Reproducible builds #28987
Labels
Comments
|
Actually, ignore reproducing the same build as the travis builder, we don't even reproduce the same build on the same system: root@208cb9fcfa68:/go-ethereum# rm ./build/bin/geth
root@208cb9fcfa68:/go-ethereum# CI=true TRAVIS=true TRAVIS_COMMIT="fe91d476ba3e29316b6dc99b6efd4a571481d888" go run ./build/ci.go install -dlgo ./cmd/geth
gotool.go:96: -dlgo version matches active Go version 1.21.6, skipping download.
>>> /usr/local/go/bin/go build -ldflags "-X github.com/ethereum/go-ethereum/internal/version.gitCommit=fe91d476ba3e29316b6dc99b6efd4a571481d888 -X github.com/ethereum/go-ethereum/internal/version.gitDate=20240213 -extldflags '-Wl,-z,stack-size=0x800000'" -tags urfave_cli_no_docs,ckzg -trimpath -v -o /go-ethereum/build/bin/geth ./cmd/geth
root@208cb9fcfa68:/go-ethereum# md5sum ./build/bin/geth
1337ffaed216a31fa9a77caf138f642f ./build/bin/geth
root@208cb9fcfa68:/go-ethereum# rm ./build/bin/geth
root@208cb9fcfa68:/go-ethereum# CI=true TRAVIS=true TRAVIS_COMMIT="fe91d476ba3e29316b6dc99b6efd4a571481d888" go run ./build/ci.go install -dlgo ./cmd/geth
gotool.go:96: -dlgo version matches active Go version 1.21.6, skipping download.
>>> /usr/local/go/bin/go build -ldflags "-X github.com/ethereum/go-ethereum/internal/version.gitCommit=fe91d476ba3e29316b6dc99b6efd4a571481d888 -X github.com/ethereum/go-ethereum/internal/version.gitDate=20240213 -extldflags '-Wl,-z,stack-size=0x800000'" -tags urfave_cli_no_docs,ckzg -trimpath -v -o /go-ethereum/build/bin/geth ./cmd/geth
root@208cb9fcfa68:/go-ethereum# md5sum ./build/bin/geth
4e5180c9678db91d506e223c9a25838a ./build/bin/geth |
|
If we disable the C building, then we get reliable builds on a single machine |
|
Got a report that these paths are present in the output: This works when imported as a library too I don't see these paths in the output binary |
|
Hi, Running this: wget https://gethstore.blob.core.windows.net/builds/geth-linux-amd64-1.13.15-c5ba367e.tar.gz
tar -xvf geth-linux-amd64-1.13.15-c5ba367e.tar.gz
cd geth-linux-amd64-1.13.15-c5ba367e
grep -a 'home/travis' geth | stringsI get four occurrences of full Travis paths in the bundle: Which are part of the read-only data.
Here are some files to reproduce more descriptive diffs using diffoscope. |
|
Right. And here's how it looks against a newer binary ( |
|
@vivi365 made a great finding here: golang/go#67011, Following that example, I did the same (but with First dockerfile, results in For
This is good, now it stripped the path So seems that particular bug is only present in ubuntu. We should bump the CI-builders. |
|
Thanks for bumping to noble, it seems to have fixed the trimpath issue! And a heads up; the dist being used might be Ubuntu focal, not noble.
Might be as they officially do not support noble (yet) https://docs.travis-ci.com/user/reference/linux/. |
|
To follow up a bit. I tried building with docker Now, if I build twice in a row, as Looking into more details, there are two causes for the diffs: The second is some sort of description data related to buildid. The upper hexdata is So, I made four builds. The first two do not have And the last two build-ids are almost identical, differing only in the last (fourth) part. |
|
More docs about the buildid: https://go.dev/src/cmd/go/internal/work/buildid.go |
|
This was a tough one! But after searching around, I found that I can get around the spurious path by adding the linker directive It shaves off 20 MB of data too, but it probably makes all stack traces suck (it might be that it only makes the C-side stack traces suck, I'm not sure). ( |
|
I tested a ctrl-c:ing it to yield a stack trace, and the (go-side) stacks do indeed appear to be readable. So using |
fjl
pushed a commit
that referenced
this issue
Aug 20, 2024
This PR implements the conclusions from #28987 (comment), that is: Building with `--strip-all` as a ld-flag to the cgo linker, to remove symbols. Without that, some spurious reference to a temporary file is included into the kzg-related library. Building with `--build-id=none`, to avoid putting a `build id` into the file.
ucwong
pushed a commit
to CortexFoundation/CortexTheseus
that referenced
this issue
Aug 20, 2024
This PR implements the conclusions from ethereum/go-ethereum#28987 (comment), that is: Building with `--strip-all` as a ld-flag to the cgo linker, to remove symbols. Without that, some spurious reference to a temporary file is included into the kzg-related library. Building with `--build-id=none`, to avoid putting a `build id` into the file.
|
So close now... Everything matches, except for the Comparing a locally built (docker on a debian) with the downloaded binary: |
holiman
added a commit
that referenced
this issue
Aug 23, 2024
The previous clearing of buildid did fully work, turns out we need to set it in `ldflags` The go buildid is the only remaining hurdle for reproducible builds, see #28987 (comment) This PR changes the go build id application note to say literally `none` golang/go#33772 (comment): > This difference is due to the .note.go.buildid section added by the linker. It can be set to something static e.g. -ldflags=-buildid= (empty string) to gain reproducibility.
|
Downloading the travis-built binary: Winding up with Now, using this With Run the docker, mount the Lo and behold: I think we're done! Thanks @vivi365 Next step would be to set up a secondary system which performs builds, and sounds an alarm if the builds fail to reproduce. |
|
Great result! Do you think we could make it reproduce without |
You are right, |
For the 386, the Ah, it seems there's still something not quite right -- I thought this was fixes by #30325 . Comparing the |
|
Make it unclean Bingo. Question is why the travis-build detects the vcs as modified |
|
That did it |
|
The rest (linux) also checks out now. The want: And the "have" |
|
Hm, oddly, though, sometimes when I build, the readelf output is identical. I used Ah, ok, the problem occurs if I install the arm-gcc stuff to early. This "ruins" the regular 386-builder: |
ucwong
pushed a commit
to CortexFoundation/CortexTheseus
that referenced
this issue
Aug 26, 2024
The previous clearing of buildid did fully work, turns out we need to set it in `ldflags` The go buildid is the only remaining hurdle for reproducible builds, see ethereum/go-ethereum#28987 (comment) This PR changes the go build id application note to say literally `none` golang/go#33772 (comment): > This difference is due to the .note.go.buildid section added by the linker. It can be set to something static e.g. -ldflags=-buildid= (empty string) to gain reproducibility.
mask-pp
pushed a commit
to mask-pp/go-ethereum
that referenced
this issue
Aug 30, 2024
The previous clearing of buildid did fully work, turns out we need to set it in `ldflags` The go buildid is the only remaining hurdle for reproducible builds, see ethereum#28987 (comment) This PR changes the go build id application note to say literally `none` golang/go#33772 (comment): > This difference is due to the .note.go.buildid section added by the linker. It can be set to something static e.g. -ldflags=-buildid= (empty string) to gain reproducibility.
davidtaikocha
added a commit
to taikoxyz/taiko-geth
that referenced
this issue
Oct 3, 2024
* params: release Geth v1.14.5 * params: begin v1.14.6 release cycle * cmd/evm/internal/t8ntool: remove unused parameter (#29930) * go.mod : tidy * cmd/clef, cmd/evm: fix markdown issues in README (#29954) * cmd/geth: remove unused param (#29952) * p2p/discover: add missing lock when calling tab.handleAddNode (#29960) * p2p: use package slices to sort in PeersInfo (#29957) * core: initialize developer genesis beacon root contract with 0 balance (#29963) * core, rlp: remove duplicated words (#29964) * cmd, core: prefetch reads too from tries if requested (#29807) * cmd/utils, consensus/beacon, core/state: when configured via stub flag: prefetch all reads from account/storage tries, terminate prefetcher synchronously. * cmd, core/state: fix nil panic, fix error handling, prefetch nosnap too * core/state: expand prefetcher metrics for reads and writes separately * cmd/utils, eth: fix noop collect witness flag --------- Co-authored-by: Péter Szilágyi <[email protected]> * core/state: rename all the AccessList receivers to 'al' (#29921) rename all the receivers to 'al' * ethconfig: regenerate config (#29970) * cmd/devp2p: fix log output (#29972) * .github: disable cache in actions run (#29926) * p2p/simulations: update doc of HTTP endpoints (#29894) * all: fix inconsistent receiver name and add lint rule for it (#29974) * .golangci.yml: enable check for consistent receiver name * beacon/light/sync: fix receiver name * core/txpool/blobpool: fix receiver name * core/types: fix receiver name * internal/ethapi: use consistent receiver name 'api' for handler object * signer/core/apitypes: fix receiver name * signer/core: use consistent receiver name 'api' for handler object * log: fix receiver name * accounts: avoid duplicate regex compilation (#29943) * fix: Optimize regular initialization * modify var name * variable change to private types * core/state, eth/protocols, trie, triedb/pathdb: remove unused error from trie Commit (#29869) * core/state, eth/protocols, trie, triedb/pathdb: remove unused error return from trie Commit * move set back to account-trie-update block scoping for easier readability * address review * undo tests submodule change * trie: panic if BatchSerialize returns an error in Verkle trie Commit * trie: verkle comment nitpicks --------- Co-authored-by: Péter Szilágyi <[email protected]> * beacon/light: fix shutdown issues (#29946) * beacon/light/request: add server test for event after unsubscribe * beacon/light/api: fixed double stream.Close() * beacon/light/request: add checks for nil event callback function * beacon/light/request: unlock server mutex while unsubscribing from parent * trie/triedb: add Reader to backend interface (#29988) * core/state/snapshot: add a missing lock (#30001) * upgrade lock usage * revert unnecessary change * go.mod: update Pebble to sort out a deleted upstream dependency (#30010) * log: fix some functions comments (#29907) updates some docstrings --------- Co-authored-by: rjl493456442 <[email protected]> * trie, triedb/pathdb: prealloc capacity for map and slice (#29986) * triedb/pathdb: use maps.Clone and maps.Keys (#29985) * common/math: fix out of bounds access in json unmarshalling (#30014) Co-authored-by: Martin Holst Swende <[email protected]> * core/state/snapshot: acquire the lock on Release (#30011) * core/state/snapshot: acquire the lock on release * core/state/snapshot: only acquire read-lock when iterating * cmd/geth, ethdb/pebble: improve database statistic (#29948) * cmd/geth, ethdb/pebble: polish method naming and code comment * implement db stat for pebble * cmd, core, ethdb, internal, trie: remove db property selector * cmd, core, ethdb: fix function description --------- Co-authored-by: prpeh <[email protected]> Co-authored-by: Gary Rong <[email protected]> * trie: don't reset tracer at the end of Commit (#30024) * trie: don't reset tracer at the end of Commit * Update trie.go --------- Co-authored-by: rjl493456442 <[email protected]> * common: using `ParseUint` instead of `ParseInt` (#30020) Since Decimal is defined as unsiged `uint64`, we should use `strconv.ParseUint` instead of `strconv.ParseInt` during unmarshalling. --------- Co-authored-by: Martin Holst Swende <[email protected]> * core/txpool/blobpool: change rw-lock to r-lock (#29989) * trie/trienode: avoid unnecessary copy (#30019) * avoid unnecessary copy * delete the never used function ProofList * eth/protocols/snap, trie/trienode: polish the code --------- Co-authored-by: Gary Rong <[email protected]> * p2p/rlpx: 2KB maximum size for handshake messages (#30029) Co-authored-by: Felix Lange <[email protected]> * core/state/snapshot: tiny fixes (#29995) * Revert "core/state/snapshot: tiny fixes" (#30039) Revert "core/state/snapshot: tiny fixes (#29995)" This reverts commit e0e45dbc32501d7917edb07083aa1c34ab7b0fb4. * p2p/discover: improve flaky revalidation tests (#30023) * cmd/blsync: use debug.Setup for logging configuration (#30065) * .github: add lightclient as codeowner to relevant packages (#30062) * accounts/keystore: use t.TempDir in test (#30052) * internal/debug: remove unnecessary log level assignment (#30044) Log level is specified in L259 so it's unnecessary to specify it for handlers (L234, L236). * all: stateless witness builder and (self-)cross validator (#29719) * all: add stateless verifications * all: simplify witness and integrate it into live geth --------- Co-authored-by: Péter Szilágyi <[email protected]> * core/txpool/blobpool: avoid use *map as parameter. (#30048) * trie/trienode: remove unnecessary check in Summary (#30047) * eth/tracers,trie: remove unnecessary check (#30071) * trie: relocate state execution logic into pathdb package (#29861) * triedb/pathdb: fix flaky test in pathdb (#29901) * core/txpool/blobpool: improve newPriceHeap function (#30050) Co-authored-by: Felix Lange <[email protected]> * cmd/evm/internal/t8ntool: log writeTraceResult error message (#30038) * all: replace division with right shift if possible (#29911) * rpc: truncate call error data logs (#30028) Co-authored-by: Felix Lange <[email protected]> * accounts/usbwallet/trezor: upgrade to generate with protoc 27.1 (#30058) * build: add check for stale generated files (#30037) Co-authored-by: Felix Lange <[email protected]> * core/state: fix inconsistent verkle test error messages (#29753) * accounts/abi: embed Go template instead of string literal (#30098) refactor(accounts/abi): use embed pkg to split default template to file * params: release Geth v1.14.6 * params: begin v1.14.7 release cycle * params: release Geth v1.14.6 * build: upgrade -dlgo version to Go 1.22.5 (#30112) * crypto: remove hardcoded value for secp256k1.N (#30126) * go.mod: update uint256 to 1.3.0 (#30134) * eth/catalyst: fix params in failure log (#30131) * core/txpool/blobpool: revert #29989, WLock on Nonce (#30142) * params: go-ethereum v1.14.7 stable * params: begin v1.14.8 release cycle * core/state: fix prefetcher for verkle (#29760) * core/txpool/blobpool: use nonce from argument instead of tx.Nonce() (#30148) This does not change the behavior here as the nonce in the argument is tx.Nonce(). This commit helps to make the function easier to read and avoid capturing the tx in the function. * trie: add RollBackAccount function to verkle trees (#30135) * p2p: fix ip change log parameter (#30158) * cmd/utils: fix typo in flag description (#30127) * core/types: don't modify signature V when reading large chainID (#30157) * SECURITY.md: correct PGP key block formatting (#30123) * all: simplify tests using t.TempDir() (#30150) * eth/catalyst: fix (*SimulatedBeacon).AdjustTime() conversion (#30138) * trie, triedb: remove unnecessary child resolver interface (#30167) * core/txpool/legacypool: use maps.Keys and maps.Copy (#30091) * core/state: don't compute verkle storage tree roots (#30130) * core/rawdb, triedb, cmd: create an isolated disk namespace for verkle (#30105) * core, triedb/pathdb, cmd: define verkle state ancient store * core/rawdb, triedb: add verkle namespace in pathdb * p2p/discover: remove type encPubkey (#30172) The pubkey type was moved to package v4wire a long time ago. Remaining uses of encPubkey were probably left in due to laziness. * go.mod: upgrade to btcsuite/btcd/btcec v2.3.4 (#30181) * ethdb: remove snapshot (#30189) * eth/gasprice: remove default from config (#30080) * eth/gasprice: remove default from config * eth/gasprice: sanitize startPrice * rpc: use stable object in notifier test (#30193) This makes the test resilient to changes of types.Header -- otherwise the test needs to be updated each time the header structure is modified. * core/state: remove useless metrics (#30184) Originally, these metrics were added to track the largest storage wiping. Since account self-destruction was deprecated with the Cancun fork, these metrics have become meaningless. * rpc: show more error detail for `invalidMessageError` (#30191) Here we add distinct error messages for network timeouts and JSON parsing errors. Note this specifically applies to HTTP connections serving a single RPC request. Co-authored-by: Felix Lange <[email protected]> * core/tracing: update latest release version (#30211) * core/txpool: use the cached address in ValidateTransactionWithState (#30208) The address recover is executed and cached in ValidateTransaction already. It's expected that the cached one is returned in ValidateTransaction. However, currently, we use the wrong function signer.Sender instead of types.Sender which will do all the address recover again. * core/state: check db error after intermediate call (#30171) This pull request adds an additional error check after statedb.IntermediateRoot, ensuring that no errors occur during this call. This step is essential, as the call might encounter database errors. * cmd/utils: allow configurating blob pool from flags (#30203) Currently, we have 3 flags to configure blob pool. However, we don't read these flags and set the blob pool configuration in eth config accordingly. This commit adds a function to check if these flags are provided and set blob pool configuration based on them. * core/state: fix SetStorage override behavior (#30185) This pull request fixes the broken feature where the entire storage set is overridden. Originally, the storage set override was achieved by marking the associated account as deleted, preventing access to the storage slot on disk. However, since #29520, this flag is also checked when accessing the account, rendering the account unreachable. A fix has been applied in this pull request, which re-creates a new state object with all account metadata inherited. * triedb/pathdb: print out all trie owner and hash information (#30200) This pull request explicitly prints out the full hash for debugging purpose. * beacon/types, cmd/devp2p, p2p/enr: clean up uses of fmt.Errorf (#30182) * eth/tracers, internal/ethapi: remove unnecessary map pointer in state override (#30094) * internal/ethapi: fix state override test (#30228) Looks like #30094 became a bit stale after #30185 was merged and now we have a stale ref to a state override object causing CI to fail on master. * p2p/nat: return correct port for ExtIP NAT (#30234) Return the actually requested external port instead of 0 in the AddMapping implementation for `--nat extip:<IP>`. * p2p: fix flaky test TestServerPortMapping (#30241) The test specifies `ListenAddr: ":0"`, which means a random ephemeral port will be chosen for the TCP listener by the OS. Additionally, since no `DiscAddr` was specified, the same port that is chosen automatically by the OS will also be used for the UDP listener in the discovery UDP setup. This sometimes leads to test failures if the TCP listener picks a free TCP port that is already taken for UDP. By specifying `DiscAddr: ":0"`, the UDP port will be chosen independently from the TCP port, fixing the random failure. See issue #29830. Verified using ``` cd p2p go test -c -race stress ./p2p.test -test.run=TestServerPortMapping ... 5m0s: 4556 runs so far, 0 failures ``` The issue described above can technically lead to sporadic failures on systems that specify a listen address via the `--port` flag of 0 while not setting `--discovery.port`. Since the default is using port `30303` and using a random ephemeral port is likely not used much to begin with, not addressing the root cause might be acceptable. * p2p/discover: schedule revalidation also when all nodes are excluded (#30239) ## Issue If `nextTime` has passed, but all nodes are excluded, `get` would return `nil` and `run` would therefore not invoke `schedule`. Then, we schedule a timer for the past, as neither `nextTime` value has been updated. This creates a busy loop, as the timer immediately returns. ## Fix With this PR, revalidation will be also rescheduled when all nodes are excluded. --------- Co-authored-by: lightclient <[email protected]> * miner: remove outdated comment (#30248) * eth/downloader: correct sync mode logging to show old mode (#30219) This PR fixes an issue in the setMode method of beaconBackfiller where the log message was not displaying the previous mode correctly. The log message now shows both the old and new sync modes. * all: remove deprecated protobuf dependencies (#30232) The package `github.com/golang/protobuf/proto` is deprecated in favor `google.golang.org/protobuf/proto`. We should update the codes to recommended package. Signed-off-by: Icarus Wu <[email protected]> * accounts/abi/bind: add accessList support to base bond contract (#30195) Adding the correct accessList parameter when calling a contract can reduce gas consumption. However, the current version only allows adding the accessList manually when constructing the transaction. This PR can provide convenience for saving gas. * internal/debug: remove memsize (#30253) Removing because memsize will very likely be broken by Go 1.23. See https://github.com/fjl/memsize/issues/4 * eth/downloader: gofmt (#30261) Fixes a regression introduced in https://github.com/ethereum/go-ethereum/pull/30219 * cmd/evm: don't overwrite sender account (#30259) Fixes #30254 It seems like the removed CreateAccount call is very old and not needed anymore. After removing it, setting a sender that does not exist in the state doesn't seem to cause an issue. * eth/catalyst: get params.ExcessBlobGas but check with params.BlobGasUsed (#30267) Seems it is checked with the wrong argument Signed-off-by: jsvisa <[email protected]> * params: remove unused les parameters (#30268) * core/vm/runtime: ensure tracer benchmark calls `OnTxStart` (#30257) The struct-based tracing added in #29189 seems to have caused an issue with the benchmark `BenchmarkTracerStepVsCallFrame`. On master we see the following panic: ```console BenchmarkTracerStepVsCallFrame panic: runtime error: invalid memory address or nil pointer dereference [signal SIGSEGV: segmentation violation code=0x2 addr=0x40 pc=0x1019782f0] goroutine 37 [running]: github.com/ethereum/go-ethereum/eth/tracers/js.(*jsTracer).OnOpcode(0x140004c4000, 0x0, 0x10?, 0x989680, 0x1, {0x101ea2298, 0x1400000e258}, {0x1400000e258?, 0x14000155928?, 0x10173020c?}, ...) /Users/matt/dev/go-ethereum/eth/tracers/js/goja.go:328 +0x140 github.com/ethereum/go-ethereum/core/vm.(*EVMInterpreter).Run(0x14000307da0, 0x140003cc0d0, {0x0, 0x0, 0x0}, 0x0) ... FAIL github.com/ethereum/go-ethereum/core/vm/runtime 0.420s FAIL ``` The issue seems to be that `OnOpcode` expects that `OnTxStart` has already been called to initialize the `env` value in the tracer. The JS tracer uses it in `OnOpcode` for the `GetRefund()` method. This patch resolves the issue by reusing the `Call` method already defined in `runtime_test.go` which correctly calls `OnTxStart`. * ethclient: support networkID in hex format (#30263) Some chains’ network IDs use hexadecimal such as Optimism ("0xa" instead of "10"), so when converting the string to big.Int, we cannot specify base 10; otherwise, it will encounter errors with hexadecimal network IDs. * core/vm: improved stack swap performance (#30249) This PR adds the methods `Stack.swap1..16()` that faster than `Stack.swap(1..16)`. Co-authored-by: lmittmann <[email protected]> * signer/core: improve performance of isPrimitiveTypeValid function (#30274) (#30277) Precomputes valid primitive types into a map to use for validation, thus removing sprintf. * core/vm: use uint64 in memory for indices everywhere (#30252) Consistently use `uint64` for indices in `Memory` and drop lots of type conversions from `uint64` to `int64`. --------- Co-authored-by: lmittmann <[email protected]> * build: upgrade -dlgo version to Go 1.22.6 (#30273) * tests: fix TransactionTest to actually run (#30272) Due to https://github.com/ethereum/tests/releases/tag/v10.1, the format of the TransactionTest changed, but it was not properly addressed, causing the test to pass unexpectedly. --------- Co-authored-by: Martin Holst Swende <[email protected]> * eth/downloader, core/types: take withdrawals-size into account in downloader queue (#30276) Fixes a slight miscalculation in the downloader queue, which was not accurately taking block withdrawals into account when calculating the size of the items in the queue * cmd/evm: fix evm basefee (#30281) fixes #30279 -- previously we did not use the basefee from the genesis, and instead the defaults were used from `runtime.go/setDefaults`-function * go.mod: update uint256 to 1.3.1 (#30280) Release notes: https://github.com/holiman/uint256/releases/tag/v1.3.1 * beacon/engine, consensus/beacon: use params.MaximumExtraDataSize instead of hard-coded value (#29721) Co-authored-by: Felix Lange <[email protected]> Co-authored-by: Marius van der Wijden <[email protected]> Co-authored-by: lightclient <[email protected]> * p2p/simulations: remove packages (#30250) Looking at the history of these packages over the past several years, there haven't been any meaningful contributions or usages: https://github.com/ethereum/go-ethereum/commits/master/p2p/simulations?before=de6d5976794a9ed3b626d4eba57bf7f0806fb970+35 Almost all of the commits are part of larger refactors or low-hanging-fruit contributions. Seems like it's not providing much value and taking up team + contributor time. * eth/protocols/snap: cleanup dangling account trie nodes due to incomplete storage (#30258) This pull request fixes #30229. During snap sync, large storage will be split into several pieces and synchronized concurrently. Unfortunately, the tradeoff is that the respective merkle trie of each storage chunk will be incomplete due to the incomplete boundaries. The trie nodes on these boundaries will be discarded, and any dangling nodes on disk will also be removed if they fall on these paths, ensuring the state healer won't be blocked. However, the dangling account trie nodes on the path from the root to the associated account are left untouched. This means the dangling account trie nodes could potentially stop the state healing and break the assumption that the entire subtrie should exist if the subtrie root exists. We should consider the account trie node as the ancestor of the corresponding storage trie node. In the scenarios described in the above ticket, the state corruption could occur if there is a dangling account trie node while some storage trie nodes are removed due to synchronization redo. The fixing idea is pretty straightforward, the trie nodes on the path from root to account should all be explicitly removed if an incomplete storage trie occurs. Therefore, a `delete` operation has been added into `gentrie` to explicitly clear the account along with all nodes on this path. The special thing is that it's a cross-trie clearing. In theory, there may be a dangling node at any position on this account key and we have to clear all of them. * params: release go-ethereum v1.14.8 stable * params: begin v1.14.9 release cycle * go.mod: remove github.com/julienschmidt/httprouter (#30290) * build: run 'go mod tidy' check as part of lint (#30291) * core/txpool/blobpool: fix error message (#30247) the validation process only checks for 'less than', which is inconsistent with the error output * go.mod: upgrade to pebble v1.1.2 (#30297) Includes a fix for MIPS32 support. Pebble release: https://github.com/cockroachdb/pebble/releases/tag/v1.1.2 Key fix for mips32: https://github.com/cockroachdb/pebble/commit/9f3904a705d60b9832febb6c6494183d92c8f556 (also the only change from v1.1.1. * core: only compute state root once (#30299) This PR refactors the genesis initialization a bit, s.th. we only compute the blockhash once instead of twice as before (during hashAlloc and flushAlloc) This will significantly reduce the amount of memory allocated during genesis init --------- Co-authored-by: Gary Rong <[email protected]> * .golangci.yml: remove lint warning for TxLookupLimit * eth/fetcher: always expect transaction metadata in announcement (#30288) This pull request drops the legacy transaction retrieval support from before eth68, adding the restrictions that transaction metadata must be provided along with the transaction announment. * eth/ethconfig: remove LES server config (#30298) * eth/tracers/js: add coinbase addr to ctx (#30231) Add coinbase address to javascript tracer context. This PR adds the `coinbase` address to `jsTracer.ctx`, allowing access to the coinbase address (fee receipient) in custom JavaScript tracers. Example usage: ```javascript result: function(ctx) { return toAddress(ctx.coinbase); } ``` This change enables custom tracers to access coinbase address, previously unavailable, enhancing their capabilities to match built-in tracers. * eth: dial nodes from discv5 (#30302) Here I am adding a discv5 nodes source into the p2p dial iterator. It's an improved version of #29533. Unlike discv4, the discv5 random nodes iterator will always provide full ENRs. This means we can apply filtering to the results and will only try dialing nodes which explictly opt into the eth protocol with a matching chain. I have also removed the dial iterator from snap. We don't have an official DNS list for snap anymore, and I doubt anyone else is running one. While we could potentially filter for snap on discv5, there will be very few nodes announcing it, and the extra iterator would just stall the dialer. --------- Co-authored-by: lightclient <[email protected]> * beacon/light: handle endpoint URL more gracefully (#30306) blsync was failing if the light endpoint it was provided ended with a `/`. This change should handle the joining more gracefully. * core: remove withdrawal length check for state processor (#30286) The withdrawal length is already verified by the beacon consensus package, so the check in the state processor is a duplicate. * vm: simplify error handling in `vm.EVM.create()` (#30292) To allow all error paths in `vm.EVM.create()` to consume the necessary gas, there is currently a pattern of gating code on `if err == nil` instead of returning as soon as the error occurs. The same behaviour can be achieved by abstracting the gated code into a method that returns immediately on error, improving readability and thus making it easier to understand and maintain. * internal/build: include git-date on detached head (#30320) When we are building in detached head, we cannot easily obtain the same information as we can if we're in non-detached head. However, one thing we _can_ obtain is the git-hash and git-date. Currently, we omit to include the git-date into the build-info, which causes problem for reproducable builds which are on a detached head. This change fixes it to include the date-info always. * build: remove mantic from ppa builds (#30322) removes ppa-build for ubuntu `mantic` * gitignore: ignore upload-artefacts (#30325) Our `WriteArchive`, used by ci builder, creates files in the repo root,in order to upload. After we've built the amd64-builds, we create the uploads, and cause the repo to be flagged as dirty for the remaining builds. This change fixes it by adding the artefacts to gitignore. Closes #30324 * eth/catalyst: ensure period zero mode leaves no pending txs in pool (#30264) closes #29475, replaces #29657, #30104 Fixes two issues. First is a deadlock where the txpool attempts to reorg, but can't complete because there are no readers left for the new txs subscription. Second, resolves a problem with on demand mode where txs may be left pending when there are more pending txs than block space. Co-authored-by: Martin Holst Swende <[email protected]> * accounts/abi: handle ABIs with contract type parameter (#30315) convert parameter of type contract to the basic `address` type --------- Co-authored-by: Martin HS <[email protected]> * core/rawdb: drop MigrateTable (#30331) These are the leftovers from #24028. * core/vm: reuse Memory instances (#30137) This PR adds a sync.Pool to reuse instances of Memory in EVMInterpreter. * build: attempt at reproducible builds (#30321) This PR implements the conclusions from https://github.com/ethereum/go-ethereum/issues/28987#issuecomment-2296075028, that is: Building with `--strip-all` as a ld-flag to the cgo linker, to remove symbols. Without that, some spurious reference to a temporary file is included into the kzg-related library. Building with `--build-id=none`, to avoid putting a `build id` into the file. * all: update to go version 1.23.0 (#30323) This PR updates the version of go used in builds and docker to 1.23.0. Release notes: https://go.dev/doc/go1.23 More importantly, following our policy of maintaining the last two versions (which now becomes 1.23 and 1.22), we can now make use of the things that were introduced in 1.22: https://go.dev/doc/go1.22 Go 1.22 makes two changes to “for” loops. - each iteration creates new variables, - for loops may range over integers Other than that, some interesting library changes and other stuff. * rpc: add timeout to rpc client Unsubscribe (#30318) Fixes #30156 This adds a repro of the linked issue. I fixed it by adding a timeout when issuing the call to unsubscribe. * cmd/devp2p: require dns:read, dns:edit permissions for cloudflare deploy (#30326) This PR adds the `dns:read` and `dns:edit` permissions to the required set of permissions checked before deploying an ENR tree to Cloudflare. These permissions are necessary for a successful publish. **Background**: The current logic for `devp2p dns to-cloudflare` checks for `zone:edit` and `zone:read` permissions. However, when running the command with only these two permissions, the following error occurs: ``` wrong permissions on zone REMOVED-ZONE: map[#zone:edit:false #zone:read:true] ``` Adding `zone:read` and `zone:edit` to the API token led to a different error: ``` INFO [08-19|14:06:16.782] Retrieving existing TXT records on pos-nodes.hardfork.dev Authentication error (10000) ``` This suggested that additional permissions were required. I added `dns:read`, but encountered another error: ``` INFO [08-19|14:11:42.342] Retrieving existing TXT records on pos-nodes.hardfork.dev INFO [08-19|14:11:42.851] Updating DNS entries failed to publish REMOVED.pos-nodes.hardfork.dev: Authentication error (10000) ``` Finally, after adding both `dns:read` and `dns:edit` permissions, the command executed successfully with the following output: ``` INFO [08-19|14:13:07.677] Checking Permissions on zone REMOVED-ZONE INFO [08-19|14:13:08.014] Retrieving existing TXT records on pos-nodes.hardfork.dev INFO [08-19|14:13:08.440] Updating DNS entries INFO [08-19|14:13:08.440] "Updating pos-nodes.hardfork.dev from \"enrtree-root:v1 e=FSED3EDKEKRDDFMCLP746QY6CY l=FDXN3SN67NA5DKA4J2GOK7BVQI seq=1 sig=Glja2c9RviRqOpaaHR0MnHsQwU76nJXadJwFeiXpp8MRTVIhvL0LIireT0yE3ETZArGEmY5Ywz3FVHZ3LR5JTAE\" to \"enrtree-root:v1 e=AB66M4ULYD5OYN4XFFCPVZRLUM l=FDXN3SN67NA5DKA4J2GOK7BVQI seq=1 sig=H8cqDzu0FAzBplK4g3yudhSaNtszIebc2aj4oDm5a5ZE5PAg-xpCnQgVE_53CsgsqQpalD9byafx_FrUT61sagA\"" INFO [08-19|14:13:16.932] Updated DNS entries new=32 updated=1 untouched=100 INFO [08-19|14:13:16.932] Deleting stale DNS entries INFO [08-19|14:13:24.663] Deleted stale DNS entries count=31 ``` With this PR, the required permissions for deploying an ENR tree to Cloudflare now include `zone:read`, `zone:edit`, `dns:read`, and `dns:edit`. The initial check now includes all of the necessary permissions and indicates in the error message which permissions are missing: ``` INFO [08-19|14:17:20.339] Checking Permissions on zone REMOVED-ZONE wrong permissions on zone REMOVED-ZONE: map[#dns_records:edit:false #dns_records:read:false #zone:edit:false #zone:read:true] ``` * all: clean up goerli flag and config (#30289) Co-authored-by: lightclient <[email protected]> * cmd/utils,p2p: enable discv5 by default (#30327) * travis.yml: use focal for builds (#30319) * trie: use go-verkle helper for speedier (*VerkleTrie).RollBackAccount (#30242) This is a performance improvement on the account-creation rollback code required for the archive node to support verkle. It uses the utility function `DeleteAtStem` to remove code and account data per-group instead of doing it leaf by leaf. It also fixes an index bug, as code is chunked in 31-byte chunks, so comparing with the code size should use 31 as its stride. --------- Co-authored-by: Felix Lange <[email protected]> * eth/protocols/eth: handle zero-count header requests (#30305) Proper fix for handling `count=0` get header requests. https://en.wikipedia.org/wiki/Count_Zero * eth/tracers: avoid panic in state test runner (#30332) Make tracers more robust by handling `nil` receipt as input. Also pass in a receipt with gas used in the state test runner. Closes https://github.com/ethereum/go-ethereum/issues/30117. --------- Co-authored-by: Sina Mahmoodi <[email protected]> * build: fix hash for go1.23.0.linux-riscv64.tar.gz (#30335) build: fix hash for go1.23.0.linux-riscv64.tar.gz * build: make go buildid static (#30342) The previous clearing of buildid did fully work, turns out we need to set it in `ldflags` The go buildid is the only remaining hurdle for reproducible builds, see https://github.com/ethereum/go-ethereum/issues/28987#issuecomment-2306412590 This PR changes the go build id application note to say literally `none` https://github.com/golang/go/issues/33772#issuecomment-528176001: > This difference is due to the .note.go.buildid section added by the linker. It can be set to something static e.g. -ldflags=-buildid= (empty string) to gain reproducibility. * trie: avoid un-needed map copy (#30343) This change avoids the an unnecessary map copy if the preimage recording is not enabled. * beacon/blsync: better error information in test (#30336) this change reports the error instead of ignoring it * beacon/light/sync: basic tests for rangeLock (#30269) adds simple tests for lock and firstUnlocked method from rangeLock type --------- Co-authored-by: lightclient <[email protected]> * build: debug travis build (#30344) debugging travis build pipeline * gitignore: ignore build signatures (#30346) Ignore files are generated during signing of download-binaries, which 'dirty' the vcs for subsequent builds. * doc: update 2021-08-22-split-postmortem (#30351) Update 2021-08-22-split-postmortem * core: implement EIP-2935 (#29465) https://eips.ethereum.org/EIPS/eip-2935 --------- Co-authored-by: Guillaume Ballet <[email protected]> Co-authored-by: Ignacio Hagopian <[email protected]> Co-authored-by: Martin HS <[email protected]> * core: add metrics for state access (#30353) This pull request adds a few more performance metrics, specifically: - The average time cost of an account read - The average time cost of a storage read - The rate of account reads - The rate of storage reads * core/state: fix trie prefetcher for verkle (#30354) This pull request fixes the panic issue in prefetcher once the verkle is activated. * p2p/discover: fix Write method in metered connection (#30355) `WriteToUDP` was never called, since `meteredUdpConn` exposed directly all the methods from the underlying `UDPConn` interface. This fixes the `discover/egress` metric never being updated. * accounts/abi/bind, ethclient/simulated: check SendTransaction error in tests (#30349) In few tests the returned error from `SendTransaction` is not being checked. This PR checks the returned err in tests. Returning errors also revealed tx in `TestCommitReturnValue` is not actually being sent, and returns err ` only replay-protected (EIP-155) transactions allowed over RPC`. Fixed the transaction by using the `testTx` function. * core/state: semantic journalling (part 1) (#28880) This is a follow-up to #29520, and a preparatory PR to a more thorough change in the journalling system. ### API methods instead of `append` operations This PR hides the journal-implementation details away, so that the statedb invokes methods like `JournalCreate`, instead of explicitly appending journal-events in a list. This means that it's up to the journal whether to implement it as a sequence of events or aggregate/merge events. ### Snapshot-management inside the journal This PR also makes it so that management of valid snapshots is moved inside the journal, exposed via the methods `Snapshot() int` and `RevertToSnapshot(revid int, s *StateDB)`. ### SetCode JournalSetCode journals the setting of code: it is implicit that the previous values were "no code" and emptyCodeHash. Therefore, we can simplify the setCode journal. ### Selfdestruct The self-destruct journalling is a bit strange: we allow the selfdestruct operation to be journalled several times. This makes it so that we also are forced to store whether the account was already destructed. What we can do instead, is to only journal the first destruction, and after that only journal balance-changes, but not journal the selfdestruct itself. This simplifies the journalling, so that internals about state management does not leak into the journal-API. ### Preimages Preimages were, for some reason, integrated into the journal management, despite not being a consensus-critical data structure. This PR undoes that. --------- Co-authored-by: Gary Rong <[email protected]> * signer/core/apitypes: support fixed size arrays for EIP-712 typed data (#30175) When attempting to hash a typed data struct that includes a type reference with a fixed-size array, the validation process fails. According to EIP-712, arrays can be either fixed-size or dynamic, denoted by `Type[n]` or `Type[]` respectively, although it appears this currently isn't supported. This change modifies the validation logic to accommodate types containing fixed-size arrays. * consensus/beacon, core/types: add verkle witness builder (#30129) This PR adds the bulk verkle witness+proof production at the end of block production. It reads all data from the tree in one swoop and produces a verkle proof. Co-authored-by: Felix Lange <[email protected]> * trie, core/state: Nyota EIP-6800 & EIP-4762 spec updates (#30357) This PR implements changes related to [EIP-6800](https://eips.ethereum.org/EIPS/eip-6800) and [EIP-4762](https://eips.ethereum.org/EIPS/eip-4762) spec updates. A TL;DR of the changes is that `Version`, `Balance`, `Nonce` and `CodeSize` are encoded in a single leaf named `BasicData`. For more details, see the [_Header Values_ table in EIP-6800](https://eips.ethereum.org/EIPS/eip-6800#header-values). The motivation for this was simplifying access event patterns, reducing code complexity, and, as a side effect, saving gas since fewer leaf nodes must be accessed. --------- Co-authored-by: Guillaume Ballet <[email protected]> Co-authored-by: Felix Lange <[email protected]> * Include tracerConfig in created tracing test (#30364) Fixes the tracer test filler for when there is tracerConfig. * core/state: pull the verkle trie from prefetcher for empty storage root (#30369) This pull request fixes a flaw in prefetcher. In verkle tree world, both accounts and storage slots are committed into a single tree instance for state hashing. If the prefetcher is activated, we will try to pull the trie for the prefetcher for performance speedup. However, we had a special logic to skip pulling storage trie if the storage root is empty. While it's true for merkle as we have nothing to do with an empty storage trie, it's totally wrong for verkle. The consequences for skipping pulling is the storage changes are committed into trie A, while the account changes are committed into trie B (pulled from the prefetcher), boom. * funding.json: add funding information file (#30385) Adds a list of funding identifiers. * all: implement EIP-6110, execution layer triggered deposits (#29431) This PR implements EIP-6110: Supply validator deposits on chain. It also sketches out the base for Prague in the engine API types. * all: remove forkchoicer and reorgNeeded (#29179) This PR changes how sidechains are handled. Before the merge, it was possible to import a chain with lower td and not set it as canonical. After the merge, we expect every chain that we get via InsertChain to be canonical. Non-canonical blocks can still be inserted with InsertBlockWIthoutSetHead. If during the InsertChain, the existing chain is not canonical anymore, we mark it as a sidechain and send the SideChainEvents normally. * core: fix compilation error (#30394) un-borks a compilation error from a recent merge to master * all: remove funding verifier (#30391) Now that verification is done, we can remove the funding information. * node: fix flaky jwt-test (#30388) This PR fixes a flaky jwt-test. The test is a jwt "from one second in the future". The test passes; the reason for this is that the CI-system is slow, and by the time the jwt is actually evaluated, that second has passed, and it's no longer future. Alternative to #30380 * build: increase go test timeout (#30398) This increases the timeout for the go tests on ci, this should prevent travis from erroring. see: https://app.travis-ci.com/github/ethereum/go-ethereum/jobs/625803693 * core/state: state reader abstraction (#29761) This pull request introduces a state.Reader interface for state accessing. The interface could be implemented in various ways. It can be pure trie only reader, or the combination of trie and state snapshot. What's more, this interface allows us to have more flexibility in the future, e.g. the archive reader (for accessing archive state). Additionally, this pull request removes the following metrics - `chain/snapshot/account/reads` - `chain/snapshot/storage/reads` * core/state: get rid of field pointer in journal (#30361) This pull request replaces the field pointer in journal entry with the field itself, specifically the address of mutated account. While it will introduce the extra allocation cost, but it's easier for code reading. Let's measure the overhead overall to see if the change is acceptable or not. * build: upgrade -dlgo version to Go 1.23.1 (#30404) New security fix: https://groups.google.com/g/golang-announce/c/K-cEzDeCtpc * internal/ethapi: eth_multicall (#27720) This is a successor PR to #25743. This PR is based on a new iteration of the spec: https://github.com/ethereum/execution-apis/pull/484. `eth_multicall` takes in a list of blocks, each optionally overriding fields like number, timestamp, etc. of a base block. Each block can include calls. At each block users can override the state. There are extra features, such as: - Include ether transfers as part of the logs - Overriding precompile codes with evm bytecode - Redirecting accounts to another address ## Breaking changes This PR includes the following breaking changes: - Block override fields of eth_call and debug_traceCall have had the following fields renamed - `coinbase` -> `feeRecipient` - `random` -> `prevRandao` - `baseFee` -> `baseFeePerGas` --------- Co-authored-by: Gary Rong <[email protected]> Co-authored-by: Martin Holst Swende <[email protected]> * eth/fetcher: fix blob transaction propagation (#30125) This PR fixes an issue with blob transaction propagation due to the blob transation txpool rejecting transactions with gapped nonces. The specific changes are: - fetch transactions from a peer in the order they were announced to minimize nonce-gaps (which cause blob txs to be rejected - don't wait on fetching blob transactions after announcement is received, since they are not broadcast Testing: - unit tests updated to reflect that fetch order should always match tx announcement order - unit test added to confirm blob transactions are scheduled immediately for fetching - running the PR on an eth mainnet full node without incident so far --------- Signed-off-by: Roberto Bayardo <[email protected]> Co-authored-by: Gary Rong <[email protected]> * core/state/snapshot: port changes from 29995 (#30040) #29995 has been reverted due to an unexpected flaw in the state snapshot process. Specifically, it attempts to stop the state snapshot generation, which could potentially cause the system to halt if the generation is not currently running. This pull request ports the changes made in #29995 and fixes the flaw. * beacon/engine/types: remove PayloadV4 (#30415) h/t @MariusVanDerWijden for finding and fixing this on devnet 3. I made the mistake of thinking `PayloadVersion` was correlated with the `GetPayloadVX` method, but it actually tracks which version of `PayloadAttributes` were passed to `forkchoiceUpdated`. So far, Prague does not necessitate a new version of fcu, so there is no need for `PayloadV4`. Co-authored-by: Marius van der Wijden <[email protected]> * core/vm: remove panic when address is not present (#30414) Remove redundant address presence check in `makeGasSStoreFunc`. This PR simplifies the `makeGasSStoreFunc` function by removing the redundant check for address presence in the access list. The updated code now only checks for slot presence, streamlining the logic and eliminating unnecessary panic conditions. This change removes the unnecessary address presence check, simplifying the code and improving maintainability without affecting functionality. The previous panic condition was intended as a canary during the testing phases (i.e. _YOLOv2_) and is no longer needed. * beacon/light/api: fixed blsync update query (#30421) This PR fixes what https://github.com/ethereum/go-ethereum/pull/30306/ broke. Escaping the `?` in the event sub query was fixed in that PR but it was still escaped in the `updates` request. This PR adds a URL params argument to `httpGet` and fixes `updates` query formatting. * eth/filters: prevent concurrent access in test (#30401) use a mutex to prevent concurrent access to the api.filters map during `TestPendingTxFilterDeadlock` test * core/rawdb: more accurate description of freezer in docs (#30393) fixes https://github.com/ethereum/go-ethereum/issues/29793 * core/state, core/vm: Nyota contract create init simplification (#30409) Implementation of [this EIP-4762 update](https://github.com/ethereum/EIPs/pull/8867). --------- Signed-off-by: Guillaume Ballet <[email protected]> Co-authored-by: Tanishq Jasoria <[email protected]> * p2p/enode: add quic ENR entry (#30283) Add `quic` entry to the ENR as proposed in https://github.com/ethereum/consensus-specs/pull/3644 --------- Co-authored-by: lightclient <[email protected]> * core/tracing: fix copy/paste error+comments in reason listing (#30431) Signed-off-by: Guillaume Ballet <[email protected]> * core/txpool/blobpool: avoid possible zero index panic (#30430) This situation(`len(txs) == 0`) rarely occurs, but if it does, it will panic. --------- Co-authored-by: Martin HS <[email protected]> * core/rawdb: remove unused transition status state accessors (#30433) * internal: run tests in parallel (#30381) Continuation of https://github.com/ethereum/go-ethereum/pull/28546 * core/types: more easily extensible tx signing (#30372) This change makes the code slightly easier for downstream-projects to extend with more signer-types, but if functionalily equivalent to the previous code. * core, trie: prealloc capacity for maps (#30437) - preallocate capacity for map - avoid `reinject` adding empty value - use `maps.Copy` * core/tracing: fix typo in comment (#30443) minor fix * core/tracing: add verkle gas change reasons to changelog (#30444) Add changes from #30409 and #29338 to changelog. --------- Co-authored-by: Martin HS <[email protected]> Co-authored-by: Guillaume Ballet <[email protected]> * Revert "core/rawdb: remove unused transition status state accessors" (#30449) Reverts ethereum/go-ethereum#30433 * params: release go-ethereum v1.14.9 stable (#30455) * params: begin v1.14.10 release cycle (#30457) * genesis: fix dev mode alloc (#30460) Balance being null causes `getGenesisState` to fail as the balance field is required in json marshaling of an account. * core: minor fix for the log wrapper with debug purpose (#30454) After this PR, https://github.com/ethereum/go-ethereum/pull/28187, the way to set the default logger is different. This PR only updates the way to set logger in some test cases' comments that existed in the codebase (since this commit https://github.com/ethereum/go-ethereum/commit/b63e3c37a6). Although I am not sure if it a good way to leave the code in the comment, it truly makes me more efficiently to debug and fix the failing test cases. * ethdb/pebble: handle errors (#30367) * .github: add release maintainers to params/ CODEOWNERS (#30458) * build: fix macos builds by working around travis osx flaw (#30479) This should fix https://github.com/ethereum/go-ethereum/issues/30471. See investigation in https://github.com/ethereum/go-ethereum/pull/30478 for more background. * beacon, core, eth, miner: integrate witnesses into production Geth (#30069) This PR integrates witness-enabled block production, witness-creating payload execution and stateless cross-validation into the `engine` API. The purpose of the PR is to enable the following use-cases (for API details, please see next section): - Cross validating locally created blocks: - Call `forkchoiceUpdatedWithWitness` instead of `forkchoiceUpdated` to trigger witness creation too. - Call `getPayload` as before to retrieve the new block and also the above created witness. - Call `executeStatelessPayload` against another client to cross-validate the block. - Cross validating locally processed blocks: - Call `newPayloadWithWitness` instead of `newPayload` to trigger witness creation too. - Call `executeStatelessPayload` against another client to cross-validate the block. - Block production for stateless clients (local or MEV builders): - Call `forkchoiceUpdatedWithWitness` instead of `forkchoiceUpdated` to trigger witness creation too. - Call `getPayload` as before to retrieve the new block and also the above created witness. - Propagate witnesses across the consensus libp2p network for stateless Ethereum. - Stateless validator validation: - Call `executeStatelessPayload` with the propagated witness to statelessly validate the block. *Note, the various `WithWitness` methods could also *just be* an additional boolean flag on the base methods, but this PR wanted to keep the methods separate until a final consensus is reached on how to integrate in production.* --- The following `engine` API types are introduced: ```go // StatelessPayloadStatusV1 is the result of a stateless payload execution. type StatelessPayloadStatusV1 struct { Status string `json:"status"` StateRoot common.Hash `json:"stateRoot"` ReceiptsRoot common.Hash `json:"receiptsRoot"` ValidationError *string `json:"validationError"` } ``` - Add `forkchoiceUpdatedWithWitnessV1,2,3` with same params and returns as `forkchoiceUpdatedV1,2,3`, but triggering a stateless witness building if block production is requested. - Extend `getPayloadV2,3` to return `executionPayloadEnvelope` with an additional `witness` field of type `bytes` iff created via `forkchoiceUpdatedWithWitnessV2,3`. - Add `newPayloadWithWitnessV1,2,3,4` with same params and returns as `newPayloadV1,2,3,4`, but triggering a stateless witness creation during payload execution to allow cross validating it. - Extend `payloadStatusV1` with a `witness` field of type `bytes` if returned by `newPayloadWithWitnessV1,2,3,4`. - Add `executeStatelessPayloadV1,2,3,4` with same base params as `newPayloadV1,2,3,4` and one more additional param (`witness`) of type `bytes`. The method returns `statelessPayloadStatusV1`, which mirrors `payloadStatusV1` but replaces `latestValidHash` with `stateRoot` and `receiptRoot`. * travis: work around travis/osx/go1.23 setup bug (#30491) This is a work-around for a strange issue with travis, specifically, `os=osx, go: 1.23.1`. When this is used, the actual go that ends up being used is `go1.19.4 darwin/amd64 `. Using `which go`, it told me that the `go` in the path was a softlink at `/Users/travis/gopath/bin/go1.23.1 `. However, this was not true: using `command -v go`, it told me that the actual `go` that was used is a softlink at `/usr/local/bin/go`. This change rewrites the `/usr/local/bin/go` softlink to point to the binary at `/Users/travis/gopath/bin/go1.23.1`, so we get the right go-version. * cmd/utils: fix `setEtherbase` (#30488) Make `setEtherbase` fall thorugh and handle `miner.pending.feeRecipient` after showing deprecation-warning for `miner.etherbase`-flag. * core/state: fix comment of `mode` (#30490) * core/state: commit snapshot only if the base layer exists (#30493) This pull request skips the state snapshot update if the base layer is not existent, eliminating the numerous warning logs after an unclean shutdown. Specifically, Geth will rewind its chain head to a historical block after unclean shutdown and state snapshot will be remained as unchanged waiting for recovery. During this period of time, the snapshot is unusable and all state updates should be ignored/skipped for state snapshot update. * internal/ethapi/api: for simulated calls, set gaspool to max value if global gascap is 0 (#30474) In #27720, we introduced RPC global gas cap. A value of `0` means an unlimited gas cap. However, this was not the case for simulated calls. This PR fixes the behaviour. * core/rawdb: make sure specified state scheme is valid (#30499) This change exits with error if user provided a `--state.scheme` which is neither `hash` nor `path` * feat(repo): `geth/v1.14.9` upstream merge * internal/ethapi: fix gascap 0 for eth_simulateV1 (#30496) Similar to #30474. * core/tracing, core/vm: add ContractCode to the OpContext (#30466) Extends the opcontext interface to include accessor for code being executed in current context. While it is possible to get the code via `statedb.GetCode`, that approach doesn't work for initcode. * core/vm: more benchmarks for bls g1/g2-multiexp precompiles (#30459) This change adds more comprehensive benchmarks with a wider-variety of input sizes for g1 and g2 multi exponentiation. * p2p/discover: fix flaky tests writing to test.log after completion (#30506) This PR fixes two tests, which had a tendency to sometimes write to the `*testing.T` `log` facility after the test function had completed, which is not allowed. This PR fixes it by using waitgroups to ensure that the handler/logwriter terminates before the test exits. closes #30505 * deps: update supranational/blst (#30504) This update should only affect the fuzzers, as far as I know. But it seems like it might also fix some arm/macos compilation issue in https://github.com/ethereum/go-ethereum/issues/30494 Closes #30494 (I think) * core/txpool, eth/catalyst: ensure gas tip retains current value upon rollback (#30495) Here we move the method that drops all transactions by temporarily increasing the fee into the TxPool itself. It's better to have it there because we can set it back to the configured value afterwards. This resolves a TODO in the simulated backend. * feat(repo): Fix bug merge 1.14.9 (#320) * fix lint * fix bug * update generation files * core/txpool/blobpool: revert part of #30437, return all reinject-addresses * core/txpool/blobpool: add test to check internal shuffling * Revert "core/txpool, eth/catalyst: ensure gas tip retains current value upon rollback" (#30521) Reverts ethereum/go-ethereum#30495 You are free to create a proper Clear method if that's the best way. But one that does a proper cleanup, not some hacky call to set gas which screws up logs, metrics and everything along the way. Also doesn't work for legacy pool local transactions. The current code had a hack in the simulated code, now we have a hack in live txpooling code. No, that's not acceptable. I want the live code to be proper, meaningful API, meaningful comments, meaningful implementation. * params: release Geth v1.14.10 * params: begin v1.14.11 release cycle * feat: merge 1.14.10 * fix(taiko): Fix bug merge 1.14.9 (#325) * fix bug * fix bug * p2p/discover: add config option for disabling FINDNODE liveness check (#30512) This is for fixing Prysm integration tests. * core/txpool/blobpool: use types.Sender instead of signer.Sender (#30473) Use types.Sender(signer, tx) to utilize the transaction's sender cache and avoid repeated address recover. * build: use buildx to build multi-platform docker images (#30530) * eth/catalyst: use setcanonical instead of sethead in simulated fork (#30465) Fixes https://github.com/ethereum/go-ethereum/issues/30448 * cmd/geth: remove deprecated lightchaindata db (#30527) This PR removes the dependencies on `lightchaindata` db as the light protocol has been deprecated and removed from the codebase. * fix: fix lint errors * internal/ethapi: remove td field from block (#30386) implement https://github.com/ethereum/execution-apis/pull/570 * params: go-ethereum v1.14.11 stable * feat(repo): `geth/v1.14.11` upstream merge * feat(repo): `geth/v1.14.11` upstream merge --------- Signed-off-by: Icarus Wu <[email protected]> Signed-off-by: jsvisa <[email protected]> Signed-off-by: Roberto Bayardo <[email protected]> Signed-off-by: Guillaume Ballet <[email protected]> Co-authored-by: Gary Rong <[email protected]> Co-authored-by: Gealber Morales <[email protected]> Co-authored-by: ucwong <[email protected]> Co-authored-by: kukuru909 <[email protected]> Co-authored-by: Ha DANG <[email protected]> Co-authored-by: jwasinger <[email protected]> Co-authored-by: TinyFoxy <[email protected]> Co-authored-by: Péter Szilágyi <[email protected]> Co-authored-by: maskpp <[email protected]> Co-authored-by: bugmaker9371 <[email protected]> Co-authored-by: Guillaume Ballet <[email protected]> Co-authored-by: Felix Lange <[email protected]> Co-authored-by: jackyin <[email protected]> Co-authored-by: Felföldi Zsolt <[email protected]> Co-authored-by: Darioush Jalali <[email protected]> Co-authored-by: Zoro <[email protected]> Co-authored-by: Dean Eigenmann <[email protected]> Co-authored-by: Martin Holst Swende <[email protected]> Co-authored-by: Marius van der Wijden <[email protected]> Co-authored-by: prpeh <[email protected]> Co-authored-by: Halimao <[email protected]> Co-authored-by: psogv0308 <[email protected]> Co-authored-by: David Theodore <[email protected]> Co-authored-by: lightclient <[email protected]> Co-authored-by: AMIR <[email protected]> Co-authored-by: lilasxie <[email protected]> Co-authored-by: gitglorythegreat <[email protected]> Co-authored-by: Ceyhun Onur <[email protected]> Co-authored-by: Hteev Oli <[email protected]> Co-authored-by: winniehere <[email protected]> Co-authored-by: Marius Kjærstad <[email protected]> Co-authored-by: zhiqiangxu <[email protected]> Co-authored-by: Aayush Rajasekaran <[email protected]> Co-authored-by: minh-bq <[email protected]> Co-authored-by: Nathan Jo <[email protected]> Co-authored-by: Jeremy Schlatter <[email protected]> Co-authored-by: Danyal Prout <[email protected]> Co-authored-by: JeukHwang <[email protected]> Co-authored-by: Jordan Krage <[email protected]> Co-authored-by: Alexander Mint <[email protected]> Co-authored-by: Sina M <[email protected]> Co-authored-by: yukionfire <[email protected]> Co-authored-by: caseylove <[email protected]> Co-authored-by: dknopik <[email protected]> Co-authored-by: Marius G <[email protected]> Co-authored-by: lightclient <[email protected]> Co-authored-by: Seungmin Kim <[email protected]> Co-authored-by: Icarus Wu <[email protected]> Co-authored-by: ysh0566 <[email protected]> Co-authored-by: Delweng <[email protected]> Co-authored-by: stevemilk <[email protected]> Co-authored-by: Zhihao Lin <[email protected]> Co-authored-by: lmittmann <[email protected]> Co-authored-by: lmittmann <[email protected]> Co-authored-by: llkhacquan <[email protected]> Co-authored-by: taiking <[email protected]> Co-authored-by: Artyom Aminov <[email protected]> Co-authored-by: Shude Li <[email protected]> Co-authored-by: Zoo <[email protected]> Co-authored-by: Adrian Sutton <[email protected]> Co-authored-by: Dylan Vassallo <[email protected]> Co-authored-by: Arran Schlosberg <[email protected]> Co-authored-by: chen4903 <[email protected]> Co-authored-by: John Hilliard <[email protected]> Co-authored-by: Sina Mahmoodi <[email protected]> Co-authored-by: Karl Bartel <[email protected]> Co-authored-by: Oksana <[email protected]> Co-authored-by: Guillaume Ballet <[email protected]> Co-authored-by: Ignacio Hagopian <[email protected]> Co-authored-by: Nicolas Gotchac <[email protected]> Co-authored-by: markus <[email protected]> Co-authored-by: Roberto Bayardo <[email protected]> Co-authored-by: Tanishq Jasoria <[email protected]> Co-authored-by: Guillaume Michel <[email protected]> Co-authored-by: Håvard Anda Estensen <[email protected]> Co-authored-by: piersy <[email protected]> Co-authored-by: Ikko Eltociear Ashimine <[email protected]> Co-authored-by: Szupingwang <[email protected]> Co-authored-by: Karol Chojnowski <[email protected]> Co-authored-by: Ng Wei Han <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Reproducible builds
This is a little investigation into "do we have reproducible builds in geth?".
A reproducible build means that one can replicate locally a build made on e.g. a build-server.
That is, produce an exact matching binary. This is very useful to verify the integrity
of the build-servers: any remote machine can be used to watch over the builds.
The Go compiler is, supposedly, reproducible. However, go-ethereum is not pure go
ccompiler, thus we need to ensure the same compiler is usedTesting
First, I downloaded the latest build from our downloads-page. The downloads-page
lists the checksum as
8d5e138dc3eb7b08cde48966aee0ea79(note:md5is not a securecryptographic hash, but we also provide detached signatures, which offers much
better security in verifying integrity).
I then tried to create a docker container replicating the enviromment used. Details gleaned from the downloaded file:
The
.travis.ymlalso gives us some hints:Dockerfile attempt
Using a dockerfile like this:
In order to make the docker-version bundle the git data, we set the
TRAVIS,CIenv variables. Seeinternal/build/env.gofor reasons.The two builds are not exactly alike in size:
Content-wise:
VS
The text was updated successfully, but these errors were encountered: