned this pull request Sep 9, 2025

feat: add flagmap config support #1790

Closed

toddbaert reviewed Sep 9, 2025

View reviewed changes

Comment thread flagd/pkg/service/flag-evaluation/eventing.go

toddbaert requested review from beeme1mr, chrfwow, tangenti and toddbaert September 9, 2025 23:44

toddbaert approved these changes Sep 9, 2025

View reviewed changes

chrfwow reviewed Sep 10, 2025

View reviewed changes

Comment thread core/pkg/evaluator/json.go

Comment thread core/pkg/store/store.go Outdated

tangenti requested changes Sep 10, 2025

View reviewed changes

aepfli commented Sep 11, 2025

View reviewed changes

Comment thread core/pkg/store/store.go

aepfli force-pushed the refactor/change_store_interface branch from 5f712a9 to ce81476 Compare September 29, 2025 18:15

gemini-code-assist Bot reviewed Sep 29, 2025

View reviewed changes

Comment thread core/pkg/store/store.go Outdated

Comment thread flagd/pkg/service/flag-evaluation/eventing.go

Comment thread flagd/pkg/service/flag-sync/handler.go

aepfli marked this pull request as draft September 29, 2025 18:49

aepfli force-pushed the refactor/change_store_interface branch from 0b89a95 to 8cded80 Compare September 29, 2025 19:16

aepfli marked this pull request as ready for review September 29, 2025 19:20

aepfli requested review from chrfwow, tangenti and toddbaert September 30, 2025 06:39

toddbaert force-pushed the refactor/change_store_interface branch from 8cded80 to 54d118b Compare October 3, 2025 19:32

toddbaert changed the title ~~feat!: Enable parsing of array flag configurations for flagd~~ feat!: enable parsing of array flag configurations for flagd Oct 3, 2025

toddbaert force-pushed the refactor/change_store_interface branch from 54d118b to a47ddd6 Compare October 3, 2025 19:33

toddbaert approved these changes Oct 3, 2025

View reviewed changes

aepfli and others added 2 commits October 3, 2025 15:38

fixup: test race, un-needed struct, submodule tag

b646be0

Signed-off-by: Todd Baert <todd.baert@dynatrace.com>

toddbaert force-pushed the refactor/change_store_interface branch from a47ddd6 to b646be0 Compare October 3, 2025 19:38

toddbaert merged commit 97c6ffa into open-feature:main Oct 8, 2025
15 checks passed

github-actions Bot mentioned this pull request Oct 6, 2025

chore: release main #1699

Merged

This was referenced Oct 9, 2025

[DOC] Remove outdated Resync Event section for Sync sources #1809

Open

[ADR] Create ADR to Address Handling of Duplicate Flagset IDs Across Multiple Sources in flagd #1810

Open

aepfli mentioned this pull request Nov 27, 2025

fix: fixing sync return format missing flag layer, adding full e2e suite #1827

Merged

kunalworldwide mentioned this pull request May 27, 2026

docs: remove outdated State Resync Event section #1970

Open

Conversation

aepfli commented Sep 8, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Highlights

Uh oh!

netlify Bot commented Sep 8, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

✅ Deploy Preview for polite-licorice-3db33c canceled.

Uh oh!

gemini-code-assist Bot left a comment

Choose a reason for hiding this comment

Summary of Changes

Highlights

Footnotes

Uh oh!

gemini-code-assist Bot left a comment

Choose a reason for hiding this comment

Code Review

Uh oh!

Uh oh!

Uh oh!

aepfli commented Sep 8, 2025

Uh oh!

gemini-code-assist Bot left a comment

Choose a reason for hiding this comment

Code Review

Uh oh!

Uh oh!

Uh oh!

Uh oh!

toddbaert left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

tangenti left a comment

Choose a reason for hiding this comment

Uh oh!

tangenti Sep 10, 2025

Choose a reason for hiding this comment

Uh oh!

aepfli Sep 10, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

toddbaert Sep 10, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

tangenti Sep 10, 2025

Choose a reason for hiding this comment

Uh oh!

aepfli Sep 11, 2025

Choose a reason for hiding this comment

Uh oh!

tangenti Sep 29, 2025

Choose a reason for hiding this comment

Uh oh!

aepfli Sep 29, 2025

Choose a reason for hiding this comment

Uh oh!

toddbaert Sep 29, 2025

Choose a reason for hiding this comment

Uh oh!

beeme1mr Sep 29, 2025

Choose a reason for hiding this comment

Uh oh!

aepfli Sep 29, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

toddbaert commented Sep 10, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

aepfli commented Sep 29, 2025

Uh oh!

gemini-code-assist Bot left a comment

Choose a reason for hiding this comment

aepfli commented Sep 8, 2025 •

edited

Loading

netlify Bot commented Sep 8, 2025 •

edited

Loading

aepfli Sep 10, 2025 •

edited

Loading

toddbaert Sep 10, 2025 •

edited

Loading

aepfli Sep 29, 2025 •

edited

Loading

toddbaert commented Sep 10, 2025 •

edited

Loading

toddbaert left a comment •

edited

Loading