[nasa314159]

Summary

Adds a short note to the pattern-matching chapter documenting a limitation for products of vector wildcards. This is a documentation-only change; it does not modify the matcher.

Motivation / issue #300

Issue #300 reports that a pattern with repeated vector wildcards does not match a term that would match under another assignment of the same vector factors. A maintainer comment (2020-07-28) says this is a known shortcoming, that a full fix would require a large rewrite, and that it would be best to mention it in the manual instead.

Documents #300.

Documentation changes

• Add a concise statement of the limitation.
• Add the minimal reproducer from issue Pattern matching for products of vectors #300.
• Explain the expected versus actual matching.
• Note that FORM stores factors in a canonical order, so the input order is not necessarily the order used by the matcher.
• Note that the matcher has backtracking support for function factors that is not available in the same way for vector factors.
• Warn that replacing vectors with commuting functions or tensors changes algebraic semantics and is not a general workaround.
• Note that separate id statements can cover fixed small cases, but do not make the matcher try all permutations.

Validation performed

• Reproducer on FORM 5.0.2 (macOS arm64) still leaves the first term unchanged and removes the second; exit 0.
• Tested the fixed-case id workaround and input reordering; reordering the input does not help.
• Tested replacement by commuting functions/tensors; this changes the objects, so it is not presented as semantically equivalent.
• git diff --check clean.
• pdflatex (two passes) and htlatex builds completed locally; dvipng is not installed, so HTML math images were not generated. No system dependencies were installed.
• The patch changes only doc/manual/pattern.tex.

Limitations

• This documents the current behavior; it does not fix the matcher.
• The note does not specify a complete matching order or promise permutation-complete matching.
• Local documentation builds are not a substitute for upstream CI.
• Wording may need maintainer adjustment (for example, manual versus developer reference).