| Commit message (Collapse) | Author | Age |
|
|
|
| |
PiperOrigin-RevId: 208554171
|
|
|
|
| |
PiperOrigin-RevId: 206389059
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
*** Reason for rollback ***
Go back to the default constructor - instead of requiring everywhere to know the correct hash function, we'll have the default rely on global state. It will make transition easier, even if it makes the origin of the hash less obvious.
*** Original change description ***
Remove default MD5 in most of Bazel's virtual filesystems.
This forces the ex-default to be explicit in a lot of tests, but I'd rather that than have the risk of implicit md5-use in production code.
To keep this CL smaller, do not remove the default from UnixFS quite yet.
RELNOTES: None.
PiperOrigin-RevId: 206358838
|
|
|
|
|
|
|
|
|
| |
This forces the ex-default to be explicit in a lot of tests, but I'd rather that than have the risk of implicit md5-use in production code.
To keep this CL smaller, do not remove the default from UnixFS quite yet.
RELNOTES: None.
PiperOrigin-RevId: 206223521
|
|
|
|
| |
PiperOrigin-RevId: 204121958
|
|
|
|
| |
PiperOrigin-RevId: 202704472
|
|
|
|
|
|
|
|
| |
LinkedHashMultimap as well. Let ImmutableSetRuntimeCodec handle some other sets. We squash these extra types into the same codec because we want them to become immutable anyway, and they can be dealt with easily by the existing codecs.
Add some more non-Bazel classes that should be DynamicCodec'ed. Also, when a disallowed value is serialized, throw a SerializationException instead of crashing. This makes it easier to recover in tests.
PiperOrigin-RevId: 201237631
|
|
|
|
|
|
| |
default.
PiperOrigin-RevId: 201218341
|
|
|
|
|
|
| |
and have DynamicCodec add to it. It's very useful to have this trail available when debugging or whitelisting.
PiperOrigin-RevId: 201205884
|
|
|
|
| |
PiperOrigin-RevId: 201203706
|
|
|
|
|
|
| |
ConfiguredTargetValue is no longer prohibitively slow.
PiperOrigin-RevId: 200640971
|
|
|
|
| |
PiperOrigin-RevId: 200561008
|
|
|
|
|
|
| |
We could conceivably do some monkey-patching at server startup to make all lambdas act Serializable, but one step at a time.
PiperOrigin-RevId: 200509321
|
|
|
|
|
|
| |
not serialized. Tag TestCompletionValue and any ActionExecutionValue coming from a NotifyOnActionCacheHit (i.e., tests) like that. To make such values really not shared, request the ActionExecutionValue from TestCompletionFunction as opposed to the ArtifactValue (propagating the unshareable bit up seemed like too much fuss, and I have a dream of getting rid of ArtifactValue anyway).
PiperOrigin-RevId: 200504358
|
|
|
|
|
|
|
|
| |
on that future.
This allows NestedSet deserialization not to block on storage reads - in-progress deserializations are simply made a member of the NestedSets they produce.
PiperOrigin-RevId: 200440131
|
|
|
|
| |
PiperOrigin-RevId: 199965139
|
|
|
|
|
|
|
|
|
| |
When a LATIN-1 parameter file is requested, we can take advantage of the fact that JDK9 strings are (usually) stored as LATIN-1.
For UTF-8, we can still optimize for the common case where a LATIN-1 string contains only ASCII characters, as these are bit-identical between UTF-8 and LATIN-1. This would still be expected to be the vast majority of parameter file contents.
RELNOTES: None
PiperOrigin-RevId: 199816430
|
|
|
|
|
|
|
|
| |
class is allowed to be serialized in the current context. A codec can now add an explicitly allowed class that can be serialized underneath it (via SerializationContext#addExplicitlyAllowedClass), and that class's codec can check that it is explicitly allowed (via SerializationContext#checkClassExplicitlyAllowed). It is a runtime crash if a codec checks that it was explicitly allowed and finds that it wasn't. Thus, if PackageCodec is invoked without it having been explicitly allowed, we will crash, preventing Package from sneaking into a value it shouldn't be in.
This is only enabled if the codec is memoizing.
PiperOrigin-RevId: 199317936
|
|
|
|
|
|
| |
visibility in Skyframe classes.
PiperOrigin-RevId: 197665817
|
|
|
|
| |
PiperOrigin-RevId: 197508159
|
|
|
|
|
|
| |
longer worth it.
PiperOrigin-RevId: 197477783
|
|
|
|
| |
PiperOrigin-RevId: 197465288
|
|
|
|
|
|
| |
the String's underlying raw bytes. This avoids byte copying and UTF interpretation. Experiments show it is approximately 15(!) times faster than the naive StringCodec for serialization and 2 times faster for deserialization (10 times faster for non-ASCII strings).
PiperOrigin-RevId: 197441758
|
|
|
|
|
|
| |
(3-8x faster than StringCodec).
PiperOrigin-RevId: 196615858
|
|
|
|
|
|
|
|
| |
prevents Bazel from crashing on an ErrorInfo with an exception from
ActionExecutionFunction, which can contain an action, and therefore contain a
NestedSet.
PiperOrigin-RevId: 196533301
|
|
|
|
|
|
| |
and continue to use the embedded JDK as the default host_javabase.
PiperOrigin-RevId: 196471714
|
|
|
|
| |
PiperOrigin-RevId: 196310244
|
|
|
|
| |
PiperOrigin-RevId: 196127040
|
|
|
|
|
|
| |
backend writes for inner NestedSet serialization.
PiperOrigin-RevId: 195294676
|
|
|
|
| |
PiperOrigin-RevId: 195110169
|
|
|
|
| |
PiperOrigin-RevId: 195040539
|
|
|
|
| |
PiperOrigin-RevId: 195002908
|
|
|
|
| |
PiperOrigin-RevId: 194895469
|
|
|
|
|
|
| |
remote write should block on the provided future.
PiperOrigin-RevId: 194836516
|
|
|
|
| |
PiperOrigin-RevId: 194787067
|
|
|
|
| |
PiperOrigin-RevId: 194487570
|
|
|
|
|
|
| |
This should reduce memory consumption in NestedSet deserialization, which currently does not recycle Artifact instances.
PiperOrigin-RevId: 194083901
|
|
|
|
|
|
| |
serialization.
PiperOrigin-RevId: 193536486
|
|
|
|
|
|
|
|
|
|
|
|
| |
*** Reason for rollback ***
Looks like this is using more cpu than I realized. Doesn't seem fixable short-term. This may need to be cherrypicked into the next canary if it shows up in profiling. See bug for details.
*** Original change description ***
Add ValueConstants helper to handle the increasingly complex logic for value-equality-tested constants. Main hurdle is efficiently testing to see if a Collection is a value constant without trying to do work on Collections that can't possibly be value constants.
PiperOrigin-RevId: 193403226
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
*** Reason for rollback ***
Breaks //third_party/java_src/copybara/java/com/google/copybara/config:parser:
[]
*** Original change description ***
PiperOrigin-RevId: 193292991
|
| |
|
|
|
|
|
|
| |
value-equality-tested constants. Main hurdle is efficiently testing to see if a Collection is a value constant without trying to do work on Collections that can't possibly be value constants.
PiperOrigin-RevId: 193085625
|
|
|
|
|
|
| |
process-global bimap of fingerprints to NestedSet contents.
PiperOrigin-RevId: 193063717
|
| |
|
|
|
|
|
|
| |
This is required for moving --test_timeout from ExecutionOptions to BuildConfiguration.Options
PiperOrigin-RevId: 192815911
|
|
|
|
| |
PiperOrigin-RevId: 192662977
|
|
|
|
|
|
| |
when there may not be a canonical instance of the object we want (or the canonical instance isn't available at registry construction time), but we can reasonably cheaply do value equality comparisons. Strings are a good example.
PiperOrigin-RevId: 192354865
|
|
|
|
|
|
|
|
| |
constants. A lot of care is needed here because we're using reference equality. I plan to add value-equality constants in a follow-up.
Add ImmutableSortedSet marshaller because I think it might have been needed, and hey, why not.
PiperOrigin-RevId: 192326359
|
|
|
|
| |
PiperOrigin-RevId: 192325937
|