Track guessed as P from the test source. The ranker reads frontmatter, so an unset track parks a stub in Track T's queue regardless of what the body says -- correct the track: line if this is wrong.

origin/master has advanced 1 commit(s) since this sha. Re-verify at current HEAD before acting — the callback is tagged to the sha that was tested, which may no longer be the state of the tree.

regression: test-core#src:test/test_nested_dynarray_field.pas red at b936d125601e (auto-filed by twatch)

Type: regression (auto-filed by Track T watcher, host plexus). Untriaged.
Found: 2026-08-25T06:10:55Z
Test source: test/test_nested_dynarray_field.pas

Repro

tools/testmgr.py --tier native --job 'test-core#src:test/test_nested_dynarray_field.pas' at b936d125601ea26a9e570d65be152ff3a35d04a0

Range

bad b936d125601e, last good 0626344011cf, 6 commit(s) in range — the watcher narrows this by idle bisect; check tstate/TSTATE.md for the current range.

Log tail

pascal26:25: error: this value cannot be indexed — only arrays, strings and pointers can (g)
pascal26:30: error: this value cannot be indexed — only arrays, strings and pointers can (g)
pascal26:32: error: this value cannot be indexed — only arrays, strings and pointers can (g)
(tail)
pascal26:25: error: this value cannot be indexed — only arrays, strings and pointers can (g)
  near:  i   j  >>>  i  
pascal26:30: error: this value cannot be indexed — only arrays, strings and pointers can (g)
  near:  i   j  >>>  WriteLn  
pascal26:32: error: this value cannot be indexed — only arrays, strings and pointers can (g)
  near:       >>>   m12=  
pascal26:32: error: this value cannot be indexed — only arrays, strings and pointers can (g)
  near:       >>>   m22=  
pascal26:32: error: this value cannot be indexed — only arrays, strings and pointers can (g)
  near:       >>>   sum=

Stub ticket: signal only. Track T agent (face 2) enriches or a dev track takes it from the repro line.

2026-08-25 — resolved, commit ff6b904ce.

Resolution (2026-08-25)

Self-inflicted, by 6cb98e361 (bug-p-ten-constructs-fpc-rejects-are-accepted-and-silently-wrong). Two new diagnostics were guarded on the node's TYPE KIND, and a node's kind is the wrong question for an array: an array-shaped node carries its ELEMENT kind, and TypeIsOrdinal answers True for tyPointer (symtab.inc lists 17 deliberately — a pointer has a simple ordinal value). So a dynamic array read as "an ordinal that is not an array" from both guards.

the index guard now also requires tk <> tyPointer and ASTKind[node] = AN_IDENT, i.e. the FIRST subscript applied directly to the variable. pa^[2] and g.m[i][j] arrive with idx still naming the ROOT symbol while tk has walked down to the element type — the message did not even name the thing being indexed.
the Length guard now asks the SYMBOL of a bare identifier operand instead of the node's kind. Length(MakeDyn(3)), Length(gr.mm[0]) and Length(pfx^) are all chains and are all legal.

test/test_indexing_length_for_new_inc_positive.pas gained a row per shape, so these cannot come back. Writing those rows found two further pre-existing defects, both filed rather than papered over: bug-p-length-of-a-dereferenced-pointer-to-array-answers-zero and bug-p-an-array-returning-call-cannot-be-indexed-directly.

The lesson is root-cause-over-microfix's: "everything indexable has already been claimed by an arm above" was reasoning, and it was wrong. The measurement that would have caught it is the one that closed this — running the positive test against fpc.