Observed this segfault running all tests on mater, occurring
in around 2-3 out of 10 test runs.
```
FAIL: //beacon-chain/sync:go_default_test (shard 3 of 4, run 1 of 10) (see /home/j/.cache/bazel/_bazel_j/1ba834ca9d49f27aeb8f0bbb6f28fdf3/execroot/prysm/bazel-out/k8-fastbuild/testlogs/beacon-chain/sync/go_default_test/shard_3_of_4_run_1_of_10/test.log)
INFO: From Testing //beacon-chain/sync:go_default_test (shard 3 of 4, run 1 of 10):
==================== Test output for //beacon-chain/sync:go_default_test (shard 3 of 4, run 1 of 10):
panic: runtime error: invalid memory address or nil pointer dereference
[signal SIGSEGV: segmentation violation code=0x1 addr=0x0 pc=0x138eea6]
goroutine 1660 [running]:
github.com/prysmaticlabs/prysm/shared/abool.(*AtomicBool).IsSet(...)
shared/abool/abool.go:39
github.com/prysmaticlabs/prysm/beacon-chain/sync.(*Service).subscribeStaticWithSubnets.func1(0xc002dd4400, 0xc002990940, 0x17bca26, 0x1e)
beacon-chain/sync/subscriber.go:207 +0xe6
created by github.com/prysmaticlabs/prysm/beacon-chain/sync.(*Service).subscribeStaticWithSubnets
beacon-chain/sync/subscriber.go:200 +0x172
================================================================================
```
TestStaticSubnets was testing a Service with an uninitialized
chainStarted value. This commit initializes chainStarted explicitly
in all tests that construct a Service. This reduces the observed flake
rate to 0/10 runs. This was verified with:
```
./bazel.sh test //beacon-chain/sync:go_default_test --runs_per_test 10
```
* Return error on invalid range and fix tests
* Uncomment some test codes
* Update comment
* Sync with master, fixed more tests
* Rm error condition, update comments, tests
* add --enable --disable flags for validator accounts
* refactor DeleteAccountConfig into AccountConfig to be used for enable and disable feature
* add `disable` flag for validator accounts
* [wip] add method to disable account
* refactor account delete
* add disable & enable with proper filters
* fix keymanager unit tests
* update DisabledPublicKeys to be a string instead of [][]byte
* fix FetchValidatingPrivateKeys to only fetch active keys with new string format
* fix FetchValidationPrivateKeys with new DisabledPublicKeys format (as a string)
* rename file + update AccountsConfig to include Disable, Enable and Delete distinct attributes
* rename accounts_activation -> accounts_enable_disable
* revert changes from using string to [][]byte for DisabledPublicKeys
* add FetchAllValidatingPublicKeys to preserve the functionality for accounts list, backup and delete
* fix unit tests
* convert publickeys from [][]byte to str before passing it to pb message
* add unit tests for disable keys
* add unit tests for EnableAccounts
* revert WORKSPACE LLM for now
* ran gazelle
* move function to convert KeymanagerOpts to Config inside rpc and run gazelle
* add unit tests for FetchAllValidatingPublicKeys
* fix keymanageropts for InteropKey
* Fix mistake for enable accounts
* add docstring to DisableAccountsCli and EnableAccountsCli
* remove previous testnet and add toledo & pyrmont
* Update eth1data params to double
* Update spec tests tags and state field for fssz gen
* Update more spec test sha tags
* Update slashing params
* Update slashing precompute to use config instead of hardcoded 3
* Update slashing test values due to config changes
* Update configs for slashedless test
* Go mod tidy
* Add toledo config (#7743)
* Update genesis delay to one week (#7782)
* Add Pyrmont config (#7797)
* Add Pyrmont config
* Fix config
* Update genesis time to the correct value
* Remove TestExecuteStateTransition_FullBlock
* Add back missing comments
* Update spectests to v1.0.0
Co-authored-by: Raul Jordan <raul@prysmaticlabs.com>
Co-authored-by: Preston Van Loon <preston@prysmaticlabs.com>
* Immediate save of validator protection data
* fix error log
* separate delete from save
* remove logs
* rename delete into reset
* comment fix
Co-authored-by: Raul Jordan <raul@prysmaticlabs.com>
Co-authored-by: prylabs-bulldozer[bot] <58059840+prylabs-bulldozer[bot]@users.noreply.github.com>
* Warn user for missing protection db
* better warning message
Co-authored-by: Raul Jordan <raul@prysmaticlabs.com>
Co-authored-by: prylabs-bulldozer[bot] <58059840+prylabs-bulldozer[bot]@users.noreply.github.com>
* enable head sync only during long period of non-finality
* Terence's suggestion
Co-authored-by: prylabs-bulldozer[bot] <58059840+prylabs-bulldozer[bot]@users.noreply.github.com>
* Ensure request epoch is used
* Update test
* Comment
Co-authored-by: Raul Jordan <raul@prysmaticlabs.com>
Co-authored-by: prylabs-bulldozer[bot] <58059840+prylabs-bulldozer[bot]@users.noreply.github.com>
* Fix up exits logic
* comments
* tests for malformed exits
* comment fix
* add yet another unit test, check pending list with binary search
* simplify
* add test for favoring earlier exit epoch
* gaz
* removal of superficial map check
* call LogValidatorGainsAndLosses at end of epoch
* Reviewer fixes
* Reviewer fixes
* Reviewer fixes
* Export Inclusion Distance to Prometheus
* changed default value to 1
* removed default value
* Added other performance metrics
* add slot
* get rid of inclusion_slot
* Fix fmt test
* Reviewer changes
Co-authored-by: Raul Jordan <raul@prysmaticlabs.com>
Co-authored-by: terence tsao <terence@prysmaticlabs.com>
* fix
* var
Co-authored-by: Raul Jordan <raul@prysmaticlabs.com>
Co-authored-by: Victor Farazdagi <simple.square@gmail.com>
Co-authored-by: prylabs-bulldozer[bot] <58059840+prylabs-bulldozer[bot]@users.noreply.github.com>
