batcher: remove isPectra and assume Pectra is always active

[Himess]

Summary

Since the Pectra fork happened long ago, we can now assume L1 is always running Pectra. This PR simplifies the batcher code by removing the isPectra checks.

Changes

• Remove isPectra parameter from ChannelConfigProvider.ChannelConfig()
• Remove isPectra parameter from channelManager.TxData()
• Simplify BatchSubmitter.l1Tip() to no longer return isPectra boolean
• Update computeSingleCalldataTxCost() to always use totalCostFloorPerToken = 10 (Pectra pricing)
• Update all related tests to reflect Pectra-only behavior

Files Changed

• op-batcher/batcher/channel_config_provider.go
• op-batcher/batcher/channel_config.go
• op-batcher/batcher/channel_manager.go
• op-batcher/batcher/driver.go
• op-batcher/batcher/test_batch_submitter.go
• op-batcher/batcher/channel_config_provider_test.go
• op-batcher/batcher/channel_manager_test.go
• op-batcher/batcher/channel_manager_memory_test.go

Closes #18479

[Himess]

@geoknee Checked op-e2e/system/da - the L1PragueTimeOffset in eip4844_test.go is just test config to activate Prague at genesis, not a runtime check like isPectra was. No similar simplification needed there.

There's a comment about updating maxBlobsPerBlock to Prague's limit (9 vs 6), but that's more of a feature change than a simplification.

[geoknee]

/ci authorize 62490c8