| Commit message (Collapse) | Author | Age |
|
|
|
|
|
|
|
| |
The invalidator is no longer using the SkyValue, so there's no need
to return it when dirtying a node.
--
MOS_MIGRATED_REVID=100307211
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The primary user of invalidation tracking is the SkyframeBuildView,
which monitored which ConfiguredTargetValues were invalidated. It
did that so the SkyframeExecutor could limit its search for artifact
conflicts to when the set of configured targets in the build changed.
For the newly introduced set of dirtied keys,
"dirtiedConfiguredTargetKeys" in SkyframeBuildView, to be as useful as
the "dirtyConfiguredTargets" set it replaced, the
ConfiguredTargetValueInvalidationReceiver must only remove a key from
the set if it was found to be clean when it was re-evaluated. If it
was rebuilt, then the key must stay in the set, to represent that the
set of configured target values has truly changed.
This CL introduces a semantic change that hopefully has a small effect,
if any. Previously, the informInvalidationReceiver method in
InvalidatingNodeVisitor only informed the invalidationReceiver when a
non-null value was invalidated. (This is usually, but not always,
equivalent to a non-error value. The uncommon exception is that in
keep-going builds, some nodes with errors may also have values, and
the invalidator would inform the receiver when such a node was
invalidated.) Now, the receiver is informed that a node is invalidated
regardless of whether its value is null. Because the receiver uses this
information to determine whether artifact conflict search has to be
rerun, and that search is expensive, it's possible this change will
negatively impact performance. However, the only way an extra search
could be invoked is if the invalidated configured target nodes are
all in error. That seems like it would happen rarely, if at all.
Further cleanup of unused SkyValues returned by markDirty to come in
a subsequent CL.
--
MOS_MIGRATED_REVID=100304744
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=100285082
|
|
|
|
|
|
|
|
|
| |
employed by java.lang.String.
Consider the case where thread T1 calls cacheHashCode and then in the future thread T2 calls hashCode. Because the side-effects of 'cacheHashCode' were non-atomic, T2 could see a value of 'true' for 'hashCodeCached' but still see the old value (of 0) for 'hashCode' and incorrectly think the hash code is 0.
--
MOS_MIGRATED_REVID=100283097
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=99961435
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=99943491
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=99747336
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=99411089
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=99197069
|
|
|
|
|
|
|
| |
mismatch.
--
MOS_MIGRATED_REVID=98388239
|
|
|
|
|
|
|
| |
MoreObjects#toStringHelper
--
MOS_MIGRATED_REVID=98383075
|
|
|
|
|
|
|
| |
ParallelEvaluator, to allow users to customize which events they want to store.
--
MOS_MIGRATED_REVID=98299604
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
We do this by implementing equality for TaggedEvents (and all objects
it transitively includes). Before this change, if a Skyframe node
re-evaluated to the same value as in the previous build, but had
(transitive) events, change pruning would not cut off the evaluation
of its parents. This is not a big issue in practice because most nodes
that would re-evaluate to the same value (like FileValues or
GlobValues) never emit events, and others (like ActionExecutionValues)
have secondary caches that mask this effect.
Also do a drive-by fix where we were using the hash code of a nested
set instead of the shallow hash code (didn't have any bad effects in
practice because we never hash these values).
(Minor formatting clean-ups from https://bazel-review.googlesource.com/1610 )
--
Change-Id: I751a8479627f0456993c5ec8834528aeb593d736
Reviewed-on: https://bazel-review.googlesource.com/1610
MOS_MIGRATED_REVID=98115908
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=97926138
|
|
|
|
|
|
|
| |
InMemoryMemoizingEvaluator#delete
--
MOS_MIGRATED_REVID=97908428
|
|
|
|
|
|
|
|
| |
Hooks up the recently introduced interleaved loading functions to
normal graph loading.
--
MOS_MIGRATED_REVID=97679451
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=97022077
|
|
|
|
|
|
|
|
|
| |
Removing since we don't appear to need it. Replaced SkyFunctionName.computed
with create to simplify changing everything. Moved all constructors using new
to use create for consistency. Chose create over new since it was less changes.
--
MOS_MIGRATED_REVID=96914199
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=96433497
|
|
|
|
|
|
|
|
| |
Similar to previous approach, except diff-awareness agnostic. Log
SkyKeys as is for simplicity- this is all for human consumption anyway.
--
MOS_MIGRATED_REVID=96428248
|
|
|
|
|
|
|
| |
getReverseDeps(Target) from the QueryEnvironment interface, since they're no longer needed by any query functions, and also WalkableGraph#get{Direct,Reverse}Deps(SkyKey).
--
MOS_MIGRATED_REVID=96361760
|
|
|
|
|
|
|
| |
shutdown operations.
--
MOS_MIGRATED_REVID=96351438
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=96214911
|
|
|
|
|
|
|
| |
InMemoryNodeEntry#markDirty implementation.
--
MOS_MIGRATED_REVID=95758779
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=94910185
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=93647914
|
|
|
|
|
|
|
|
|
| |
encounters a node with an error.
We were doing this in most cases, but not if the error was in a node that was already done or was revalidated during change pruning.
--
MOS_MIGRATED_REVID=92521223
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=92398218
|
|
|
|
|
|
|
| |
needed. Also clean up some lint warnings/unresolved references.
--
MOS_MIGRATED_REVID=92380618
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
SkyKey argument hashcodes are not stable, because they're composed
of values whose hashcodes may not be stable, such as Java enums.
Therefore a SkyKey's own hashcode isn't stable. And this is fine,
but if we try to serialize and then deserialize the SkyKey with its
cached hashcode, the deserialized SkyKey's cached hashcode won't
match a normally constructed SkyKey, despite them being equal.
Because a SkyKey will deserialize with a default value of 0 for its
cached hashcode, this change also introduces a transient boolean
guard to note whether the correct hashcode has been calculated and
cached.
--
MOS_MIGRATED_REVID=91985674
|
|
|
|
|
|
|
|
|
| |
implementation.
This slightly shortens the massive ParallelEvaluator file, and also allows alternate Environment implementations to avoid repeating the boilerplate.
--
MOS_MIGRATED_REVID=90830529
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=90418160
|
|
|
|
|
|
|
| |
Specifies function return values. Replaces old "builder" terminology. Removes bad advice.
--
MOS_MIGRATED_REVID=90417023
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=90103351
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=89710872
|
|
|
|
|
|
|
| |
and use it in the underlying implementation of SkyFunction.Environment#getValues. It's reasonable for an alternative graph implementation to have a more efficient implementation of getBatch than the naive serial implementation.
--
MOS_MIGRATED_REVID=89708027
|
|
|
|
|
|
|
| |
to do so is very likely to obfuscate the original crash.
--
MOS_MIGRATED_REVID=89138275
|
|
|
|
|
|
|
|
|
| |
SkyFunctionException#validateExceptionType. In practice, this function is only called with subtypes of Exception. Also, this changes emphasizes that the type of an error wrapped in SkyFunctionExceptions must be a subtype of Exception.
Similarly, restrict SchedulerException to only wrapping Exceptions. This is true in practice, but was missed in an earlier change.
--
MOS_MIGRATED_REVID=89127540
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=88758470
|
|
|
|
|
|
|
|
|
| |
This allows the pending action counter to be decremented correctly when in --keep_going mode.
As part of this, there's a small refactor in ParallelEvaluator that also fixes a potential bug, that nodes in error were signaling their parents with the graph version as opposed to their actual version. Because errors never compare equal (ErrorInfo doesn't override equality), this isn't an issue in practice. But it could be in the future.
--
MOS_MIGRATED_REVID=88395500
|
|
|
|
|
|
|
| |
serializable; alternative graph implementations no longer need these.
--
MOS_MIGRATED_REVID=88003503
|
|
|
|
|
|
|
| |
the ordered list of grouped deps.
--
MOS_MIGRATED_REVID=87969259
|
|
|
|
|
|
|
| |
for alternate graph implementations.
--
MOS_MIGRATED_REVID=87755843
|
|
|
|
|
|
|
|
|
| |
This environment eagerly preloads the transitive closure of a specified query "universe", and so may not be as efficient as the standard query for limited-scope queries. It is activated when the universe is specified and ordered results are not requested (since it is currently unable to order results).
Tests were modified/added to exercise this environment where deemed interesting. Some ugly hacks were done to add coverage in AbstractQueryTest and friends, because currently even if the full depot is loaded (using //...), individual target patterns most likely won't be present in the graph. A better way to deal with this situation, suggested by felly, is probably to extract target pattern resolution logic to an auxiliary function so that query is able to resolve target patterns without mutating the graph, and then call into the read-only graph with the resolved patterns. That may be done in a follow-up, in which case the "scope" of every query could be //... .
--
MOS_MIGRATED_REVID=87257028
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=87009938
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=85946859
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=85945243
|
|
--
MOE_MIGRATED_REVID=85702957
|