@remigius42/morg 0.4.0 → 0.6.0

package/dist/markdownToOrg.js

@@ -7,6 +7,11 @@ import { uniorgStringify } from "uniorg-stringify";
 import { visit } from "unist-util-visit";
 import { transformMdastToUniorgAst } from "./core/mdastToUniorg/index.js";
 import { detectMarkdownStyle, STYLE_KEYWORD } from "./core/markdownStyle.js";
+import { escapeOrgMarkup } from "./core/markupBoundary.js";
+import { escapeLineSyntax } from "./core/lineSyntax.js";
+import { escapeFootnoteReferences } from "./core/footnoteReferences.js";
+import { escapeTablePipes } from "./core/tablePipes.js";
+import { requireBracedScripts } from "./core/bracedScripts.js";
 /**
  * Converts a Markdown string to an Org-mode string.
  * @param markdown The Markdown string to convert.
@@ -40,7 +45,7 @@ export function convertMarkdownToOrg(markdown, options = {}) {
         canonical
     ]));
     restoreOrgisms(uniorgAst, canonicalKeys);
-    // Phase 2c: formatting-as-structure — adjacent lists need two blank
+    // Phase 2c: formatting-as-structure. Adjacent lists need two blank
     // lines between them, or org's parser merges them into one list.
     separateAdjacentLists(uniorgAst);
     // Phase 2d: record the source's own style markers, so the return trip
@@ -52,6 +57,19 @@ export function convertMarkdownToOrg(markdown, options = {}) {
     if (options.preset?.applyToUniorg) {
         uniorgAst = options.preset.applyToUniorg(uniorgAst);
     }
+    // Phase 3b: literal footnote references, paragraph lines org would
+    // read as line syntax, literal markers org would read as markup, and
+    // markup touching a word character, need a zero-width space escape;
+    // a pipe in a table cell an entity
+    escapeTablePipes(uniorgAst);
+    escapeFootnoteReferences(uniorgAst);
+    escapeLineSyntax(uniorgAst);
+    escapeOrgMarkup(uniorgAst);
+    // Phase 3c: md text has no sub/superscripts; keep org from reading
+    // bare underscores and carets as such. Checked on the text as
+    // rendered: after the preset, whose text rewrites (wikilink aliases)
+    // org parses too, and after the escapes, next to which org reads them
+    requireBracedScripts(uniorgAst);
     // Phase 4: Render uniorg-ast to Org-mode string
     const processor = unified().use(uniorgStringify);
     const orgContent = processor.stringify(uniorgAst);

package/dist/normalize.d.ts

@@ -1,7 +1,7 @@
 import type { MarkdownToOrgOptions, OrgToMarkdownOptions } from "./options.js";
 /**
  * The full option set of both directions: normalization uses the same
- * configuration as conversion — Convergence is per-config (see ADR
+ * configuration as conversion, because Convergence is per-config (ADR
  * 0002), so a file must be normalized with the exact config (preset,
  * style, key names, toggles) it will be converted with.
  */
@@ -9,7 +9,7 @@ export type NormalizeOptions = MarkdownToOrgOptions & OrgToMarkdownOptions;
 /**
  * Normalizes a Markdown string to morg's canonical form: one full round
  * trip (`md → org → md`), whose output is a fixed point (see ADR 0001).
- * This canonicalizes, it does not just re-style — org-isms and md-isms
+ * This canonicalizes, it does not just re-style: org-isms and md-isms
  * are rewritten the same way a conversion would rewrite them.
  * @param markdown The Markdown string to normalize.
  * @param options Warning callback and dialect preset.

package/dist/normalize.js

@@ -3,7 +3,7 @@ import { convertOrgToMarkdown } from "./orgToMarkdown.js";
 /**
  * Normalizes a Markdown string to morg's canonical form: one full round
  * trip (`md → org → md`), whose output is a fixed point (see ADR 0001).
- * This canonicalizes, it does not just re-style — org-isms and md-isms
+ * This canonicalizes, it does not just re-style: org-isms and md-isms
  * are rewritten the same way a conversion would rewrite them.
  * @param markdown The Markdown string to normalize.
  * @param options Warning callback and dialect preset.

package/dist/options.d.ts

@@ -33,7 +33,7 @@ export interface MarkdownToOrgOptions {
     /**
      * Custom names for org-ism `key::` lines, canonical → custom (e.g.
      * `{ todo: "state" }`). Must match the mapping the file was written
-     * with — Convergence is per-config (ADR 0002).
+     * with, since Convergence is per-config (ADR 0002).
      */
     orgismKeys?: Record<string, string>;
     /** Called for each construct dropped without an equivalent. */
@@ -46,7 +46,7 @@ export interface MarkdownToOrgOptions {
  * Canonical form is parameterized by these (see ADR 0001): round trips
  * must use the same style, and files formatted under one style are not
  * a fixed point under another. Defaults: `-` bullet, `*` emphasis /
- * `*` strong (i.e. `**bold**`), backtick fences, `-` rule — prettier's
+ * `*` strong (i.e. `**bold**`), backtick fences, `-` rule: prettier's
  * choices except emphasis (`emphasis: "_"` aligns with prettier).
  */
 export interface MarkdownStyleOptions {
@@ -78,7 +78,7 @@ export interface OrgToMarkdownOptions {
      * Render Org constructs without a Markdown equivalent as raw HTML
      * (`<u>`, `<sup>`, `<sub>`, `<dl>`) instead of keeping their org markup
      * verbatim. HTML round-trips as a preserved md-ism (export blocks and
-     * snippets), not back to native org markup — a one-way door unless
+     * snippets), not back to native org markup, a one-way door unless
      * the return trip enables its inverse, `interpretHtml`.
      * Default: `false`.
      */

package/dist/orgToMarkdown.js

@@ -1,11 +1,24 @@
 import { unified } from "unified";
-import uniorgParse from "uniorg-parse";
 import remarkStringify from "remark-stringify";
 import remarkGfm from "remark-gfm";
 import remarkFrontmatter from "remark-frontmatter";
 import remarkMath from "remark-math";
 import { transformUniorgAstToMdast } from "./core/uniorgToMdast/index.js";
 import { takeRecordedStyle } from "./core/markdownStyle.js";
+import { unescapeOrgMarkup } from "./core/markupBoundary.js";
+import { unescapeLineSyntax } from "./core/lineSyntax.js";
+import { unescapeFootnoteReferences } from "./core/footnoteReferences.js";
+import { parseOrg } from "./core/bracedScripts.js";
+import { dropUnderscoreBulletGuards, guardUnderscoreBullets } from "./core/underscoreBullets.js";
+// a list item's paragraph after its nested list needs a blank line, or
+// md reads it as a lazy continuation of the nested list's last item
+function separateTextAfterNestedList(left, right, parent) {
+    return parent.type === "listItem" &&
+        left.type === "list" &&
+        right.type === "paragraph"
+        ? 1
+        : undefined;
+}
 /**
  * Converts an Org-mode string to a Markdown string.
  * @param org The Org-mode string to convert.
@@ -14,10 +27,17 @@ import { takeRecordedStyle } from "./core/markdownStyle.js";
  */
 export function convertOrgToMarkdown(org, options = {}) {
     // Phase 1: Parse Org-mode to uniorg-ast
-    let uniorgAst = unified().use(uniorgParse).parse(org);
-    // Phase 1b: a recorded style is morg's own (ADR 0004) — consume it so
+    // md text has no scripts, so ^:{} is implied there and consumed here
+    // (see markdownToOrg); uniorg misreads `_.` lines (see underscoreBullets)
+    let uniorgAst = parseOrg(guardUnderscoreBullets(org));
+    dropUnderscoreBulletGuards(uniorgAst);
+    // Phase 1b: a recorded style is morg's own (ADR 0004), so consume it so
     // it does not travel on as frontmatter; explicit options still win
     const recordedStyle = takeRecordedStyle(uniorgAst);
+    // Phase 1c: markdown needs no zero-width space escapes (inverse of md→org)
+    unescapeOrgMarkup(uniorgAst);
+    unescapeLineSyntax(uniorgAst);
+    unescapeFootnoteReferences(uniorgAst);
     // Phase 2: Extract dialect preset conventions, if any
     if (options.preset?.extractFromUniorg) {
         uniorgAst = options.preset.extractFromUniorg(uniorgAst);
@@ -46,11 +66,15 @@ export function convertOrgToMarkdown(org, options = {}) {
         rule: "-",
         ...recordedStyle,
         ...options.markdownStyle,
+        join: [separateTextAfterNestedList],
         handlers: {
             // key:: value blocks and preset inline passthroughs (e.g.
-            // wikilinks) are emitted verbatim, unescaped
+            // wikilinks) are emitted verbatim, unescaped; only a pipe inside
+            // a table cell is escaped, or it would split the cell
             keyValue: (node) => node.value,
-            verbatimInline: (node) => node.value
+            verbatimInline: (node, _parent, state) => state.stack.includes("tableCell")
+                ? node.value.replaceAll("|", "\\|")
+                : node.value
         }
     })
         .use(remarkGfm)

package/dist/presets/logseq.js

@@ -191,7 +191,7 @@ function repairHighlights(uniorgAst) {
 }
 // Logseq page and block references: [[page]] stays a wikilink,
 // [[page][label]] becomes [label]([[page]]), [[((uuid))][label]]
-// becomes [label](((uuid))) — emitted unescaped via verbatim-inline
+// becomes [label](((uuid))), emitted unescaped via verbatim-inline
 const BLOCK_REF_RE = /^\(\(.*\)\)$/;
 function pageRefValue(label, target) {
     if (!label) {

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@remigius42/morg",
-  "version": "0.4.0",
+  "version": "0.6.0",
   "description": "Bidirectional Markdown ↔ Org-mode converter with round-trip convergence",
   "keywords": [
     "org-mode",
@@ -46,6 +46,7 @@
     "prepare": "husky",
     "prepack": "npm run build",
     "version": "node scripts/release-changelog.mjs && git add CHANGELOG.md",
+    "postversion": "printf '\\nNext:\\n  git push origin v%s\\n  wait for the Release run to pass\\n  git push\\n' \"$npm_package_version\"",
     "build": "tsc -p tsconfig.build.json",
     "build:web": "vite build web",
     "dev:web": "vite web",
@@ -61,6 +62,8 @@
     "lint:fix": "npm run lint:fix:prettier && npm run lint:fix:eslint",
     "lint:fix:prettier": "prettier --write '**/*'",
     "lint:fix:eslint": "eslint --fix '**/*.{ts,mts,js}'",
+    "perf": "npm run build && node scripts/perf/perf.mjs",
+    "perf:record": "node scripts/perf/perf.mjs --record",
     "test:unit": "vitest",
     "test:unit:ci": "vitest run --reporter default --coverage",
     "test:coverage": "vitest run --coverage",