Skip to content

Add exact text semantic reference and oracle - #34

Draft
isomorphisms wants to merge 4 commits into
mainfrom
exact-text-semantic-reference
Draft

Add exact text semantic reference and oracle#34
isomorphisms wants to merge 4 commits into
mainfrom
exact-text-semantic-reference

Conversation

@isomorphisms

Copy link
Copy Markdown
Owner

Summary

  • add a deliberately simple Idriç exact-substring semantic reference
  • expose contains_string, overlapping zero-based character match_positions, and match_count
  • add deterministic fixtures for absent/beginning/middle/end, overlap, one-byte, repeated-prefix, and empty-pattern behavior
  • add a focused CI job compiling/running the oracle with IB's pinned Idriç compiler
  • document the distinction from PR Add filesystem Float32 vector backend #17's Float32 vector exact scan and the intended ARM/Thumb optimization boundary

Why this slice first

PR #9 described a future grep-like extraction pass but did not implement one. This makes that ordinary-text operation real without prematurely baking an ARM trick into browser semantics.

The current ARM/Thumb backend is still straight-line and numerical; PR #8 is a branch/dispatch acceptance-spec change, not string/loop lowering. The browser oracle therefore lands as a separate slice. Target-specific work remains tracked in isomorphisms/Idric#21 and isomorphisms/idric-arm-thumb#9, #10, and #11.

Semantics

  • case-sensitive Char substring matching
  • overlaps count
  • positions are zero-based character positions
  • empty pattern matches every character boundary, including the end

The first ARM comparison should use ASCII fixtures so byte and character offsets coincide. Any later Unicode-optimized API must state its offset units explicitly.

Refs #33.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant