Update geth version for integration tests #1301
Conversation
njgheorghita
force-pushed
the
geth-integration
branch
7 times, most recently
from
bb625ba
to
72687ff
Compare
| @@ -8,7 +8,8 @@ | |||
| Web3ModuleTest, | |||
| ) | |||
|
|
|||
| GETH_SIGNED_TX = b'\xf8j\x80\x85\x040\xe24\x00\x82R\x08\x94\xdcTM\x1a\xa8\x8f\xf8\xbb\xd2\xf2\xae\xc7T\xb1\xf1\xe9\x9e\x18\x12\xfd\x01\x80\x86\xee\xca\xc4f\xe1\x16\xa0\xbb7^\x1f\xf0\x03(P\x07|\x053Q\xd3M\xf1\x83\xe9\xdcp\xdc\x02\xb4\xe7`\x85\xcd\x84\xdb\xb4\xd0\xaa\xa07\x8cl\xd7\xa6R\x01\xfaW\x0e\x0f\xc1_$\xdf`\x8dO\x18\x1dC\xbc\x87\x8fud\xd2R*W\xfd4' # noqa: E501 | |||
| GETH_17_SIGNED_TX = b'\xf8j\x80\x85\x040\xe24\x00\x82R\x08\x94\xdcTM\x1a\xa8\x8f\xf8\xbb\xd2\xf2\xae\xc7T\xb1\xf1\xe9\x9e\x18\x12\xfd\x01\x80\x86\xee\xca\xc4f\xe1\x16\xa0\xbb7^\x1f\xf0\x03(P\x07|\x053Q\xd3M\xf1\x83\xe9\xdcp\xdc\x02\xb4\xe7`\x85\xcd\x84\xdb\xb4\xd0\xaa\xa07\x8cl\xd7\xa6R\x01\xfaW\x0e\x0f\xc1_$\xdf`\x8dO\x18\x1dC\xbc\x87\x8fud\xd2R*W\xfd4' # noqa: E501 | |||
| GETH_18_SIGNED_TX = b'\xf8i\x80\x84;\x9a\xca\x00\x82R\x08\x94\xdcTM\x1a\xa8\x8f\xf8\xbb\xd2\xf2\xae\xc7T\xb1\xf1\xe9\x9e\x18\x12\xfd\x01\x80\x86\xee\xca\xc4f\xe1\x16\xa0W\x91\xe6a\xb7l\x93,\xd3\xdd?;\xa8\x0e\xd3\xbe\x04\x89(\xd49\x1e+\xd6\xee\x88k\xfb\xe7\x83\xeb(\xa0\x04\x9a%`\x91uc\x1a\xdd\xb8\x96\xbb\x10\xb5v|\xcf\x03\xb6\x90\xbb\x93x\xfef\xae\xe1L\xbc7\xafJ' # noqa: E501 | |||
|
|
|||
There was a problem hiding this comment.
I'm having a hard time figuring out why these are returning different values for a signed tx. If the parameters are the same (i.e. gas price, to / from, tx...) shouldn't the signed transaction be the same? Why could different versions of geth be returning different values?
There was a problem hiding this comment.
I'd decode it (rlp.decode(...)) and see what is different.
| if 'v1.8.22' in web3.clientVersion: | ||
| # https://github.com/ethereum/go-ethereum/commit/51db5975cc5fb88db6a0dba1826b534fd4df29d7 | ||
| pytest.xfail('eth_submitHashrate endpoint updated in 1.8.22 to ethash_submitHashRate') | ||
| super().test_eth_submitHashrate(web3) |
There was a problem hiding this comment.
From what I can tell - geth no longer supports the eth_submitHashrate endpoint. What's the strategy for web3.eth.submitHashrate? Better to just leave it alone and let users running the latest geth run into "this method is not available" errors?
There was a problem hiding this comment.
Will you update this to package the go-ethereum data directory in a zip-file and update tests to manage the unzipping of that fixture. I think it's a much cleaner approach than what I did historically and makes for more manageable diffs.
One thing that @kclowes and I discussed was ensuring that only trusted parties are allowed to update those zipped fixtures since they pose attack surface for someone to get malicious code of some form into our codebase.
njgheorghita
force-pushed
the
geth-integration
branch
from
72687ff
to
ff97b12
Compare
njgheorghita
force-pushed
the
geth-integration
branch
from
1fc13f4
to
0239271
Compare
|
@pipermerriam ping for final review |
What was wrong?
update geth version for integration tests from
1.8.1to1.8.22(inspired by convo here)update go version used in integration tests
removed unused geth 1.6 fixture data
xfail test for
eth_submitHashratein1.8.22geth tests as it appears it has been dropped in favor ofethash_submitHashRateCute Animal Picture