Dev Updates
Development in the open. PyneCore and PyneComp releases land here as they happen — what got fixed, what got added, and now and then what we got wrong — along with progress on work that is not finished yet. Synced directly from our Discord dev-updates channel.
PyneCore’s full changelog lives on GitHub; PyneComp’s progress you will find here and nowhere else. More granular and more technical than News — this is the engineering log, not the announcements.
PyneCore v6.9.0 released — Pine int is now a static type over a double-only runtime,…
Every int-typed value travels as an integral float; na(int) is a plain nan, truncation happens at the consuming slot, storage keeps the value (array.new_int(1, 7 / 4) holds 1.75) - Static Pine type inference stamps the…
Read more →PyneComp v6.0.65 — a script that uses a numeric ternary as a condition, like if a and…
Converting or compiling a Pine script where a numeric ? : expression with an na branch is used inside an if condition failed with an internal error (HTTP 400 from the API). It now converts exactly the way TradingView…
Read more →PyneComp v6.0.64 — seven fixes for scripts that used to convert or compile into Python…
A Pine v5 script dividing two integer constants (10 / 3) kept Pine 5's truncating integer result instead of silently turning into a decimal one after conversion — this changed values wherever a script compared against…
Read more →PyneCore v6.8.14 released — twelve TradingView-parity fixes across the strategy engine,…
Same-instant reversals are booked long side first when both entry orders waited in the book together; a calc_on_order_fills-issued leg keeps its chronological order - pyramiding is judged when strategy.entry runs,…
Read more →PyneComp v6.0.63 — four conversion bugs that silently changed what a strategy does are…
Converting a v4 strategy whose strategy.entry() or strategy.order() took a computed long argument produced a script that traded the wrong way — a DCA bot closing with long=(IS_LONG?false:true) opened a doubling long…
Read more →PyneCore v6.8.13 released — the remaining measured TradingView divergences in the…
TradingView's second intrabar margin-call stage is reproduced: 848/848 margin-call bars exact on BINANCE:BTCUSDT 30m, and Rocket Grid Algorithm goes bit-exact at 411/411 trades - request.security() with…
Read more →PyneCore v6.8.12 released — the strategy statistics surface is complete and the equity…
The last missing statistics variables are in: strategy.avg_trade_percent, avg_winning_trade_percent, avg_losing_trade_percent, grossprofit_percent, grossloss_percent, max_contracts_held_all/long/short,…
Read more →PyneComp v6.0.61 — three conversion and compilation bugs that silently changed what a…
A variable declared as var x = input.something(...) now keeps the value you assign to it with :=. Before, it snapped back to the input's default value on the next bar, so any counter or state seeded from an input was…
Read more →PyneCore v6.8.11 released — loop call sites now share one instance the way TradingView…
Loop iterations share ONE instance per call site: user var state accumulates across a bar's iterations, callee builtin machines roll back to bar start — except ta.percentile_*, ta.highest/ta.lowest and math.random,…
Read more →PyneCore v6.8.10 released — a strategy-engine release: how TradingView sizes, books and…
strategy.risk.max_drawdown now enforces against the realized (closed-equity) high-water mark and only on bars that book realized P&L — floating drawdown alone no longer halts trading - Exit orders settle against the…
Read more →