Buf CLI

Rules and categories

The rules and categories described here belong to the v1 and v2 configurations. If you're still using v1beta1 configuration files and haven't migrated yet, refer to the previous reference.

The Buf CLI applies individual lint rules across your Protobuf schema, reporting any violations as errors. This page describes the available categories and the individual rules within each category. See the overview for usage and the buf.yaml reference for configuration options.

Categories

The Buf CLI provides three top-level categories of increasing strictness, which are most of the lint rules you'll probably want to apply:

It also provides two categories outside of the strictness hierarchy that enforce additional useful constraints:

MINIMAL

The MINIMAL category represents what we consider to be fundamental rules for modern Protobuf development. Not applying them can lead to many bad situations across Protobuf plugins, especially plugins that aren't built into protoc itself, and there's no downside to applying them. They're described in more detail below:

The MINIMAL category verifies that all files with package foo.bar.baz.v1 are in the directory foo/bar/baz/v1 (relative to the buf.yaml file), and that only one such directory exists. protoc doesn't enforce file structure in any way, but you're likely to have a rough time with many Protobuf plugins across various languages if you don't do this. Many languages such as Go and Java explicitly or effectively enforce such a file structure.

For example, consider this tree:

.
β”œβ”€β”€ buf.yaml
└── foo
    └── bar
        β”œβ”€β”€ bat
        β”‚Β Β  └── v1
        β”‚Β Β      └── bat.proto // package foo.bar.bat.v1
        └── baz
            └── v1
                β”œβ”€β”€ baz.proto         // package foo.bar.baz.v1
                └── baz_service.proto // package foo.bar.baz.v1

Arranging the files this way also has the effect of allowing imports to self-document their package. For example, you can discern that the import foo/bar/bat/v1/bat.proto has types in the package foo.bar.bat.v1.

BASIC

The BASIC category includes everything from the MINIMAL category, and adds rules that are widely accepted as standard Protobuf style. These rules should generally be applied for all Protobuf schemas.

The additional rules in the BASIC category are:

DEFAULT

The DEFAULT category includes everything from the BASIC category and some additional rules that represent our recommendations for modern Protobuf development. True to its name, DEFAULT is also the default set of lint rules used by the Buf CLI if the buf.yaml file has no lint settings configured.

The additional rules in the DEFAULT category are:

COMMENTS

This is an extra top-level category that enforces the presence of comments on various parts of your Protobuf schema. It includes these rules:

Only leading comments count towards passing these rules. The rules are separated by object type so that you can selectively enforce which parts of your schema contain comments. You can select individual rules in the COMMENTS category like this:

buf.yaml
version: v2
lint:
  use:
    - DEFAULT
    - COMMENT_ENUM
    - COMMENT_MESSAGE

UNARY_RPC

This is an extra top-level category that outlaws streaming RPCs. It includes these rules:

Some RPC protocols don't allow streaming RPCs, such as Twirp. This category enforces that no developer accidentally adds a streaming RPC if your setup doesn't support them. Additionally, streaming RPCs have a number of issues in general usage. See this discussion for more details.

Rules

COMMENT_ENUM

Categories: COMMENTS

This rule checks that enums have non-empty comments.

COMMENT_ENUM_VALUE

Categories: COMMENTS

This rule checks that enum values have non-empty comments.

COMMENT_FIELD

Categories: COMMENTS

This rule checks that fields have non-empty comments.

COMMENT_MESSAGE

Categories: COMMENTS

This rule checks that messages have non-empty comments.

COMMENT_ONEOF

Categories: COMMENTS

This rule checks that oneofs have non-empty comments.

COMMENT_RPC

Categories: COMMENTS

This rule checks that RPCs have non-empty comments.

COMMENT_SERVICE

Categories: COMMENTS

This rule checks that services have non-empty comments.

DIRECTORY_SAME_PACKAGE

Categories: MINIMAL, BASIC, DEFAULT

This rule checks that all files in a given directory are in the same package.

ENUM_FIRST_VALUE_ZERO

Categories: BASIC, DEFAULT

This rule enforces that the first enum value is the zero value, which is a proto3 requirement on build, but isn't required in proto2 on build. The rule enforces that the requirement is also followed in proto2.

This example would result in a linting error if the rule is active:

syntax = "proto2";

enum Scheme {
  // *** DON'T DO THIS ***
  SCHEME_FTP = 1;
  SCHEME_UNSPECIFIED = 0;
}

ENUM_NO_ALLOW_ALIAS

Categories: BASIC, DEFAULT

This rule outlaws aliased enums like this:

enum Foo {
  option allow_alias = true;
  FOO_UNSPECIFIED = 0;
  FOO_ONE = 1;
  FOO_TWO = 1; // *** DON'T DO THIS ***
}

The Protobuf allow_alias option lets multiple enum values have the same number. This can lead to issues when working with the JSON representation of Protobuf, which is a first-class citizen of proto3. If you get a serialized Protobuf value over the wire in binary format, the specific enum value it applies to is unknown, and JSON usually serializes enum values by name. This can lead to hard-to-track bugs if you declare an alias and expect names to be interchangeable.

Instead of having an alias, we recommend deprecating your current enum and making a new one with the enum value name you want, or just sticking with the current name for your enum value.

ENUM_PASCAL_CASE

Categories: BASIC, DEFAULT

This rule checks that enums are PascalCase.

ENUM_VALUE_PREFIX

Categories: DEFAULT

This rule requires that all enum value names are prefixed with the enum name. For example:

enum Foo {
  FOO_UNSPECIFIED = 0;
  FOO_ONE = 1;
}

Protobuf enums use C++ scoping rules, which makes it impossible to have two enums in the same package with the same enum value name (an exception is when enums are nested, in which case this rule applies within the given message). Though you might assume that a given enum value name is always unique across a package, schemas can develop over years, and there are countless examples of developers having to compromise on their enum names due to backwards compatibility issues. For example, you might have this enum:

enum Scheme {
  // Right off the bat, you can't use "UNSPECIFIED" in multiple enums
  // in the same package, so you always would have to prefix this anyway.
  SCHEME_UNSPECIFIED = 0;
  HTTP = 1;
  HTTPS = 2;
  ...
}

Two years later, you have an enum in the same package you want to add, but can't:

// This is a made up example, bear with us.
enum SecureProtocol {
  SECURE_PROTOCOL_UNSPECIFIED = 0;
  // If this enum is in the same package as Scheme, this produces
  // a protoc compile-time error!
  HTTPS = 1;
  ...
}

ENUM_VALUE_UPPER_SNAKE_CASE

Categories: BASIC, DEFAULT

This rule checks that enum values are UPPER_SNAKE_CASE.

ENUM_ZERO_VALUE_SUFFIX

Categories: DEFAULT

This rule requires that all enum values have a zero value with a defined suffix. By default, it verifies that the zero value of all enums ends in _UNSPECIFIED, but the suffix is configurable.

enum Foo {
  FOO_UNSPECIFIED = 0;
}

All enums should have a zero value. proto3 doesn't differentiate between set and unset fields, so if an enum field isn't explicitly set, it defaults to the zero value. If an explicit zero value isn't part of the enum definition, this defaults to the actual zero value of the enum. For example, if you had the following .proto file, any Uri with scheme not explicitly set would default to SCHEME_FTP:

enum Scheme {
  // *** don't DO THIS ***
  SCHEME_FTP = 0
}

message Uri {
  Scheme scheme = 1;
}

FIELD_LOWER_SNAKE_CASE

Categories: BASIC, DEFAULT

This rule checks that field names are lower_snake_case.

FIELD_NOT_REQUIRED

Categories: BASIC, DEFAULT

This rule checks that field is not configured as required. This means that using the "required" label in proto2 sources is not allowed and using the feature field_presence = LEGACY_REQUIRED is not allowed in Editions sources.

This is a new rule that can only be used with v2 configuration files.

FILE_LOWER_SNAKE_CASE

Categories: DEFAULT

This rule says that all .proto files must be named as lower_snake_case.proto. This is the widely accepted standard.

IMPORT_NO_PUBLIC

Categories: BASIC, DEFAULT

This rule outlaws declaring imports as public. If you didn't know that was possible, forget what you just learned in this sentence.

IMPORT_NO_WEAK

Categories: BASIC, DEFAULT

Similar to the IMPORT_NO_PUBLIC rule, this rule outlaws declaring imports as weak. If you didn't know that was possible, forget what you just learned in this sentence.

IMPORT_USED

Categories: BASIC, DEFAULT

This rule checks that all the imports declared across your Protobuf files are actually used. This .proto file, for example, would fail:

syntax = "proto3";

package payments.v1;

import "product.proto"; // Unused import

message Payment {
  string payment_id = 1;
  // other fields
}

MESSAGE_PASCAL_CASE

Categories: BASIC, DEFAULT

This rule checks that messages are PascalCase.

ONEOF_LOWER_SNAKE_CASE

Categories: BASIC, DEFAULT

This rule checks that oneof names are lower_snake_case.

PACKAGE_DEFINED

Categories: MINIMAL, BASIC, DEFAULT

This rule checks that all files have a package declaration.

PACKAGE_DIRECTORY_MATCH

Categories: MINIMAL, BASIC, DEFAULT

This rule checks that all files are in a directory that matches their package name.

PACKAGE_LOWER_SNAKE_CASE

Categories: BASIC, DEFAULT

This rule checks that packages are lower_snake_case.

PACKAGE_NO_IMPORT_CYCLE

Categories: DEFAULT (only for v2 configuration files, otherwise uncategorized)

This is adetects package import cycles. The Protobuf compiler outlaws circular file imports, but it's still possible to introduce package cycles, such as these:

.
β”œβ”€β”€ bar
β”‚   β”œβ”€β”€ four.proto
β”‚   └── three.proto
└── foo
    β”œβ”€β”€ one.proto
    └── two.proto
# foo/one.proto
syntax = "proto3";

package foo;

import "bar/three.proto";

message One {
    bar.Three three = 3;
}
# bar/four.proto
syntax = "proto3";

package bar;

import "foo/one.proto";

message Four {
    foo.One one = 1;
}

These packages successfully compile, but this file structure introduces problems for languages that rely on package-based imports, such as Go. If possible, this rule should always be configured.

PACKAGE_SAME_<file_option>

Categories: BASIC, DEFAULT

The Buf CLI doesn't lint file option values, as explained in the What we left out section below. However, it's important to have consistent file option values across all files in a given Protobuf package if you do use them.

  • PACKAGE_SAME_CSHARP_NAMESPACE checks that all files with a given package have the same value for the csharp_namespace option.
  • PACKAGE_SAME_GO_PACKAGE checks that all files with a given package have the same value for the go_package option.
  • PACKAGE_SAME_JAVA_MULTIPLE_FILES checks that all files with a given package have the same value for the java_multiple_files option.
  • PACKAGE_SAME_JAVA_PACKAGE checks that all files with a given package have the same value for the java_package option.
  • PACKAGE_SAME_PHP_NAMESPACE checks that all files with a given package have the same value for the php_namespace option.
  • PACKAGE_SAME_RUBY_PACKAGE checks that all files with a given package have the same value for the ruby_package option.
  • PACKAGE_SAME_SWIFT_PREFIX checks that all files with a given package have the same value for the swift_prefix option.

Each of these rules verify that if a given file option is used in one file in a given package, it's used in every file in that package.

For example, if you have file foo_one.proto:

// foo_one.proto
syntax = "proto3";

package foo.v1;

option go_package = "foov1";
option java_multiple_files = true;
option java_package = "com.foo.v1";

Another file foo_two.proto with package foo.v1 must have these three options set to the same value, and the other options unset:

// foo_two.proto
syntax = "proto3";

package foo.v1;

option go_package = "foov1";
option java_multiple_files = true;
option java_package = "com.foo.v1";

PACKAGE_SAME_DIRECTORY

Categories: MINIMAL, BASIC, DEFAULT

This rule checks that all files with a given package are in the same directory.

PACKAGE_VERSION_SUFFIX

Categories: DEFAULT

This rule enforces that the last component of a package must be a version of the form v\d+, v\d+test.*, v\d+(alpha|beta)\d*, or v\d+p\d+(alpha|beta)\d*, where numbers are >=1.

Valid examples:

foo.v1
foo.v2
foo.bar.v1
foo.bar.v1alpha
foo.bar.v1alpha1
foo.bar.v1alpha2
foo.bar.v1beta
foo.bar.v1beta1
foo.bar.v1beta2
foo.bar.v1p1alpha
foo.bar.v1p1alpha1
foo.bar.v1p1alpha2
foo.bar.v1p1beta
foo.bar.v1p1beta1
foo.bar.v1p1beta2
foo.bar.v1test
foo.bar.v1testfoo

One of the core promises of Protobuf schema development is to never have breaking changes in your APIs. There are scenarios, however, where you do want to properly version your schema. Instead of making changes, the proper method is to make a completely new Protobuf package that's a copy of your existing Protobuf package, serve both packages server-side, and manually migrate your callers. This rule enforces that all packages have a version attached so that it's clear when a package represents a new version.

PROTOVALIDATE

Categories: DEFAULT

This rule requires that all protovalidate constraints specified are valid.

For a buf.validate.field to be valid, it must ensure:

  • skipped is the only field if set.
  • At most one of required and ignore_empty are set.
  • required isn't set if the field belongs to a oneof.
  • Neither required nor ignore_empty is set if the field is an extension.
  • Its CEL constraints are valid.
  • Its type specific rules, such as (buf.validate.field).int32, are valid.

For a buf.validate.message to be valid, it must ensure:

  • disabled is the only field if set.
  • Its CEL constraints are valid.

For a set of CEL constraints on a message or field to be valid, each constraint must:

  • Have a CEL expression that compiles successfully and evaluates to a string or boolean. These are the only two types that the protovalidate runtime allows, and it's a runtime error for a CEL expression to evaluate to another type.
  • Have a non-empty message if the CEL expression evaluates to a boolean value. This message is used by the protovalidate runtime report validation failure.
  • Have an empty message if the CEL expression evaluates to a string value. The validation failure message in this case is the value this CEL expression evaluates to, while message won't be used in any way.
  • Have a non-empty id, consisting of only alphanumeric characters, _, - and .. The id must be unique within the buf.validate.message or buf.validate.field it's specified on. A unique id is useful for debugging and locating the CEL constraint that fails, and can be used as a key for i18n.

For a set of rules specified on a field, such as (buf.validate.field).int32, to be valid, it must additionally:

  • Have a type compatible with the type it validates: (buf.validate.field).int32 rules can only be set on a field of type int32 or google.protobuf.Int32Value. A type mismatch causes a runtime error.
  • Permit some value: setting contains: "foo" and not_contains: "foo" isn't valid because it rejects all values.
  • Have no obviously redundant rules. For example, it's redundant to set lt: 5 and const: 3.

Numeric rules, timestamp rules and duration rules

  • The field to validate must match the rules type or its corresponding wrapper type (if any).
  • If a lower bound (gt or gte) and an upper bound(lt or lte) are both specified, they must not be equal. If they are both inclusive (gte and lte), they must be replaced by const. Otherwise, all values are invalid.
  • Durations and timestamps defined in options, such as (buf.validate.field).timestamp.lt, must be valid.
  • If the rule is timestamp:
    • within must be a positive duration.
    • lt_now and gt_now must not both be specified.

String rules

  • The field to validate must be string or google.protobuf.StringValue.
  • If len is specified, min_len or max_len must not be specified. If both are specified, min_len must be lower than max_len.
  • If len_bytes is specified, min_bytes or max_bytes must not be specified. If both are specified, min_bytes must be lower than max_bytes.
  • If min_len and max_bytes are both defined, min_len must be less than or equal to max_bytes. It's impossible for a string to have 3 or more UTF-8 characters while having less than 2 bytes.
  • If min_bytes and max_len are both defined, min_bytes must be less than or equal to 4 times max_len. It's impossible for a string to have 2 or less UTF-8 characters while having 9 or more bytes, since each UTF-8 character takes at most 4 bytes.
  • If prefix, suffix, or contains is specified, its length must not exceed max_len and max_bytes. Otherwise, all values are invalid.
  • Any value of prefix, suffix and contains must not contain, or be a substring of not_contains, if they're both specified.
  • If strict is set to false, well_known_regex must also be specified.
  • If pattern is specified, is must be a valid regular expression in RE2 syntax.

Bytes rules

  • The field to validate must be bytes or google.protobuf.BytesValue.
  • If len is specified, min_len or max_len must not be specified. If both are specified, min_len must be lower than max_len.
  • If any of prefix, suffix and contains is specified, its length must not exceed max_len. Otherwise, all values are invalid.
  • If pattern is specified, is must be a valid regular expression in RE2 syntax.

Map rules

  • The field to validate must be a map.
  • min_pairs must not be higher than max_pairs.
  • The set of rules in keys must be valid and compatible with the map field's key type.
  • The set of rules in values must be valid and compatible with the map field's value type.

Repeated rules

  • The field to validate must have label repeated.
  • min_items must not be higher than max_items.
  • The set of rules in items must be compatible with the field's type.
  • If unique is set to true, the field must be a scalar or a wrapper type.

RPC_NO_CLIENT_STREAMING

Categories: UNARY_RPC

This rule checks that RPCs aren't client streaming.

RPC_NO_SERVER_STREAMING

Categories: UNARY_RPC

This rule checks that RPCs aren't server streaming.

RPC_PASCAL_CASE

Categories: BASIC, DEFAULT

This rule checks that RPCs are PascalCase.

RPC_REQUEST_STANDARD_NAME RPC_RESPONSE_STANDARD_NAME RPC_REQUEST_RESPONSE_UNIQUE

Categories: DEFAULT

These rules enforce the message name of RPC request/responses, and that all request/responses are unique.

One of the single most important rules to enforce in modern Protobuf development is to have a unique request and response message for every RPC. Separate RPCs shouldn't have their request and response parameters controlled by the same Protobuf message, and if you share a Protobuf message between multiple RPCs, this results in multiple RPCs being affected when fields on this Protobuf message change. Even in straightforward cases, best practice is to always have a wrapper message for your RPC request and response types. The Buf CLI enforces this with these three rules by verifying that:

  • All request and response messages are unique across your Protobuf schema.
  • All request and response messages are named after the RPC, either by naming them MethodNameRequest, MethodNameResponse or ServiceNameMethodNameRequest, ServiceNameMethodNameResponse.

This service definition, for example, abides by these rules:

// request/response message definitions omitted for brevity

service FooService {
  rpc Bar(BarRequest) returns (BarResponse) {}
  rpc Baz(FooServiceBazRequest) returns (FooServiceBazResponse) {}
}

Though we don't recommend it, we provide a few configuration options to loosen these restrictions somewhat:

SERVICE_PASCAL_CASE

Categories: BASIC, DEFAULT

This rule checks that services are PascalCase.

SERVICE_SUFFIX

Categories: DEFAULT

This rule enforces that all services are suffixed with Service. For example:

service FooService {}
service BarService {}
service BazService {}

Service names inherently end up having a lot of overlap with package names, and service naming often ends up inconsistent as a result across a larger Protobuf schema. Enforcing a consistent suffix takes away some of this inconsistency.

The suffix is configurable. For example, if you have this configuration in your buf.yaml:

buf.yaml
version: v2
lint:
  service_suffix: Endpoint

then the SERVICE_SUFFIX rule enforces this naming instead:

service FooEndpoint {}
service BarEndpoint {}
service BazEndpoint {}

What we left out

We think that the above lint rules represent a set that sufficiently enforces consistent and maintainable Protobuf schemas, while still enabling your organization to make design decisions. However, there are some rules we purposefully didn't write that deserve mention.

File option values

The Buf CLI doesn't include linting for specific file option values. It's not that we don't think consistency across these file options is importantβ€”in fact, we think it simplifies Protobuf stub consumption. One of our core principles is that language-specific file options shouldn't be part of your core Protobuf schemaβ€”your Protobuf schema should only describe language-independent elements as much as possible.

The values for most file options, in fact, should be deduced in a stable and deterministic manner. For example, we think that java_package should likely be a constant prefix followed by the package name as a suffix. Your go_package should use the last component of your package name. And java_multiple_files should always be true. These aren't defaults for backwards-compatibility reasons, but if you're using a tool like the Buf CLI to produce your stubs, you shouldn't have to think about any of this.

This is exactly why we've created managed mode, which sets all of these file options on the fly with buf generate.

The Buf CLI still enforces that specific file options are the same across a given package through the BASIC and DEFAULT categories described above. We do find this to be important, regardless of what values you choose. Fortunately, with managed mode you can remove file option declarations from your Protobuf files altogether.

Custom options

There are no lint rules for widely used custom options such as google.api options or protoc-gen-validateβ€”we currently only support the standard set of file options. Contact us if this is a big need for your organization.

Naming opinions

We stay away from enforcing naming opinions, such as package name restrictions (beyond versioning requirements and lower_snake_case), or field naming such as google.protobuf.Duration name standardization. This is to provide maximum usefulness of the DEFAULT category out of the box.

Adding or requesting new rules

If you'd like a new rule added, contact us to discuss it. We'll add rules if we think they're maintainable and could have widespread value. Most rules can be easily added, and although Buf is OSS, it's usually more efficient for us to add it ourselves.

Style guide

The style guide provides a concise document that includes all rules in the DEFAULT category, as well as additional recommendations that aren't enforced by the linter. We provide this for ease of consumption across your teams.