Split Seq into ReadSeq, Seq, and ResizableSeq
Seq required every implementor to provide growth methods even when they made no sense — a fixed-size mutable container can't grow, but it couldn't implement Seq without pretending it could. The Seq interface has been split into three interfaces:
ReadSeq[A](unchanged) -- read-only access by index.Seq[A](narrowed) -- read and write by index.ResizableSeq[A](new) -- read, write, grow, and shrink.
Array, String, and List now implement ResizableSeq instead of Seq.
Code that uses Seq and calls growth methods must change Seq to ResizableSeq.
Fix release build link failure when building with GCC
Building ponyc with GCC in release mode failed at two points: the ponyc binary itself failed to link, and programs compiled by a GCC-built ponyc failed to link with undefined runtime symbols. Both had the same cause — GCC's LTO format (GIMPLE) is not compatible with LLVM tools. The build system was enabling LTO globally regardless of compiler, producing GIMPLE objects that the embedded ld.lld linker and llvm-ar archiver could not read. LTO is now disabled entirely for GCC builds.
Remove compile-time error for constant integer division by zero
The compiler no longer rejects constant integer division or remainder by zero. let x: I64 = 1 / 0 was a compile error, but let x = I64(1) / I64(0) was not — the check only caught bare literals.
Fix I128 constant division by -1 producing wrong result
let r: I128 = 0 / -1 produced I128.min_value() instead of 0. The compiler's constant-folding path compared 128-bit values through 64-bit intermediaries, and I128.min_value() has all-zero lower 64 bits — making I128(0) and I128.min_value() look equal. Any I128 constant that is a multiple of 2^64 divided by -1 was affected.
Fix backpressure muting cascade
Under fan-in topologies — many actors sending to a shared slow receiver — the runtime's binary muting cascaded backward through the entire actor graph. When the receiver became overloaded, every actor that sent to it was immediately muted. Their senders then saw muted targets and were muted in turn. Within one scheduling interval, most of the system was frozen, dropping throughput from tens of millions of messages per second to near zero.
The runtime now uses graduated per-actor batch sizes instead of binary muting. Each actor's batch size adjusts dynamically: senders to overloaded receivers are throttled proportionally (one step per send) rather than muted outright, and overloaded receivers process larger batches to catch up. Actors are only fully muted after repeated sends to overloaded receivers within a single scheduling run. This eliminates the cascade — throttled actors remain scheduled, so their senders never see them as muted.
[0.76.0] - 2026-10-14
Fixed
- Fix release build link failure when building with GCC (PR #6277)
- Fix I128 constant division by -1 producing wrong result (PR #6282)
- Fix backpressure muting cascade (PR #6285)