Update felt schema to be less strict

[fracek]

Usage related changes

Update the schema used to validate felts to the most recent version
included in starknet-specs.
This schema is less strict, and allows felts that are not zero-padded.

Fix #401

Development related changes

No changes

Checklist:

• Applied formatting - ./scripts/format.sh
• No linter errors - ./scripts/lint.sh
• Performed code self-review
• Rebased to the last commit of the target branch (or merged it into my branch)
• Documented the changes
• Linked the issues which this PR resolves
• Updated the tests
• All tests are passing - ./scripts/test.sh

[fracek]

Looks like the response validation doesn't like the regex I added (copied from starknet-specs) and returns an error for every request.

{
	"error": {
		"code": -32603,
		"message": "Devnet tried to return invalid value: \"Unevaluated properties are not allowed ('block_hash', 'block_number', 'new_root', 'parent_hash', 'sequencer_address', 'status', 'timestamp', 'transactions' were unexpected)\""
	},
	"id": 0,
	"jsonrpc": "2.0"
}

[penovicp]

I can see the regex used was based on the recent change in spec, however, it looks like it might be missing some parentheses. ^0x(0|[a-fA-F1-9]{1}[a-fA-F0-9]{0,62}) checks for 0x followed by 0 or the rest, the intent was probably for it to be ^0x(0|[a-fA-F1-9]{1})([a-fA-F0-9]{0,62})$ so it checks for 0x followed by 0|[a-fA-F1-9]{1} followed by [a-fA-F0-9]{0,62}.

I'm unsure if the regex should maybe be ^0x0?[a-fA-F0-9]{1,63}$ so the leading zero is only optional and doesn't affect the count of significant digits.

[ClementWalter]

@penovicp see my correponding PR #413 with the ? which seems to me the most easy and straighforward change

[FabijanC]

@ClementWalter @penovicp
Doesn't ^0x0?[a-fA-F0-9]{1,63}$ then allow 64 digits after 0x? Should that be considered legal?

EDIT: Sorry, that's how it's supposed to be, I had a wrong number in my head

[FabijanC]

Personally I'm leaning towards penovicp's idea implement in Clement's #413

[fracek]

Agree! Let's keep it simple!

[FabijanC]

As pointed out here, we should reopen this PR and close #413.

@fracek Would you know how to resolve the issue arising with this PR?

[fracek]

The error originates when validating the response of calling starknet_getBlockWithTxs on the genesis block. I haven't investigated too much why it happens.

[FabijanC]

@mikiw Can you take a look

[mikiw]

The error originates when validating the response of calling starknet_getBlockWithTxs on the genesis block. I haven't investigated too much why it happens.

I'll take a look now.

[mikiw]

============================================ short test summary info =============================================
FAILED test/rpc/test_rpc_blocks.py::test_get_block_with_tx_hashes[hash] - KeyError: 'result'
FAILED test/rpc/test_rpc_blocks.py::test_get_block_with_tx_hashes[number] - KeyError: 'result'
FAILED test/rpc/test_rpc_blocks.py::test_get_block_with_tx_hashes[tag] - KeyError: 'result'
FAILED test/rpc/test_rpc_blocks.py::test_get_block_with_tx_hashes[tag_pending] - KeyError: 'result'
FAILED test/rpc/test_rpc_blocks.py::test_get_block_with_tx_hashes_raises_on_incorrect_block_id[block_id1] - assert {'code': -326...]{0,62})$\'"'} == {'code': 24, ...ck not found'}
FAILED test/rpc/test_rpc_blocks.py::test_get_block_with_txs[hash] - KeyError: 'result'
FAILED test/rpc/test_rpc_blocks.py::test_get_block_with_txs[number] - KeyError: 'result'
FAILED test/rpc/test_rpc_blocks.py::test_get_block_with_txs[tag] - KeyError: 'result'
FAILED test/rpc/test_rpc_blocks.py::test_get_block_with_txs[tag_pending] - KeyError: 'result'
FAILED test/rpc/test_rpc_blocks.py::test_get_block_with_txs_raises_on_incorrect_block_id[block_id1] - assert {'code': -326...]{0,62})$\'"'} == {'code': 24, ...ck not found'}
FAILED test/rpc/test_rpc_blocks.py::test_get_block_transaction_count[hash] - KeyError: 'result'
FAILED test/rpc/test_rpc_blocks.py::test_get_block_transaction_count_raises_on_incorrect_block_id[block_id1] - assert {'code': -326...]{0,62})$\'"'} == {'code': 24, ...ck not found'}
==================================== 12 failed, 7 passed in 136.81s (0:02:16) ====================================

I got these test results. Checking how to fix it.

[mikiw]

In tests we use dummy data like 0x1 I advise disabling validation in those tests to avoid many changes from 0x1 to 0x000000000000000000000000000000000000000000000000000000000000001. So validation tests will only be in test_rpc_schema.py.

[mikiw]

I made a 2fecdd9 commit with some test fixes and now just 3 tests are failing in this test file.

[mikiw]

Actually, I reverted that, seems that fix can be made only by modifying the rpc_felt wrapper.

[mikiw]

Actually, I tried that but only one test was fixed. This was my change:

    if value == 0:
        return f"{0:#0{65}x}"
    return f"{value:#0{65}x}"

Any ideas?

[FabijanC]

Superseded by #467