" Ticked " piece of state, either LedgerState or ChainDepState

Ticking refers to the passage of time (the ticking of the clock). When a piece of state is marked as ticked, it means that time-related changes have been applied to the state. There are exactly two methods in the interface that do that: tickChainDepState and applyChainTickLedgerResult.

Also note that a successful forecast forecastFor (ledgerViewForecastAt cfg st) slot must equal protocolLedgerView cfg (applyChainTick cfg slot st). Thus a LedgerView can only be projected from a Ticked state, but cannot itself be ticked.

Some examples of time related changes:

• Scheduled delegations might have been applied in Byron
• New leader schedule computed for Shelley
• Transition from Byron to Shelley activated in the hard fork combinator.
• Nonces switched out at the start of a new epoch.

Forecast annotated with details about the ledger it was derived from

Change a telescope of a forecast into a forecast of a telescope

Canonical TxIn

The Ledger and Consensus team discussed the fact that we need to be able to reach the TxIn key for an entry from any era, regardless of the era in which it was created, therefore we need to have a "canonical" serialization that doesn't change between eras. For now we are requiring that a HardForkBlock has only one associated TxIn type as a stop-gap, but Ledger will provide a serialization function into something more efficient.

This choice for HardForkTxOut imposes some complications on the code.

We deliberately chose not to have all values in the tables be TxOut era because this would require us to traverse and translate the whole UTxO set on era boundaries. To avoid this, we are holding a NS WrapTxOut xs instead.

Whenever we are carrying a LedgerState (HardForkBlock xs) mk (or ExtLedgerState), the tables are the ones inside the particular ledger state in the Telescope of the HardForkState.

However, when we are carrying LedgerTables (HardForkBlock xs) mk we are instead carrying these tables, where the TxOut is an NS. This means that whenever we are extracting these tables, we are effectively duplicating the UTxO set (Map) inside, to create an identical one where every element has been translated to the most recent era and unwrapped from the NS.

To prevent memory explosion, try to only perform one of this transformations, for example:

• when applying blocks, inject the tables for the transactions only once, and extract them only once.
• when performing queries on the tables (that use QFTraverseTables), operate with the tables at the hard fork level until the very end, when you have to promote them to some specific era.

(image code)

>>> :{
>>> either (error . show) pure =<<
>>> renderToFile "docs/haddocks/hard-fork-tables.svg" defaultEnv (tikz ["positioning", "arrows"]) "\\node at (4.5,4.8) {\\small{LedgerTables (LedgerState (HardForkBlock xs))}};\
>>> \ \\draw (0,0) rectangle (9,5);\
>>> \ \\node (rect) at (1.5,4) [draw,minimum width=1cm,minimum height=0.5cm] {TxIn};\
>>> \ \\node (oneOf) at (3.5,4) [draw=none] {NS};\
>>> \ \\draw (rect) -> (oneOf);\
>>> \ \\node (sh) at (6.5,4) [draw,minimum width=1cm,minimum height=0.5cm] {BlockATxOut};\
>>> \ \\node (al) at (6.5,3) [draw,minimum width=1cm,minimum height=0.5cm] {BlockBTxOut};\
>>> \ \\node (my) at (6.5,2) [draw=none,minimum width=1cm,minimum height=0.5cm] {...};\
>>> \ \\node (ba) at (6.5,1) [draw,minimum width=1cm,minimum height=0.5cm] {BlockNTxOut};\
>>> \ \\draw (oneOf) -> (sh);\
>>> \ \\draw (oneOf) -> (al);\
>>> \ \\draw (oneOf) -> (ba);\
>>> \ \\draw (3,0.5) rectangle (8,4.5);"
>>> :}

>>> :{
>>> either (error . show) pure =<<
>>> renderToFile "docs/haddocks/hard-fork-tables-per-block.svg" defaultEnv (tikz ["positioning", "arrows"]) "\\node at (5,4.8) {\\small{LedgerState (HardForkBlock xs)}};\
>>> \ \\draw (0,0) rectangle (10,5);\
>>> \ \\node (oneOf2) at (2,4) [draw=none] {HardForkState};\
>>> \ \\node (bb) at (5,4) [draw,minimum width=1cm,minimum height=0.5cm] {BlockAState};\
>>> \ \\node (bt) at (8,4) [draw,minimum width=1cm,minimum height=0.5cm] {BlockATables};\
>>> \ \\node (sb) at (5,3) [draw,minimum width=1cm,minimum height=0.5cm] {BlockBState};\
>>> \ \\node (st) at (8,3) [draw,minimum width=1cm,minimum height=0.5cm] {BlockBTables};\
>>> \ \\node (db) at (5,2) [draw=none,minimum width=1cm,minimum height=0.5cm] {...};\
>>> \ \\node (dt) at (8,2) [draw=none,minimum width=1cm,minimum height=0.5cm] {...};\
>>> \ \\node (bab) at (5,1) [draw,minimum width=1cm,minimum height=0.5cm] {BlockNState};\
>>> \ \\node (bat) at (8,1) [draw,minimum width=1cm,minimum height=0.5cm] {BlockNTables};\
>>> \ \\draw (oneOf2) -> (bb);\
>>> \ \\draw (bb) -> (bt);\
>>> \ \\draw (oneOf2) -> (sb);\
>>> \ \\draw (sb) -> (st);\
>>> \ \\draw (oneOf2) -> (bab);\
>>> \ \\draw (bab) -> (bat);"
>>> :}