eyeprolog 1.2.17 → 1.2.18

package/conformance-report.md

@@ -11,7 +11,7 @@ This report summarizes the file-based conformance corpus under `test/conformance
 | builtins | 11 | 0 | 0 | 0 | 11 |
 | context | 11 | 0 | 0 | 0 | 11 |
 | control | 15 | 0 | 0 | 0 | 15 |
-| iso | 167 | 218 | 0 | 0 | 385 |
+| iso | 168 | 218 | 0 | 0 | 386 |
 | lists | 52 | 3 | 0 | 0 | 55 |
 | modules | 2 | 0 | 0 | 0 | 2 |
 | negation | 8 | 0 | 19 | 0 | 27 |
@@ -23,4 +23,4 @@ This report summarizes the file-based conformance corpus under `test/conformance
 | terms | 26 | 3 | 0 | 0 | 29 |
 | unification | 18 | 0 | 0 | 0 | 18 |
 | variables | 16 | 9 | 0 | 0 | 25 |
-| **Total** | **482** | **269** | **19** | **21** | **791** |
+| **Total** | **483** | **269** | **19** | **21** | **792** |

package/package.json

@@ -3,7 +3,7 @@
   "publishConfig": {
     "access": "public"
   },
-  "version": "1.2.17",
+  "version": "1.2.18",
   "description": "EyeProlog turns facts and rules into answers and proofs.",
   "type": "module",
   "main": "./index.js",

package/src/iso.js

@@ -1556,7 +1556,15 @@ function parseIsoNumber(text) {
   let sign = '';
 
   if (text[position] === '-') {
-    if (/[\u0009-\u000d\u0020]/.test(text[position + 1] ?? '')) {
+    const next = text[position + 1] ?? '';
+    // Every token class may carry leading layout (6.4).  Thus a negative
+    // number may have layout between the name `-` and its numeric token.  A
+    // `%...\n` comment can start immediately after `-` because `%` cannot
+    // continue a graphic name token.  In contrast `/*...*/` cannot start
+    // there without separating layout: `/` *can* continue the graphic token,
+    // and the eager-consumer rule therefore keeps `-/**/1` ill-formed (the
+    // number_chars continuation corpus case 24).
+    if (/[\u0009-\u000d\u0020]/.test(next) || next === '%') {
       position = skipNumberLayout(text, position + 1);
       sign = '-';
     }

package/test/conformance/README.md

@@ -102,7 +102,7 @@ Selected cases are adapted from the ISO and standard-core suites of Logtalk,
 Scryer Prolog, Trealla Prolog, and SWI-Prolog. Their upstream identifiers and licenses
 are recorded in [THIRD_PARTY.md](THIRD_PARTY.md).
 
-The corpus has 385 cases in `iso/` and 791 file-based conformance cases in
+The corpus has 386 cases in `iso/` and 792 file-based conformance cases in
 total. The generated `conformance-report.md` is the authoritative source for
 current category totals. Together with regression, documentation-sync, API,
 example, and book-example checks, `npm test` is the release gate.

package/test/conformance/cases/iso/number_chars_minus_line_comment.pl

@@ -0,0 +1,8 @@
+% ISO 13211-1 8.16.7/8.16.8 with 6.3.1 and token layout from 6.4.
+%% goal: answer(X0, X1, X2, X3)
+
+answer(CharZero, CharOne, CodeZero, CodeOne) :-
+    number_chars(CharZero, "-%\n0"),
+    number_chars(CharOne, "-% comment\n1"),
+    number_codes(CodeZero, [45, 37, 10, 48]),
+    number_codes(CodeOne, [45, 37, 32, 99, 111, 109, 109, 101, 110, 116, 10, 49]).

package/test/conformance/expected/iso/number_chars_minus_line_comment.pl

@@ -0,0 +1 @@
+answer(0, -1, 0, -1).

package/test/fixtures/number_chars_cont_quad.pl

@@ -39,6 +39,9 @@
 73 ?- number_chars(N,"(0)").
       syntax_error(...).
 
+74 ?- number_chars(N,"-%\n0").
+      N = 0.
+
 4  ?- number_chars(1,"a").
       syntax_error(...).
 

package/test/run-regression.mjs

@@ -491,13 +491,51 @@ c4 ?- call((!;1)).
       run: () => {
         const filename = path.join(testRoot, 'fixtures', 'number_chars_cont_quad.pl');
         const source = fs.readFileSync(filename, 'utf8');
+        const numbered = new Set([...source.matchAll(/^(\d+)\s+\?-/gm)].map((match) => Number(match[1])));
+        assertEqual(numbered.size, 74, 'numbered case total');
+        for (let id = 1; id <= 74; id++) {
+          if (!numbered.has(id)) throw new Error(`number_chars continuation case #${id} is missing`);
+        }
         const result = publicApi.runQuads(Program.parseSources([{
           text: source,
           filename,
         }]));
-        assertEqual(result.total, 77, 'answer-description total');
-        assertEqual(result.passed, 77, 'answer-description passed');
-        assertEqual(result.stdout, 'quads: 77 run, 77 passed, 0 failed.\n', 'quad report');
+        assertEqual(result.total, 78, 'answer-description total');
+        assertEqual(result.passed, 78, 'answer-description passed');
+        assertEqual(result.stdout, 'quads: 78 run, 78 passed, 0 failed.\n', 'quad report');
+      },
+    },
+    {
+      name: 'number conversion accepts line-comment layout after a minus token',
+      run: () => {
+        const source = String.raw`
+?- number_chars(N,"-%\n0").
+   N = 0.
+?- number_chars(N,"-% comment\n1").
+   N = -1.
+?- number_codes(N,[45,37,10,48]).
+   N = 0.
+?- number_codes(N,[45,37,32,99,111,109,109,101,110,116,10,49]).
+   N = -1.
+`;
+        const result = publicApi.runQuads(source);
+        assertEqual(result.total, 4, 'quad total');
+        assertEqual(result.passed, 4, 'quad passed');
+        assertEqual(result.stdout, 'quads: 4 run, 4 passed, 0 failed.\n', 'quad report');
+
+        // Keep the eager-consumer distinction from continuation case #24:
+        // an adjacent bracketed comment can be consumed as part of a graphic
+        // token after `-`, so it is not equivalent to the `%` line comment.
+        for (const goal of ['number_chars(N,"-/**/1")', 'number_codes(N,[45,47,42,42,47,49])']) {
+          let caught = null;
+          try {
+            publicApi.run('', { goal });
+          } catch (error) {
+            caught = error;
+          }
+          if (caught == null) throw new Error(`${goal} should throw`);
+          assertIncludes(String(caught?.message ?? caught), 'syntax_error(number)', goal);
+        }
       },
     },
     {

package/the-art-of-eyeprolog.md

@@ -5758,12 +5758,16 @@ quoted_atom("ab").           % quoted_atom(ab)
 
 Conversions accept partial output lists when the atomic input is known, but
 constructing an atom or number requires a complete proper list with no unbound
-elements. Numeric parsing accepts leading ISO layout characters, an optional
-sign, decimal fractions, and decimal exponents; it rejects trailing material
-and non-finite values. The regression gate vendors all 73 numbered cases from
-Ulrich Neumerkel's contemporary `number_chars/2` comparison, including the
-Cor.2 error-precedence cases; `number_codes/2` shares the same numeric parser
-and has mirrored coverage for the parenthesized-number regression.
+elements. Numeric parsing accepts ISO layout before tokens, including layout between a
+minus token and the following numeric token. A single-line `%...` comment may
+therefore follow `-` directly because `%` cannot continue a graphic token; an
+adjacent bracketed comment in `-/**/1` remains a syntax error under the eager
+token-consumer rule. Decimal fractions and decimal exponents are supported;
+trailing material and non-finite values are rejected. The regression gate
+vendors all 74 numbered cases from Ulrich Neumerkel's contemporary
+`number_chars/2` comparison, including the Cor.2 error-precedence cases;
+`number_codes/2` shares the same numeric parser and has mirrored coverage for
+the recent numeric-syntax regressions.
 
 ### Streams and unit I/O
 
@@ -7020,7 +7024,7 @@ precedence still need one-by-one closure. `test/conformance/ISO-MATRIX.md`
 maps language families to representative executable cases.
 
 The complete suite must pass before release. The file-based conformance corpus
-contains 791 cases, including 385 focused ISO
+contains 792 cases, including 386 focused ISO
 cases derived from the success, failure, mode, and error behavior in
 ISO/IEC 13211-1 clauses 7 and 8, Part 2 modules, and Part 3 grammar rules.
 Separate exact-output suites check 189 normal