Sigma Σ hardfork rules

[x100111010]

• for 8 bps k = 102
• target_time_per_block reduced from 1000ms to 125ms
• Coinbase maturity: ~100 blocks at 1 bps -> 800 blocks at 8 bps
• Mergeset size limit (L): 204 (2k)
• max_block_parents: 10 -> 16
• gen_testnet11_genesis test expects bps=10 (const)
• Activation of earlier KIPs: KIP-4, KIP-9, KIP-13, KIP-10

---

bb64053

With increased blockrate syncing becomes quite slow and due to astrobwt our header verification takes more time to process, to compensate this we reduce the finality duration to 9h and also the pruning duration to around 20.32h. (represents social consensus tolerance rather than cryptographic security)

pub const NEW_FINALITY_DURATION: u64 = 32_400; // 9 hours = 259200 Blocks
pub const NEW_PRUNING_DURATION: u64 = 73_141; // 20.32 hours = 585128 Blocks

---

a7f0dbd

The transient storage mass intruced in KIP-0013 prevents excessive storage growth from payload abuse.
Since our pruning depth and finality depth are different due to different consensus parameters, we adjust the factor to 3.

Block size limit: 125KB (Kaspa with the transient mass factor of 4) -> 167KB (+33% capacity)
Worst-case storage: ~141GB (vs original 775GB threat)

---

Subsidy calculation has been rethought to preserve our monetary policy with 8x faster blocks. Instead of each block getting a fixed reward amount (which would accelerate emission by 8x), we now treat rewards as per-second values. When blocks come 8x faster, each one gets 1/8th of the original reward. So to calculate the actual block reward, we divide these values by the current BPS rate (1 before activation, 8 after).

if daa_score < activation {
    seconds = (daa_score - deflation_start);  // 1 BPS: blocks = seconds
} else {
    seconds = (activation - deflation_start) + 
              (daa_score - activation) / 8;   // 8 BPS
}

• Total rewards delta: +60 SPR (see tests::calc_high_bps_total_rewards_delta())

https://github.com/x100111010/rusty-spectre/blob/sigma_2/consensus/src/processes/coinbase.rs#L78

There's a small 60 SPR supply increase because we use ceiling division.

---

88638cf which was consolidated into sigma_activation 33ede2a

Adds a hardfork logic matrix_activation which improves randomness in heavy_hash (switch to XOR-based bit mixing in heavyhash for header_version == 2 instead of original bit shifting). spectre-miner was already updated to support this algorithm improvement and will automatically detect header version 2, this way we also avoid hardcoding any fixed daa_score or timestamp