Ignore spent funding cell(s) in channel gossip message

[bolajahmad]

Closes #530

Description

Currently, during ChannelAnnouncement, the funding TX is fetched [using get_transaction](https://github.com/nervosnetwork/fiber/blob/develop/crates/fiber-lib/src/ckb/client.rs#L220C4-L227C6).

This approach only checks that the Funding Outpoint was Committed. But a Committed Outpoint does not guarantee availability of the Cell.

This PR adds that layer of security.

• Ensure the Outpoint is Committed
• Ensure that the underlying Cell is still Live
• No point in broadcasting a dead cell.

Tests

cargo test -p fnn --all-features \
ckb::tests::test_utils::test_mock_ckb_chain_client_reports_dead_live_cell -- --exact \
 && cargo test -p fnn --all-features  \
fiber::tests::gossip::test_reject_channel_announcement_when_funding_outpoint_is_dead \ 
-- --exact

cargo fmt --all -- --check
cargo clippy -p fnn --features rocksdb --all-targets -- -D warnings
git diff --cached --check

[chainTe]

Security Review

Result: No reportable security findings in the intended change.
Reviewed range: 3c25bcf16200e5d641dcd9b79f086f391e976172..8562d79ca726099fc2cb0295f10e197241a37196

Validation:

• Existing broad tests were not run locally; CI is expected to cover them.
• Targeted build validation: cargo check -p fnn --features sqlite fails because crates/fiber-lib/src/fiber/gossip.rs calls is_funding_outpoint_live(...), but that function is not defined in this PR.
• cargo check -p fnn --features rocksdb could not reach the PR code in this local environment because RocksDB binding generation failed on a missing local Clang header (stdbool.h).

Notes:

• The added live-cell gate is security-positive in intent, but the helper implementation needs to be added (or the call updated) before this can be buildable and fully validated.

[quake]

@bolajahmad we added block_hash to the rpc get_live_cell in ckb 0.208 ( nervosnetwork/ckb#5269 )

I think we can use this new field to improve the existing live cell checking code.

[bolajahmad]

@bolajahmad we added block_hash to the rpc get_live_cell in ckb 0.208 ( nervosnetwork/ckb#5269 )

I think we can use this new field to improve the existing live cell checking code.

If I understand correctly and based on the fixes I made, the new block_hash field in the response has the same effect as checking if the cell is Some. If anything,I wonder if it's also okay to push this change and follow up with another commit to upgrade to new v5.1.0. If I'm missing something, please let me know!

[quake]

@bolajahmad we added block_hash to the rpc get_live_cell in ckb 0.208 ( nervosnetwork/ckb#5269 )
I think we can use this new field to improve the existing live cell checking code.

If I understand correctly and based on the fixes I made, the new block_hash field in the response has the same effect as checking if the cell is Some. If anything,I wonder if it's also okay to push this change and follow up with another commit to upgrade to new v5.1.0. If I'm missing something, please let me know!

lgtm 👍