Silverscript Tightens Syntax On The Road To v1

July and early August moved Silverscript from post-Toccata cleanup into an explicit push toward a v1 surface. The channel did not treat this as quiet polish. Ori Newman posted breaking-change notices, merged builtins, and said the point of the PR storm was simple: "Gotta get to v1 :)" ^{[1] [2]}

The densest stretch landed between August 2 and August 9. Newman merged Michael Sutton's earlier template-hash work, then shipped a run of language changes: entrypoint syntax shortened to entry, typed array literals and numeric hex literals, explicit struct types in literals and destructuring, removal of .reverse(), renaming of IfElse to Ternary, renaming of the compilation result from script to bytecode, introspection opcode renames, and stricter bitwise rules for bytes and equal-sized byte arrays. ^{[3] [4] [5] [6] [7] [8]} (PR #194) (PR #193) (PR #199) (PR #198) (PR #196) (PR #206)

The rename from script to bytecode is a useful tell. Newman said the old compilation-target name collided with the higher-level SilverScript language itself, and asked whether asm or bytecode was clearer. IzioDev preferred bytecode as the more straightforward output name. Newman agreed. ^{[9] [10] [11] [12]} (PR #196)

Builtins moved in parallel. Sutton's Blake function PR was merged on July 19. On August 7, Newman merged elldeeone's g16.verify builtin, the low-level Silverscript path for Groth16 verification through the rusty-kaspa ZK precompile. The signature lands as g16.verify(byte[] verifyingKey, byte[] proof, byte[32] ...publicInputs), with verification failure aborting script execution. ^{[13] [14]} (PR #150) (PR #138)

Signature naming also got cleaned up. Newman asked whether checkSigFromStack should become checkSigArg, arguing that "stack" is an implementation detail at the Silverscript layer. The thread settled on checkMsgSig, with the classic checkSig path left alone. ^{[15] [16] [17] [18] [19]} (PR #202)

Earlier July already pointed the same direction. Sutton posted a breaking-change notice for contracts using readInputStateWithTemplate or related template flows after PR #143 merged. He then opened PR #144 for review, noting that someone had asked for the feature in issue #140 after the fact. That PR later merged on August 2 as a new builtin. ^{[20] [21] [22] [3]} (PR #143) (PR #144)

IzioDev also tightened two small protocol-facing details on July 21: template hash as Blake3, and dispatch tags as the truncated first four Blake3 bytes of the function name. Those PRs matter because the later KCC-0001 ABI discussion cares exactly about entrypoints, dispatch tags, and whether a single-entrypoint contract should still carry a selector. ^[23] (PR #163)

The August 9 note is the current status line. Newman flagged more breaking changes in PR #206 around explicit scalar byte and integer conversions. Sutton asked whether this was the last syntax change. Newman answered carefully: he does not expect more, but he is auditing the code before releasing v1 and may still find edge cases that justify small syntax changes. ^{[8] [24] [25]} (PR #206)

The claim stays narrow. Silverscript is not claiming a finished, audited application platform. What the July-August record shows is a compiler and language surface being deliberately frozen into something releaseable: fewer CashScript leftovers, more explicit types, clearer signature builtins, first-class Groth16 verification, and an open admission that v1 audit can still force small late changes.

All sources link to public messages in the Kaspa Core R&D (public) Telegram channel.