deps: make syn and serde_json be workspac dependencies, manualy upgrade bitflags from 1.3 to 2.9

[eval-exec]

What problem does this PR solve?

Issue Number: close #4844

Related changes

• deps: make sync and serde_json be ckb workspace dependencies
• deps: upgrade bitflags from 1.3 to 2.9, fix chore(deps): bump bitflags from 1.3.2 to 2.9.0 #4844

Check List

Tests

• Unit test
• Integration test
• Manual test (add detailed scripts or steps below)
• No code ci-runs-only: [ quick_checks,linters ]

Side effects

• None

Release note

None: Exclude this PR from the release note.

[driftluo]

bitflag's serialization scheme has a breaking change(In my impression, the endianness sorting has changed) from 1.0 to 2.0, which will cause errors in identifying and reading data stored in peer stores. You cannot simply upgrade it. You need to change the default serialization scheme to 1.0 mode, otherwise, it will cause problems with ckb connection and discovery.

All other places use internal data. Only the network needs to interact and store data, and consistency must be guaranteed.

[eval-exec]

Serialize, Deserialize has a breakchange: https://github.com/bitflags/bitflags/releases/tag/2.0.0

But the AddrInfo#flags is a u64, not bitflag::Flag

Need to review all actions related to the serialization of CKB and bitflags again.
Or need to reconsider whether upgrading bitflags from 1.0 to 2.0 is truly necessary?

[driftluo]

Just verify whether the conversion between 1.0 and 2.0 is the same in our usage. I remember it is bitflag to u64, then u64 to bitflag.

[eval-exec]

bitflags's Flag value are encoded to ckb-network in bellow three message:
1:

ckb/util/gen-types/schemas/extensions.mol

Lines 460 to 464 in 17bd37a

	table Identify {
	flag: Uint64, // Flag
	name: Bytes, // Network Name
	client_version: Bytes,
	}

2:

ckb/util/gen-types/schemas/protocols.mol

Lines 43 to 48 in 17bd37a

	table GetNodes2 {
	version: Uint32,
	count: Uint32,
	listen_port: PortOpt,
	required_flags: Uint64,
	}

3:

ckb/util/gen-types/schemas/protocols.mol

Lines 64 to 67 in 17bd37a

	table Node2 {
	addresses: BytesVec,
	flags: Uint64,
	}

These flags are Uint64, which is created by flags.bits().pack().
Both bitflags@1 and bitflags@2 using bits method:

    #[doc = r" Returns the raw value of the flags currently stored."]
    #[inline]
    pub const fn bits(&self) -> u64 {
        self.bits
    }

This method return same u64 value for both version, so I beleave upgrade bitflag@1 to bitflag@2 is safe.