| Commit message (Collapse) | Author | Age |
|
|
|
| |
PiperOrigin-RevId: 183859414
|
|
|
|
| |
PiperOrigin-RevId: 183826311
|
|
|
|
|
|
|
|
| |
BuildConfiguration.Fragment>> set of Fragment classes that is part of the BuildConfigurationValue.Key. This class allows us to compute a fingerprint of the wrapped ImmutableSortedSet, making equality comparisons fast. The number of additional wrapper objects is the number of distinct sets of fragment classes, so 1. (In fact, we don't even need to compute a fingerprint, since reference equality does the job for us here, but we do it just to be conservative.)
This CL has a performance benefit for Bazel currently, but has a bigger performance benefit in the following changes, where there are more BuildConfigurationValue.Key objects to compare.
PiperOrigin-RevId: 183090122
|
|
|
|
|
|
|
|
| |
ImmutableSortedSet wherever possible, and use a known explicit ImmutableSortedSet in the case of two sets being equal. This is mainly a cosmetic cleanup for the sequel changes.
Also rename test-only methods in SkyframeExecutor to indicate that, and do a drive-by clean-up of a test that reported hard crashes confusingly because it wrapped RuntimeExceptions.
PiperOrigin-RevId: 182984572
|
|
|
|
| |
PiperOrigin-RevId: 182099803
|
|
|
|
|
|
|
| |
HostTransition can't be migrated yet because it depends on
BuildConfiguration.
PiperOrigin-RevId: 180842784
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
config.transitions.ConfigurationTransitionProxy.
The "proxy" part is to dissuade people from writing:
void myfunc(ConfigurationTransition transition)
signatures casually.
Maybe that's actually a better name than "Transition". But I'd rather
rename Transition to ConfigurationTransition in its own change if we
want to do that.
PiperOrigin-RevId: 180285321
|
|
|
|
| |
PiperOrigin-RevId: 179948322
|
|
|
|
| |
PiperOrigin-RevId: 179936355
|
|
|
|
|
|
|
|
|
|
|
| |
If an aspect is applied to a rule+aspect node, all attributes are merged
into ctx.rule.attr collection, and the first one with the same name wins
(in particular, rule attribute will always win over aspect attribute).
This is backwards-compatible, and unlikely to cause problems in
practice.
RELNOTES: Aspects-on-aspect see and propagate over aspect attributes.
PiperOrigin-RevId: 179675660
|
|
|
|
|
|
|
|
| |
Blaze had its own class to avoid GC from varargs array creation for the precondition happy path. Guava now (mostly) implements these, making it unnecessary to maintain our own.
This change was almost entirely automated by search-and-replace. A few BUILD files needed fixing up since I removed an export of preconditions from lib:util, which was all done by add_deps. There was one incorrect usage of Preconditions that was caught by error prone (which checks Guava's version of Preconditions) that I had to change manually.
PiperOrigin-RevId: 175033526
|
|
|
|
|
| |
RELNOTES: Late-bound attributes are exposed to skylark. This is a new API (`configuration_field()`) to depend on certain configuration-defined targets from skylark rules.
PiperOrigin-RevId: 174534104
|
|
|
|
|
|
|
| |
Configurable"
RELNOTES: None.
PiperOrigin-RevId: 171751391
|
|
|
|
|
|
|
|
|
| |
This requires moving the convenience constructor using RuleConfiguredTarget to be owned by RuleConfiguredTarget.
This refactoring is required by later work to allow SplitTransitionProvider to use configurable attributes. This would require packages/Attribute.java -> analysis/ConfiguredAttributeMapper.java, where in general, the 'analysis' package depends on the 'packages' package. This is the easiest way to prevent a circular dependency.
RELNOTES: None.
PiperOrigin-RevId: 171741620
|
|
|
|
|
|
|
|
| |
Also clarify the interfaces *TransitionResolver* - which determines what
transition to apply to an input configuration and *ConfigurationResolver*
- which determines the output configuration from that transition.
PiperOrigin-RevId: 169311986
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Currently, there is no way to enforce that LateBoundDefaults only access
the fragments that they declare. This means that LateBoundDefaults can
fail to declare fragments at all, or declare the wrong ones, and still
have no troubles.
But when trimming, these fragments must be declared, because otherwise
they will not necessarily be available.
This change refactors LateBoundDefault to declare a single fragment type,
not a set. All existing LateBoundDefaults use sets with a single element
or no elements at all for their set of fragment classes, so this does not
limit anything being done currently.
To account for LateBoundDefaults which do not use configuration at all,
typically those which only want to access the configured attribute map,
it is possible for Void to be the fragment class which is requested.
To account for LateBoundDefaults which need to access methods of the
BuildConfiguration instance itself, it is possible for BuildConfiguration
to be the fragment class which is requested; however, this is unsafe, so
it is only a temporary state until a way to do this without also giving
access to all of the fragments can be added.
Drive-by refactoring: LateBoundDefaults' values are now typed. All actual
production LateBoundDefaults were Label or List<Label> typed, through the
LateBoundLabel and LateBoundLabelList subclasses. These subclasses have
been removed, and LateBoundDefault has two type parameters, one for the
type of its input, and one for the type of its output.
RELNOTES: None.
PiperOrigin-RevId: 169242278
|
|
|
|
|
|
|
|
| |
With dynamic configurations we no longer need a special
class to apply transitions: a simple patch transition
bound to an attribute works just as well.
PiperOrigin-RevId: 168442602
|
|
|
|
|
|
| |
Part of the static config cleanup effort.
PiperOrigin-RevId: 168270713
|
|
|
|
|
|
| |
available from BuildViewTestCase#getRuleContext
PiperOrigin-RevId: 168231020
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This accomplishes a few goals:
1. Removes the outdated BuildConfiguration.StaticConfigurationApplier code.
2. Removes the TransitionApplier abstraction completely. This was an awkward
bridge meant to support both static and dynamic implementations.
3. Moves transition logic to its own dedicated class: ConfigurationResolver.
This no longer belongs in BuildConfiguration, which we ultimately want to
become a simple <key, value> map.
Part of the static config cleanup effort.
PiperOrigin-RevId: 165736955
|
|
|
|
|
|
|
|
| |
This is always true.
Part of the static config cleanup effort.
PiperOrigin-RevId: 165628823
|
|
|
|
|
|
|
|
| |
BuildConfigurationCollection.Transitions.
Part of the static config cleanup effort.
PiperOrigin-RevId: 165607492
|
|
|
|
|
|
|
| |
Part of #2219.
Change-Id: Id4929d5ddcd57b4635af5e513eb9a09f16a78e71
PiperOrigin-RevId: 162634398
|
|
|
|
|
|
|
| |
This fixes a bug that put them out of sync from how all other
attributes manage the host configuration.
PiperOrigin-RevId: 154992583
|
|
|
|
|
|
| |
--
PiperOrigin-RevId: 148342788
MOS_MIGRATED_REVID=148342788
|
|
|
|
|
|
|
|
| |
providers.
--
PiperOrigin-RevId: 147526961
MOS_MIGRATED_REVID=147526961
|
|
|
|
|
|
|
|
|
| |
This functionality is never used, have never been exposed to Skylark
and is a continuous pain to maintain and test.
--
PiperOrigin-RevId: 145079832
MOS_MIGRATED_REVID=145079832
|
|
|
|
|
|
|
|
|
| |
For now, only for aspects, but eventually expand to Attribute's
mandatory providers as well.
--
PiperOrigin-RevId: 144063369
MOS_MIGRATED_REVID=144063369
|
|
|
|
|
|
| |
--
PiperOrigin-RevId: 143991903
MOS_MIGRATED_REVID=143991903
|
|
|
|
|
|
| |
--
PiperOrigin-RevId: 142709934
MOS_MIGRATED_REVID=142709934
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
If an Aspect registered an action that an extra-action is
shadowing, its name is used when creating the extra-action's ID and
name.
Since recently, an aspect can see other aspects applied to the same
target. This CL record the names of other aspects applied to the target
as well, disambiguating the action owners.
--
PiperOrigin-RevId: 142264153
MOS_MIGRATED_REVID=142264153
|
|
|
|
|
|
|
| |
against rules. Aspects can attach to a rule if at least one set of required providers are present on the rule.
--
MOS_MIGRATED_REVID=140605023
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=139189444
|
|
|
|
|
|
|
| |
RELNOTES: Do not propagate aspect to its own attributes when using '*'.
--
MOS_MIGRATED_REVID=138194456
|
|
|
|
|
|
|
|
|
|
|
|
| |
The comments for DependencyResolver.dependentNodeMap implied that visiting of
visibility attributes could be controlled with a visitVisibility
parameter. However, no such parameter exists, and the code unconditionally
visits visibility attributes.
--
Change-Id: Iecf803aed771423b2f25e911c041e0017b77d29d
Reviewed-on: https://bazel-review.googlesource.com/#/c/6211/
MOS_MIGRATED_REVID=135454173
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=134378592
|
|
|
|
|
|
|
| |
The only place we now don't handle InterruptedException is in the action graph created after analysis, since I'm not sure that will be around for that much longer.
--
MOS_MIGRATED_REVID=130327770
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
OrderedSetMultimap. This maintains insertion order while eliminating duplicates.
Certain rules, in particular, otherwise break this invariant:
https://github.com/bazelbuild/bazel/blo[]e3e28274cca5b87f48abe33884edb84016dd3/src/main/java/com/google/devtools/build/lib/skyframe/ConfiguredTargetFunction.java#L403
There's no reason (to my knowledge) to need multiple instances of the same <Attribute, Dependency> pair.
More context from Google code review:
(Michael Staib):
> There are many things which pass around a dependentNodeMap or help construct one or modify one. We want an interface which has the right guarantees.
> ListMultimap is not the right interface because it has no guarantee of unique elements, which we want - we don't want the problem that this CL ran into, and there's no reason (that we know of, to be confirmed) that anyone would want multiple identical Dependencies.
> SetMultimap is not the right interface because it has no guarantee of deterministic iteration order or efficient iteration, which we want - dependency order sometimes matters (e.g., Java classpath or C++ link order).
> We agreed that the best way to get what we want is to define our own interface with its own simultaneous uniqueness and iterability guarantees. Unspoken in the discussion was why we wouldn't just use LinkedHashMultimap as the thing we pass around. IMO the reason for that is that we don't care that it be a LinkedHashMultimap specifically; if tomorrow Guava comes out with a faster cooler map that has deterministic and efficient iteration and guarantees element uniqueness, we want it.
> In this case we're going to make the "interface" be a (final?) class: OrderedSetMultimap, an extension of ForwardingSetMultimap which delegates to LinkedHashMultimap, an implementation which does support both of those guarantees.
> I had mentioned in the conversation that none of the Multimap implementations make guarantees about key iteration order, but this is not true - LinkedHashMultimap preserves key insertion order. We should perhaps declare this as part of the OrderedSetMultimap contract as well.
--
MOS_MIGRATED_REVID=130037643
|
|
|
|
|
|
|
|
|
| |
With this change, dynamic configs are at full feature parity for
split transitions (minus some small differences in composed
transitions from BuildConfigurationCollection.configurationHook).
--
MOS_MIGRATED_REVID=129802414
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
with
dynamic configs.
The old code incorrectly applied a no-op instead of hard-setting a specific config.
Testing: This is a prereq piece for the main change adding dynamic split transitions. Once
we have that change, standard Bazel tests over implemented late-bound split attributes
(e.g. AppleBinaryRule: ":cc_toolchain") will provide proper coverage. There's no easy way
to test this directly since the affected code won't really work until the dynamic split change is in.
--
MOS_MIGRATED_REVID=129278253
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
for general readability.
Major changes:
- Remove the intermediate Attribute -> LabelAndConfiguration multimap (computed in resolveAttributes). Instead, feed discovered values directly into the final Attribute -> Dependency map via a new RuleResolver interface.
- Remove all references to LabelAndConfiguration. The configuration is always the owning rule's configuration except for two special cases: late-bound attributes with splits and late-bound attributes with LateBoundDefault.useHostConfiguration. The original interface made this very unclear and required a lot of awkward and sometimes incorrect logic. The new interface only involves configurations for the cases that actually need them.
- Remove an ugly hack caused by BuildConfiguration.evaluateTransition mixing poorly with LateBoundDefault.useHostConfiguration (https://github.com/bazelbuild/bazel/blo[]e172693c27f3efc95ed163e43a9f0a7a6fb4017/src/main/java/com/google/devtools/build/lib/analysis/DependencyResolver.java#L488).
- Remove a hack that applies split transitions twice because of BuildConfiguration.evaluateTransition mixing poorly with late-bound split attributes (https://github.com/bazelbuild/bazel/blo[]e172693c27f3efc95ed163e43a9f0a7a6fb4017/src/main/java/com/google/devtools/build/lib/analysis/DependencyResolver.java#L319). This happens to be innocent now but won't be when nested splits are possible.
- Solidifies the API contract for Attribute.LateBoundDefault.useHostConfiguration.
- Applies clearer naming and more consistent ordering to method parameters.
- Better documentation.
This is all also prep work for dynamic split transitions.
tl;dr: late-bound attributes are legitimately special. Treat them that way to make the rest of DependencyResolver cleaner and hack-free.
--
MOS_MIGRATED_REVID=128582618
|
|
|
|
|
|
|
| |
This will be used to replace RedirectChaser so that we don't need to load packages during configuration creation anymore.
--
MOS_MIGRATED_REVID=121935989
|
|
|
|
|
|
|
|
|
| |
select() is the new abi_deps.
RELNOTES[INC]: Removed cc_* attributes: abi / abi_deps. Use select() instead.
--
MOS_MIGRATED_REVID=121704422
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
NativeAspectClass.
This a large refactoring of the aspects, currently we have the following:
- AspectClasses: The interface AspectClass is a implemented by either
SkylarkAspectClass or NativeAspectClass<NativeAspectFactory>.
They are wrappers for the AspectFactories and they hold the information about
the Class<> of the factory.
- AspectFactories (FooAspect.java): Represented by the interfaces
ConfiguredAspectFactory and NativeAspectFactory, also by
the interface ConfiguredNativeAspectFactory which is the union of the two
aforementioned interfaces.
All aspects implement ConfiguredNativeAspectFactory except Skylark aspects
which implement only ConfiguredAspectFactory.
After this CL the distinction between NativeAspectFactories and NativeAspectClasses
dissappear, namely aspect that extends NativeAspectClass is considered native
and if it implements ConfiguredAspectFactory it is configured.
Therefore the interfaces NativeAspectFactory and ConfiguredNativeAspectFactory
both disappear.
With this refactoring the aspectFactoryMap in the ConfiguredRuleClassProvider
changes its type from (String -> Class<? extends NativeAspectClass>)
to (String -> NativeAspectClass) which means it is now able to have an instance
of the aspect instead of its Class only.
By doing this, it is now possible to pass parameters when creating an
aspect in the ConfiguredRuleClassProvider.
--
MOS_MIGRATED_REVID=120819647
|
|
|
|
|
|
|
|
|
|
|
| |
bind() is assumed to be able to provide any provider. This is suboptimal, but beats the alternative of traversing the dependency graph to an arbitrary depth.
The reason for the removal of the iteration ability in TransitiveInfoCollection is that now aspects can be attached to BindConfiguredTarget, too, which is not a RuleConfiguredTarget. Whereas I could have implemented the iterator, it was used only in BindConfiguredTarget anyway, so there didn't seem to be much reason to.
Some work towards #952.
--
MOS_MIGRATED_REVID=120549877
|
|
|
|
|
|
|
| |
attribute based on the Rule itself (the transition may thus be determined based on the values of other attributes of the rule)
--
MOS_MIGRATED_REVID=120275649
|
|
|
|
|
|
|
| |
Its old name was confusing because resolve() and getDefault() do radically different things: getDefault() returns a good enough lie for when BuildConfiguration is not available, and resolve() resolves the dependency when we do have a BuildConfiguration.
--
MOS_MIGRATED_REVID=120212630
|
|
|
|
|
|
|
|
|
|
| |
Aspect becomes a triple (AspectClass, AspectDefinition,
AspectParameters) and loses its equals() method.
After this CL, SkylarkAspectClass.getDefintion still exists and is
deprecated.
--
MOS_MIGRATED_REVID=119159653
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Previously, this would get thrown when referring to the same package
from both the main and default repositories:
java.lang.IllegalArgumentException: Multiple entries with same key: tools/cpp=/home/brian/971-Robot-Code and tools/cpp=/home/brian/971-Robot-Code
at com.google.common.collect.ImmutableMap.checkNoConflict(ImmutableMap.java:136)
at com.google.common.collect.RegularImmutableMap.checkNoConflictInKeyBucket(RegularImmutableMap.java:98)
at com.google.common.collect.RegularImmutableMap.fromEntryArray(RegularImmutableMap.java:84)
at com.google.common.collect.ImmutableMap$Builder.build(ImmutableMap.java:295)
at com.google.devtools.build.lib.buildtool.BuildTool.transformPackageRoots(BuildTool.java:301)
at com.google.devtools.build.lib.buildtool.BuildTool.buildTargets(BuildTool.java:209)
at com.google.devtools.build.lib.buildtool.BuildTool.processRequest(BuildTool.java:334)
at com.google.devtools.build.lib.runtime.commands.TestCommand.doTest(TestCommand.java:119)
at com.google.devtools.build.lib.runtime.commands.TestCommand.exec(TestCommand.java:104)
at com.google.devtools.build.lib.runtime.BlazeCommandDispatcher.exec(BlazeCommandDispatcher.java:371)
at com.google.devtools.build.lib.runtime.BlazeRuntime$3.exec(BlazeRuntime.java:1016)
at com.google.devtools.build.lib.server.RPCService.executeRequest(RPCService.java:65)
at com.google.devtools.build.lib.server.RPCServer.executeRequest(RPCServer.java:434)
at com.google.devtools.build.lib.server.RPCServer.serve(RPCServer.java:229)
at com.google.devtools.build.lib.runtime.BlazeRuntime.serverMain(BlazeRuntime.java:975)
at com.google.devtools.build.lib.runtime.BlazeRuntime.main(BlazeRuntime.java:772)
at com.google.devtools.build.lib.bazel.BazelMain.main(BazelMain.java:55)
And this would get thrown for any packages in the main repository loaded
from other repositories:
java.lang.RuntimeException: Unrecoverable error while evaluating node 'PACKAGE:@//tools/build_rules/go/toolchain' (requested by nodes )
at com.google.devtools.build.skyframe.ParallelEvaluator$Evaluate.run(ParallelEvaluator.java:982)
at com.google.devtools.build.lib.concurrent.AbstractQueueVisitor$WrappedRunnable.run(AbstractQueueVisitor.java:499)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
at java.lang.Thread.run(Thread.java:745)
Caused by: java.lang.IllegalArgumentException: Invalid BUILD file name for package '@//tools/build_rules/go/toolchain': /home/brian/bazel/tools/build_rules/go/toolchain/BUILD
at com.google.devtools.build.lib.packages.Package.finishInit(Package.java:299)
at com.google.devtools.build.lib.packages.Package$Builder.finishBuild(Package.java:1308)
at com.google.devtools.build.lib.skyframe.PackageFunction.compute(PackageFunction.java:501)
at com.google.devtools.build.skyframe.ParallelEvaluator$Evaluate.run(ParallelEvaluator.java:933)
... 4 more
Sponsor's comment: note the abundance of new Label.resolveRepositoryRelative() calls. They are ugly, but it's only making existing ugliness explicit. Yes, we should fix it, especially in the implementation of configurable attributes.
Refs #940
--
Change-Id: I8bd7f7b00bec58a7157507595421bc50c81b404c
Reviewed-on: https://bazel-review.googlesource.com/#/c/2591
MOS_MIGRATED_REVID=117429733
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
https://github.com/bazelbuild/bazel/blo[]fbbd6a32b95ba746f09dae1eaeaccf675cd5b3/src/main/java/com/google/devtools/build/lib/packages/Attribute.java#L1045
This allows the default value computation for latebound attributes to consider
the values of configurable attributes. This is most directly useful for user-definable Skylark defaults, which have full access to the values of all non-latebound attributes.
Without this change, this kind of scenario crashes Bazel. For example:
------------------
select_rules.bzl:
------------------
def _impl(ctx):
ctx.file_action(
output=ctx.outputs.out_file,
content=ctx.attr.string_value,
)
return struct()
# Bug does not manifest without using this as a default.
def _derived_value(attrs, _):
return Label("//some:dep")
selector_rule = rule(
implementation=_impl,
attrs={
"string_value": attr.string(default=""),
"out_file": attr.output(),
"_derived": attr.label(default=_derived_value),
},
output_to_genfiles=True,
)
def selector_macro(name, out_file="", string_value=""):
# This will fail with selectors.
selector_rule(
name="%s_skylark" % name,
string_value=string_value,
out_file=out_file + ".skylark",
)
# This does not.
native.genrule(
name="%s_genrule" % name,
cmd="echo '" + string_value + "' > $@",
outs=[out_file + ".genrule"],
)
native.filegroup(
name=name,
srcs=[":%s_genrule" % name, "%s_skylark" % name],
)
------------------
BUILD.bzl:
------------------
config_setting(
name = "selector",
values = {"compilation_mode": "opt"},
)
selector_macro(
name = "this_rule",
string_value = """soup? """ + select({
":selector": "no, thank you.",
"//conditions:default": "yes, please!!",
}),
out_file = "this_rule.txt",
)
--
MOS_MIGRATED_REVID=114326474
|