Upgrade CEL from 0.10.1 to 0.11.0 by pkwarren · Pull Request #345 · bufbuild/protovalidate-java

pkwarren · 2025-09-08T19:23:30Z

Update to the latest CEL version, which required changing to use internal CEL null/bytes types to fix an overloaded method and the conformance tests.

While evaluateCanonicalTypesToNativeValues isn't the default yet, according to the release notes it will be the default soon so switched everything to use it.

Update to the latest CEL version, which required changing to use internal CEL null/bytes types to fix an overloaded method and the conformance tests. While `evaluateCanonicalTypesToNativeValues` isn't the default yet, according to the release notes it will be the default soon so switched everything to use it.

pkwarren · 2025-09-08T19:23:48Z

 api(libs.protobuf.java)
- implementation(libs.cel) {
- // https://github.com/google/cel-java/issues/748
- exclude(group = "com.google.protobuf", module = "protobuf-javalite")


Upstream issue is fixed in this release.

pkwarren · 2025-09-08T19:24:19Z

 BytesIn invalid = BytesIn.newBuilder().setVal(ByteString.copyFromUtf8("bar")).build();
 ValidationResult validate = validator.validate(invalid);
- assertThat(validate.isSuccess()).isTrue();
+ assertThat(validate.getViolations()).isEmpty();


Just showing expected true, got false isn't great for debugging. Made it use isEmpty so it will print the violations.

pkwarren · 2025-09-08T19:25:07Z

 }
- for (int i = 0; i < param.size(); i++) {
- if (param.byteAt(i) != receiver.byteAt(i)) {
+ byte[] paramBytes = param.toByteArray();


It would be nice if there was a byteAt method on CelByteString to avoid having to allocate a byte array to access a single byte.

pkwarren · 2025-09-08T19:30:11Z

 .addCompilerLibraries(validateLibrary)
 .addRuntimeLibraries(validateLibrary)
+ .setOptions(
+ CelOptions.DEFAULT.toBuilder().evaluateCanonicalTypesToNativeValues(true).build())


This opts in to the new behavior which uses CEL types instead of Protobuf types for NullValue / ByteString. This is apparently coming soon as the default, so since we had to update our overloads to use the new types just opted in to migrate everything to use them.

pkwarren · 2025-09-08T19:31:00Z

+
+ @Override
+ public String toString() {
+ return proto.toString();


This made debugging more painful - added a toString which will dump the contents of the proto Validation message.

pkwarren · 2025-09-08T19:23:30Z

Update to the latest CEL version, which required changing to use internal CEL null/bytes types to fix an overloaded method and the conformance tests.

While evaluateCanonicalTypesToNativeValues isn't the default yet, according to the release notes it will be the default soon so switched everything to use it.

Update to the latest CEL version, which required changing to use internal CEL null/bytes types to fix an overloaded method and the conformance tests. While `evaluateCanonicalTypesToNativeValues` isn't the default yet, according to the release notes it will be the default soon so switched everything to use it.

pkwarren · 2025-09-08T19:23:48Z

 api(libs.protobuf.java)
- implementation(libs.cel) {
- // https://github.com/google/cel-java/issues/748
- exclude(group = "com.google.protobuf", module = "protobuf-javalite")


Upstream issue is fixed in this release.

pkwarren · 2025-09-08T19:24:19Z

 BytesIn invalid = BytesIn.newBuilder().setVal(ByteString.copyFromUtf8("bar")).build();
 ValidationResult validate = validator.validate(invalid);
- assertThat(validate.isSuccess()).isTrue();
+ assertThat(validate.getViolations()).isEmpty();


Just showing expected true, got false isn't great for debugging. Made it use isEmpty so it will print the violations.

pkwarren · 2025-09-08T19:25:07Z

 }
- for (int i = 0; i < param.size(); i++) {
- if (param.byteAt(i) != receiver.byteAt(i)) {
+ byte[] paramBytes = param.toByteArray();


It would be nice if there was a byteAt method on CelByteString to avoid having to allocate a byte array to access a single byte.

pkwarren · 2025-09-08T19:30:11Z

 .addCompilerLibraries(validateLibrary)
 .addRuntimeLibraries(validateLibrary)
+ .setOptions(
+ CelOptions.DEFAULT.toBuilder().evaluateCanonicalTypesToNativeValues(true).build())


This opts in to the new behavior which uses CEL types instead of Protobuf types for NullValue / ByteString. This is apparently coming soon as the default, so since we had to update our overloads to use the new types just opted in to migrate everything to use them.

pkwarren · 2025-09-08T19:31:00Z

+
+ @Override
+ public String toString() {
+ return proto.toString();


This made debugging more painful - added a toString which will dump the contents of the proto Validation message.

pkwarren · 2025-09-08T19:23:30Z

Update to the latest CEL version, which required changing to use internal CEL null/bytes types to fix an overloaded method and the conformance tests.

While evaluateCanonicalTypesToNativeValues isn't the default yet, according to the release notes it will be the default soon so switched everything to use it.

Update to the latest CEL version, which required changing to use internal CEL null/bytes types to fix an overloaded method and the conformance tests. While `evaluateCanonicalTypesToNativeValues` isn't the default yet, according to the release notes it will be the default soon so switched everything to use it.

pkwarren · 2025-09-08T19:23:48Z

 api(libs.protobuf.java)
- implementation(libs.cel) {
- // https://github.com/google/cel-java/issues/748
- exclude(group = "com.google.protobuf", module = "protobuf-javalite")


Upstream issue is fixed in this release.

pkwarren · 2025-09-08T19:24:19Z

 BytesIn invalid = BytesIn.newBuilder().setVal(ByteString.copyFromUtf8("bar")).build();
 ValidationResult validate = validator.validate(invalid);
- assertThat(validate.isSuccess()).isTrue();
+ assertThat(validate.getViolations()).isEmpty();


Just showing expected true, got false isn't great for debugging. Made it use isEmpty so it will print the violations.

pkwarren · 2025-09-08T19:25:07Z

 }
- for (int i = 0; i < param.size(); i++) {
- if (param.byteAt(i) != receiver.byteAt(i)) {
+ byte[] paramBytes = param.toByteArray();


It would be nice if there was a byteAt method on CelByteString to avoid having to allocate a byte array to access a single byte.

pkwarren · 2025-09-08T19:30:11Z

 .addCompilerLibraries(validateLibrary)
 .addRuntimeLibraries(validateLibrary)
+ .setOptions(
+ CelOptions.DEFAULT.toBuilder().evaluateCanonicalTypesToNativeValues(true).build())


This opts in to the new behavior which uses CEL types instead of Protobuf types for NullValue / ByteString. This is apparently coming soon as the default, so since we had to update our overloads to use the new types just opted in to migrate everything to use them.

pkwarren · 2025-09-08T19:31:00Z

+
+ @Override
+ public String toString() {
+ return proto.toString();


This made debugging more painful - added a toString which will dump the contents of the proto Validation message.

pkwarren · 2025-09-08T19:23:30Z

Update to the latest CEL version, which required changing to use internal CEL null/bytes types to fix an overloaded method and the conformance tests.

While evaluateCanonicalTypesToNativeValues isn't the default yet, according to the release notes it will be the default soon so switched everything to use it.

Update to the latest CEL version, which required changing to use internal CEL null/bytes types to fix an overloaded method and the conformance tests. While `evaluateCanonicalTypesToNativeValues` isn't the default yet, according to the release notes it will be the default soon so switched everything to use it.

pkwarren · 2025-09-08T19:23:48Z

 api(libs.protobuf.java)
- implementation(libs.cel) {
- // https://github.com/google/cel-java/issues/748
- exclude(group = "com.google.protobuf", module = "protobuf-javalite")


Upstream issue is fixed in this release.

pkwarren · 2025-09-08T19:24:19Z

 BytesIn invalid = BytesIn.newBuilder().setVal(ByteString.copyFromUtf8("bar")).build();
 ValidationResult validate = validator.validate(invalid);
- assertThat(validate.isSuccess()).isTrue();
+ assertThat(validate.getViolations()).isEmpty();


Just showing expected true, got false isn't great for debugging. Made it use isEmpty so it will print the violations.

pkwarren · 2025-09-08T19:25:07Z

 }
- for (int i = 0; i < param.size(); i++) {
- if (param.byteAt(i) != receiver.byteAt(i)) {
+ byte[] paramBytes = param.toByteArray();


It would be nice if there was a byteAt method on CelByteString to avoid having to allocate a byte array to access a single byte.

pkwarren · 2025-09-08T19:30:11Z

 .addCompilerLibraries(validateLibrary)
 .addRuntimeLibraries(validateLibrary)
+ .setOptions(
+ CelOptions.DEFAULT.toBuilder().evaluateCanonicalTypesToNativeValues(true).build())


This opts in to the new behavior which uses CEL types instead of Protobuf types for NullValue / ByteString. This is apparently coming soon as the default, so since we had to update our overloads to use the new types just opted in to migrate everything to use them.

pkwarren · 2025-09-08T19:31:00Z

+
+ @Override
+ public String toString() {
+ return proto.toString();


This made debugging more painful - added a toString which will dump the contents of the proto Validation message.

pkwarren · 2025-09-08T19:23:30Z

Update to the latest CEL version, which required changing to use internal CEL null/bytes types to fix an overloaded method and the conformance tests.

While evaluateCanonicalTypesToNativeValues isn't the default yet, according to the release notes it will be the default soon so switched everything to use it.

Update to the latest CEL version, which required changing to use internal CEL null/bytes types to fix an overloaded method and the conformance tests. While `evaluateCanonicalTypesToNativeValues` isn't the default yet, according to the release notes it will be the default soon so switched everything to use it.

pkwarren · 2025-09-08T19:23:48Z

 api(libs.protobuf.java)
- implementation(libs.cel) {
- // https://github.com/google/cel-java/issues/748
- exclude(group = "com.google.protobuf", module = "protobuf-javalite")


Upstream issue is fixed in this release.

pkwarren · 2025-09-08T19:24:19Z

 BytesIn invalid = BytesIn.newBuilder().setVal(ByteString.copyFromUtf8("bar")).build();
 ValidationResult validate = validator.validate(invalid);
- assertThat(validate.isSuccess()).isTrue();
+ assertThat(validate.getViolations()).isEmpty();


Just showing expected true, got false isn't great for debugging. Made it use isEmpty so it will print the violations.

pkwarren · 2025-09-08T19:25:07Z

 }
- for (int i = 0; i < param.size(); i++) {
- if (param.byteAt(i) != receiver.byteAt(i)) {
+ byte[] paramBytes = param.toByteArray();


It would be nice if there was a byteAt method on CelByteString to avoid having to allocate a byte array to access a single byte.

pkwarren · 2025-09-08T19:30:11Z

 .addCompilerLibraries(validateLibrary)
 .addRuntimeLibraries(validateLibrary)
+ .setOptions(
+ CelOptions.DEFAULT.toBuilder().evaluateCanonicalTypesToNativeValues(true).build())


This opts in to the new behavior which uses CEL types instead of Protobuf types for NullValue / ByteString. This is apparently coming soon as the default, so since we had to update our overloads to use the new types just opted in to migrate everything to use them.

pkwarren · 2025-09-08T19:31:00Z

+
+ @Override
+ public String toString() {
+ return proto.toString();


This made debugging more painful - added a toString which will dump the contents of the proto Validation message.

pkwarren · 2025-09-08T19:23:30Z

Update to the latest CEL version, which required changing to use internal CEL null/bytes types to fix an overloaded method and the conformance tests.

While evaluateCanonicalTypesToNativeValues isn't the default yet, according to the release notes it will be the default soon so switched everything to use it.

Update to the latest CEL version, which required changing to use internal CEL null/bytes types to fix an overloaded method and the conformance tests. While `evaluateCanonicalTypesToNativeValues` isn't the default yet, according to the release notes it will be the default soon so switched everything to use it.

pkwarren · 2025-09-08T19:23:48Z

 api(libs.protobuf.java)
- implementation(libs.cel) {
- // https://github.com/google/cel-java/issues/748
- exclude(group = "com.google.protobuf", module = "protobuf-javalite")


Upstream issue is fixed in this release.

pkwarren · 2025-09-08T19:24:19Z

 BytesIn invalid = BytesIn.newBuilder().setVal(ByteString.copyFromUtf8("bar")).build();
 ValidationResult validate = validator.validate(invalid);
- assertThat(validate.isSuccess()).isTrue();
+ assertThat(validate.getViolations()).isEmpty();


Just showing expected true, got false isn't great for debugging. Made it use isEmpty so it will print the violations.

pkwarren · 2025-09-08T19:25:07Z

 }
- for (int i = 0; i < param.size(); i++) {
- if (param.byteAt(i) != receiver.byteAt(i)) {
+ byte[] paramBytes = param.toByteArray();


It would be nice if there was a byteAt method on CelByteString to avoid having to allocate a byte array to access a single byte.

pkwarren · 2025-09-08T19:30:11Z

 .addCompilerLibraries(validateLibrary)
 .addRuntimeLibraries(validateLibrary)
+ .setOptions(
+ CelOptions.DEFAULT.toBuilder().evaluateCanonicalTypesToNativeValues(true).build())


This opts in to the new behavior which uses CEL types instead of Protobuf types for NullValue / ByteString. This is apparently coming soon as the default, so since we had to update our overloads to use the new types just opted in to migrate everything to use them.

pkwarren · 2025-09-08T19:31:00Z

+
+ @Override
+ public String toString() {
+ return proto.toString();


This made debugging more painful - added a toString which will dump the contents of the proto Validation message.

pkwarren · 2025-09-08T19:23:30Z

Update to the latest CEL version, which required changing to use internal CEL null/bytes types to fix an overloaded method and the conformance tests.

While evaluateCanonicalTypesToNativeValues isn't the default yet, according to the release notes it will be the default soon so switched everything to use it.

Update to the latest CEL version, which required changing to use internal CEL null/bytes types to fix an overloaded method and the conformance tests. While `evaluateCanonicalTypesToNativeValues` isn't the default yet, according to the release notes it will be the default soon so switched everything to use it.

pkwarren · 2025-09-08T19:23:48Z

 api(libs.protobuf.java)
- implementation(libs.cel) {
- // https://github.com/google/cel-java/issues/748
- exclude(group = "com.google.protobuf", module = "protobuf-javalite")


Upstream issue is fixed in this release.

pkwarren · 2025-09-08T19:24:19Z

 BytesIn invalid = BytesIn.newBuilder().setVal(ByteString.copyFromUtf8("bar")).build();
 ValidationResult validate = validator.validate(invalid);
- assertThat(validate.isSuccess()).isTrue();
+ assertThat(validate.getViolations()).isEmpty();


Just showing expected true, got false isn't great for debugging. Made it use isEmpty so it will print the violations.

pkwarren · 2025-09-08T19:25:07Z

 }
- for (int i = 0; i < param.size(); i++) {
- if (param.byteAt(i) != receiver.byteAt(i)) {
+ byte[] paramBytes = param.toByteArray();


It would be nice if there was a byteAt method on CelByteString to avoid having to allocate a byte array to access a single byte.

pkwarren · 2025-09-08T19:30:11Z

 .addCompilerLibraries(validateLibrary)
 .addRuntimeLibraries(validateLibrary)
+ .setOptions(
+ CelOptions.DEFAULT.toBuilder().evaluateCanonicalTypesToNativeValues(true).build())


This opts in to the new behavior which uses CEL types instead of Protobuf types for NullValue / ByteString. This is apparently coming soon as the default, so since we had to update our overloads to use the new types just opted in to migrate everything to use them.

pkwarren · 2025-09-08T19:31:00Z

+
+ @Override
+ public String toString() {
+ return proto.toString();


This made debugging more painful - added a toString which will dump the contents of the proto Validation message.

pkwarren · 2025-09-08T19:23:30Z

Update to the latest CEL version, which required changing to use internal CEL null/bytes types to fix an overloaded method and the conformance tests.

While evaluateCanonicalTypesToNativeValues isn't the default yet, according to the release notes it will be the default soon so switched everything to use it.

Update to the latest CEL version, which required changing to use internal CEL null/bytes types to fix an overloaded method and the conformance tests. While `evaluateCanonicalTypesToNativeValues` isn't the default yet, according to the release notes it will be the default soon so switched everything to use it.

pkwarren · 2025-09-08T19:23:48Z

 api(libs.protobuf.java)
- implementation(libs.cel) {
- // https://github.com/google/cel-java/issues/748
- exclude(group = "com.google.protobuf", module = "protobuf-javalite")


Upstream issue is fixed in this release.

pkwarren · 2025-09-08T19:24:19Z

 BytesIn invalid = BytesIn.newBuilder().setVal(ByteString.copyFromUtf8("bar")).build();
 ValidationResult validate = validator.validate(invalid);
- assertThat(validate.isSuccess()).isTrue();
+ assertThat(validate.getViolations()).isEmpty();


Just showing expected true, got false isn't great for debugging. Made it use isEmpty so it will print the violations.

pkwarren · 2025-09-08T19:25:07Z

 }
- for (int i = 0; i < param.size(); i++) {
- if (param.byteAt(i) != receiver.byteAt(i)) {
+ byte[] paramBytes = param.toByteArray();


It would be nice if there was a byteAt method on CelByteString to avoid having to allocate a byte array to access a single byte.

pkwarren · 2025-09-08T19:30:11Z

 .addCompilerLibraries(validateLibrary)
 .addRuntimeLibraries(validateLibrary)
+ .setOptions(
+ CelOptions.DEFAULT.toBuilder().evaluateCanonicalTypesToNativeValues(true).build())


This opts in to the new behavior which uses CEL types instead of Protobuf types for NullValue / ByteString. This is apparently coming soon as the default, so since we had to update our overloads to use the new types just opted in to migrate everything to use them.

pkwarren · 2025-09-08T19:31:00Z

+
+ @Override
+ public String toString() {
+ return proto.toString();


This made debugging more painful - added a toString which will dump the contents of the proto Validation message.

pkwarren requested review from rodaine, srikrsna-buf and timostamm September 8, 2025 19:23

Further cleanup to use CEL types
a02479e

pkwarren commented Sep 8, 2025
View reviewed changes

srikrsna-buf approved these changes Sep 8, 2025
View reviewed changes

pkwarren merged commit e32891d into mainSep 8, 2025
4 checks passed

pkwarren deleted the pkw/cel-0.11.0-upgrade branch September 8, 2025 19:47

timostamm mentioned this pull request Nov 25, 2025
[BUG] Validation of message with com.google.protobuf.Timestamp fails with java.time.DateTimeException if seconds/nanos exceed Instant bounds #379
Open

pkwarren requested review from rodaine, srikrsna-buf and timostamm September 8, 2025 19:23

Further cleanup to use CEL types
a02479e

pkwarren commented Sep 8, 2025
View reviewed changes

srikrsna-buf approved these changes Sep 8, 2025
View reviewed changes

pkwarren merged commit e32891d into mainSep 8, 2025
4 checks passed

pkwarren deleted the pkw/cel-0.11.0-upgrade branch September 8, 2025 19:47

timostamm mentioned this pull request Nov 25, 2025
[BUG] Validation of message with com.google.protobuf.Timestamp fails with java.time.DateTimeException if seconds/nanos exceed Instant bounds #379
Open

pkwarren requested review from rodaine, srikrsna-buf and timostamm September 8, 2025 19:23

Further cleanup to use CEL types
a02479e

pkwarren commented Sep 8, 2025
View reviewed changes

srikrsna-buf approved these changes Sep 8, 2025
View reviewed changes

pkwarren merged commit e32891d into mainSep 8, 2025
4 checks passed

pkwarren deleted the pkw/cel-0.11.0-upgrade branch September 8, 2025 19:47

timostamm mentioned this pull request Nov 25, 2025
[BUG] Validation of message with com.google.protobuf.Timestamp fails with java.time.DateTimeException if seconds/nanos exceed Instant bounds #379
Open

pkwarren requested review from rodaine, srikrsna-buf and timostamm September 8, 2025 19:23

Further cleanup to use CEL types
a02479e

pkwarren commented Sep 8, 2025
View reviewed changes

srikrsna-buf approved these changes Sep 8, 2025
View reviewed changes

pkwarren merged commit e32891d into mainSep 8, 2025
4 checks passed

pkwarren deleted the pkw/cel-0.11.0-upgrade branch September 8, 2025 19:47

timostamm mentioned this pull request Nov 25, 2025
[BUG] Validation of message with com.google.protobuf.Timestamp fails with java.time.DateTimeException if seconds/nanos exceed Instant bounds #379
Open

pkwarren requested review from rodaine, srikrsna-buf and timostamm September 8, 2025 19:23

Further cleanup to use CEL types
a02479e

pkwarren commented Sep 8, 2025
View reviewed changes

srikrsna-buf approved these changes Sep 8, 2025
View reviewed changes

pkwarren merged commit e32891d into mainSep 8, 2025
4 checks passed

pkwarren deleted the pkw/cel-0.11.0-upgrade branch September 8, 2025 19:47

timostamm mentioned this pull request Nov 25, 2025
[BUG] Validation of message with com.google.protobuf.Timestamp fails with java.time.DateTimeException if seconds/nanos exceed Instant bounds #379
Open

pkwarren requested review from rodaine, srikrsna-buf and timostamm September 8, 2025 19:23

Further cleanup to use CEL types
a02479e

pkwarren commented Sep 8, 2025
View reviewed changes

srikrsna-buf approved these changes Sep 8, 2025
View reviewed changes

pkwarren merged commit e32891d into mainSep 8, 2025
4 checks passed

pkwarren deleted the pkw/cel-0.11.0-upgrade branch September 8, 2025 19:47

timostamm mentioned this pull request Nov 25, 2025
[BUG] Validation of message with com.google.protobuf.Timestamp fails with java.time.DateTimeException if seconds/nanos exceed Instant bounds #379
Open

pkwarren requested review from rodaine, srikrsna-buf and timostamm September 8, 2025 19:23

Further cleanup to use CEL types
a02479e

pkwarren commented Sep 8, 2025
View reviewed changes

srikrsna-buf approved these changes Sep 8, 2025
View reviewed changes

pkwarren merged commit e32891d into mainSep 8, 2025
4 checks passed

pkwarren deleted the pkw/cel-0.11.0-upgrade branch September 8, 2025 19:47

timostamm mentioned this pull request Nov 25, 2025
[BUG] Validation of message with com.google.protobuf.Timestamp fails with java.time.DateTimeException if seconds/nanos exceed Instant bounds #379
Open

pkwarren requested review from rodaine, srikrsna-buf and timostamm September 8, 2025 19:23

Further cleanup to use CEL types
a02479e

pkwarren commented Sep 8, 2025
View reviewed changes

srikrsna-buf approved these changes Sep 8, 2025
View reviewed changes

pkwarren merged commit e32891d into mainSep 8, 2025
4 checks passed

pkwarren deleted the pkw/cel-0.11.0-upgrade branch September 8, 2025 19:47

timostamm mentioned this pull request Nov 25, 2025
[BUG] Validation of message with com.google.protobuf.Timestamp fails with java.time.DateTimeException if seconds/nanos exceed Instant bounds #379
Open

Uh oh!

Conversation

pkwarren commented Sep 8, 2025

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Uh oh!

Conversation

pkwarren commented Sep 8, 2025

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Uh oh!

Conversation

pkwarren commented Sep 8, 2025

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Uh oh!

Conversation

pkwarren commented Sep 8, 2025

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment