blob: 18886695a43869190b51156cd5759ecea8a1b4ea [file] [log] [blame]
// Copyright 2014 The Bazel Authors. All rights reserved.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package com.google.devtools.build.lib.packages;
import com.google.auto.value.AutoValue;
import com.google.common.collect.ImmutableList;
import com.google.devtools.build.lib.cmdline.Label;
import com.google.devtools.build.lib.events.Location;
import com.google.devtools.build.lib.syntax.Type;
import java.util.Collection;
import javax.annotation.Nullable;
/**
* The interface for accessing a {@link Rule}'s attributes.
*
* <p>Since what an attribute lookup should return can be ambiguous (e.g. for configurable
* attributes, should we return a configuration-resolved value or the original, unresolved
* selection expression?), different implementations can apply different policies for how to
* fulfill these methods. Calling code can then use the appropriate implementation for whatever
* its particular needs are.
*/
public interface AttributeMap {
/**
* Returns the name of the rule; this is equivalent to {@code getLabel().getName()}.
*/
String getName();
/**
* Returns the label of the rule.
*/
Label getLabel();
/**
* Returns the name of the rule class.
*/
String getRuleClassName();
/**
* Returns true if an attribute with the given name exists.
*/
boolean has(String attrName);
/**
* Returns true if an attribute with the given name exists with the given type.
*
* <p>Don't use this version unless you really care about the type.
*/
<T> boolean has(String attrName, Type<T> type);
/**
* Returns the value of the named rule attribute, which must be of the given type. This may
* be null (for example, for an attribute with no default value that isn't explicitly set in
* the rule - see {@link Type#getDefaultValue}).
*
* <p>If the rule doesn't have this attribute with the specified type, throws an
* {@link IllegalArgumentException}.
*/
@Nullable
<T> T get(String attributeName, Type<T> type);
/**
* Returns true if the given attribute is configurable for this rule instance, false
* if it isn't configurable or doesn't exist.
*/
boolean isConfigurable(String attributeName);
/**
* Returns the names of all attributes covered by this map.
*/
Iterable<String> getAttributeNames();
/**
* Returns the type of the given attribute, if it exists. Otherwise returns null.
*/
@Nullable
Type<?> getAttributeType(String attrName);
/**
* Returns the attribute definition whose name is {@code attrName}, or null
* if not found.
*/
@Nullable Attribute getAttributeDefinition(String attrName);
/**
* Returns true iff the specified attribute is explicitly set in the target's definition (as
* opposed to being omitted and taking on its default value from the rule definition).
*
* <p>Note that this returns true in the case where the attribute is explicitly set to the same
* value as its default. Therefore, this method breaks encapsulation in the sense that it
* describes *how* a target is defined rather than just *what* its attribute values are.
*
* <p>CAUTION: It is a good idea to avoid relying on this method if possible. It's confusing to
* users that setting an attribute to (for example) an empty list is different from not setting it
* at all. It also breaks some use cases, such as programmatically copying a target definition via
* {@code native.existing_rules}. Specifically, the Starlark code doing the copying will observe
* the attribute on the existing target whether or not it was set explicitly, and then set that
* value explicitly on the new target. This can cause the two targets to behave differently, and
* can be a difficult bug to track down. (See #7071, b/122596733).
*/
boolean isAttributeValueExplicitlySpecified(String attributeName);
/** Returns the {@link Location} at which the attribute was defined. */
Location getAttributeLocation(String attrName);
/**
* Returns a {@link Collection} with a {@link DepEdge} for every attribute that contains labels in
* its value (either by *being* a label or being a collection that includes labels).
*/
Collection<DepEdge> visitLabels() throws InterruptedException;
/**
* {@code (Label, Attribute)} pair describing a dependency edge.
*
* <p>The {@link Label} is the target node of the {@code (Rule, Label)} edge. The source node
* should already be known. The {@link Attribute} is the attribute giving the edge.
*/
@AutoValue
abstract class DepEdge {
public abstract Label getLabel();
public abstract Attribute getAttribute();
static DepEdge create(Label label, Attribute attribute) {
return new AutoValue_AttributeMap_DepEdge(label, attribute);
}
}
// TODO(bazel-team): These methods are here to support computed defaults that inherit
// package-level default values. Instead, we should auto-inherit and remove the computed
// defaults. If we really need to give access to package-level defaults, we should come up with
// a more generic interface.
String getPackageDefaultHdrsCheck();
Boolean getPackageDefaultTestOnly();
String getPackageDefaultDeprecation();
ImmutableList<String> getPackageDefaultCopts();
}