Commit Graph

13 Commits

Author SHA1 Message Date
Kinneyzhang
95ed2b32d1 feat: buffer-level justification — region commands and auto-justify mode
The renderer (shared by both engines) is now lossless: synthesized
glue carries the original text it replaced (ekp-glue), soft line
breaks carry the whitespace swallowed around the break
(ekp-soft-break), break hyphens are marked (ekp-soft-hyphen), and
paragraph-edge whitespace survives as zero-display ekp-hidden text.
Orphaned ekp--combine-glues-and-boxes / ekp--interleave removed.

New ekp-region.el:
- ekp-justify-region / ekp-unjustify-region: in-place justification
  with exact structural restore (character- and property-exact),
  robust to edits made while justified
- ekp-auto-justify-mode: keeps the buffer justified at the window
  width; debounced re-flow on window resize, incremental
  per-paragraph re-justification after edits (served by the
  paragraph cache)
- fix found in live GUI testing: buffer-local members of
  window-size-change-functions receive the WINDOW as argument and
  may run with an unrelated buffer current; the handler now resolves
  window and buffer explicitly (regression test included)

Tests: 47 ERT (36 core + 11 region) passing; 300-case fuzz 0
failures (C/elisp parity unchanged); byte-compile clean with
error-on-warn. Verified interactively in GUI Emacs: justified at
1403px maximized, auto-reflowed to 614px on frame resize.

docs: interactive-use section in readme.md / readme_zh.md
ci: byte-compile list includes ekp-region.el

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-26 20:23:03 +08:00
Kinneyzhang
112b3a0e52 refactor!: overhaul KP core — correctness, C parity, performance, tests, docs
- fix: ekp-param-set silently reset after first justify (now persists;
  ekp-param-reset added)
- fix: narrow-width CJK returned empty string (data loss); two-pass
  emergency-break strategy guarantees output for any input
- fix: K-P penalties never synced to C module; space-box metrics
  divergence between C and Elisp engines
- fix: para cache ignored ekp-latin-lang (stale hyphenation after
  language switch) and used collision-prone sxhash keys
- fix: fullwidth letters/digits misclassified as CJK punctuation
- fix: combining chars split from their base char in the tokenizer
- fix: punctuation-wrapped words (word!/(word)/word;) never hyphenated
- fix: renderer double-counted stripped space widths; negative glue
  clamped; batch/tty font detection no longer crashes
- feat: real looseness support via (position × line-count) DP
- perf: O(1) line metrics and gap counts via prefix arrays (inner loop
  previously allocated O(n) subsequences → O(n³) total); box measurement
  dedupe; eq fast-path para lookup; prebuilt per-para glue arrays
  → zh justify 7547ms → 96ms (compiled elisp) / 57ms (C);
    range-justify 68.5s → 0.48s / 34ms; C module itself 3–19× faster
- test: 36 batch-safe ERT tests + 300-case property fuzz (C/elisp
  byte-identical output, zero content loss) replacing ad-hoc suite
- docs: readme/readme_zh/DEVELOPER/DEVELOPER_ZH/ekp_c-README rewritten
  to match the implementation; phase handoff in .phrase/phases/

BREAKING: requires Emacs 29.1+; C module must be rebuilt (v1.1, new
arities); ekp-threshold-factor / ekp-flagged-penalty /
ekp-forced-break-penalty removed; Rust module stubs removed.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-26 18:44:02 +08:00
Kinneyzhang
06e28b078d add test 2026-01-25 20:12:38 +08:00
copilot-swe-agent[bot]
32021fbe35 Implement Looseness, Threshold Pruning, and Flagged Penalties
Feature 1: Looseness (Complete Implementation)
- Added alt-paths hash table to track alternative paths by (position, line-count)
- Modified DP loop to track all paths reaching each position
- ekp--dp-trace-breaks-with-looseness uses alt-paths to find closest match
- Added ekp--dp-trace-alt-path helper with safety limit

Feature 2: Threshold Pruning
- Added ekp-threshold-factor variable (default 0 = disabled)
- Paths with demerits > best × (1 + factor) are skipped
- Tracks best-end-demerits during DP for pruning decisions

Feature 3: Flagged Penalties (Forced Breaks)
- Added flagged-positions field to ekp-para struct
- Added ekp-flagged-penalty variable (negative = preferred)
- Added ekp--flagged-p with O(log n) binary search
- Modified DP to always accept flagged breaks

Tests:
- ekp-test-unit--hyphenate-p-binary-search
- ekp-test-unit--flagged-p-binary-search
- ekp-test-unit--alt-paths-hash
- ekp-test-unit--threshold-factor
- ekp-test-unit--flagged-penalty
- Updated ekp-test-unit--struct-access for new fields

Co-authored-by: Kinneyzhang <38454496+Kinneyzhang@users.noreply.github.com>
2026-01-25 11:35:30 +00:00
Kinneyzhang
6421f6b1fb refactor ekp cache and code improve 2026-01-24 22:40:53 +08:00
Kinneyzhang
68b46f8337 code improve 2026-01-24 20:44:38 +08:00
Kinneyzhang
ca9349e99f fix bug when text is blank string 2025-09-26 00:18:37 +08:00
Kinneyzhang
820e73920a update test 2025-08-22 00:15:33 +08:00
Kinneyzhang
37baccded8 keep original text properties of string after typesetting 2025-08-22 00:06:14 +08:00
Kinneyzhang
b6ee3dedcc fix bug when a line only has one word ends with hyphen 2025-07-29 15:43:23 +08:00
Kinneyzhang
eb78e5316f fix _ local variable bug 2025-07-28 16:58:30 +08:00
Kinneyzhang
b6f6616b81 add chinese example 2025-07-28 00:33:53 +08:00
Kinneyzhang
a265357e41 update readme 2025-07-27 21:35:44 +08:00