@asmlift/core 0.8.0 → 0.8.1

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@asmlift/core",
-  "version": "0.8.0",
+  "version": "0.8.1",
   "license": "MIT",
   "type": "module",
   "description": "Match decompile an assembly function to C or Pascal",

package/src/rank.ts

@@ -2234,8 +2234,12 @@ export function rankBy<S extends { score: number; rows?: number }>(
   if (results.length === 0) {
     // Naming the withheld count matters here: "no scorable candidate" with a null cause reads as a
     // scorer failure, and a list that was entirely proof-gated is a different thing entirely.
+    // NOT `firstLine`: a compile failure puts the COMMAND on line one and the compiler's own
+    // complaint after it, so summarising to one line published the thing the reader already knows
+    // and dropped the only sentence that says what went wrong. The per-candidate `dropped` list
+    // above stays one line each — it is a roster, not a diagnosis.
     const why =
-      lastScoreErr !== null ? firstLine(lastScoreErr) : `${withheld.length} candidate(s) withheld, none scored`;
+      lastScoreErr !== null ? fullMessage(lastScoreErr) : `${withheld.length} candidate(s) withheld, none scored`;
     throw new NoScorableCandidateError(`no scorable candidate for '${symbol}': ${why}`, dropped, withheld, {
       cause: lastScoreErr,
     });
@@ -2343,3 +2347,7 @@ function castCount(source: string): number {
 function firstLine(e: unknown): string {
   return e instanceof Error ? e.message.split('\n')[0] : String(e);
 }
+
+function fullMessage(e: unknown): string {
+  return e instanceof Error ? e.message : String(e);
+}