From 53b20e6adc962cb7bc317fa508fdd543760d2ef8 Mon Sep 17 00:00:00 2001
From: "Steven R. Loomis" <srl295@gmail.com>
Date: Thu, 13 Feb 2025 16:47:12 -0600
Subject: [PATCH] feat(developer): use peggy grammar in from/to transform

Fixes: #13175
---
 developer/src/kmc-ldml/.gitignore               |  1 +
 developer/src/kmc-ldml/build.sh                 |  3 +++
 developer/src/kmc-ldml/package.json             |  2 +-
 .../src/compiler/ldml-compiler-messages.ts      |  5 +++++
 developer/src/kmc-ldml/src/compiler/tran.ts     | 17 +++++++++++++++++
 developer/src/kmc-ldml/src/util/abnf/abnf.ts    |  6 ++++++
 developer/src/kmc-ldml/tsconfig.json            |  3 ++-
 package-lock.json                               |  7 ++++++-
 8 files changed, 41 insertions(+), 3 deletions(-)

diff --git a/developer/src/kmc-ldml/.gitignore b/developer/src/kmc-ldml/.gitignore
index c27c0f4e3b..0a7af4d6dd 100644
--- a/developer/src/kmc-ldml/.gitignore
+++ b/developer/src/kmc-ldml/.gitignore
@@ -1 +1,2 @@
 src/util/abnf/*.pegjs
+src/util/abnf/*.js
diff --git a/developer/src/kmc-ldml/build.sh b/developer/src/kmc-ldml/build.sh
index abd213e7c4..5335ca69c0 100755
--- a/developer/src/kmc-ldml/build.sh
+++ b/developer/src/kmc-ldml/build.sh
@@ -54,9 +54,12 @@ function do_build_abnf() {
     base=$(basename "$file" .abnf)
     peg="$base.pegjs"
     outfile="./src/util/abnf/$peg"
+    outjs="./src/util/abnf/$base.js"
     if [ ! -f "$outfile" ]; then
       printf "${COLOR_GREY}abnf_gen ${COLOR_PURPLE}${base}.abnf -> ${peg}${COLOR_RESET}"
       npx -p abnf abnf_gen  "$file" -o "$outfile"
+      printf "${COLOR_GREY}abnf_gen ${COLOR_PURPLE}${peg} -> ${base}.js${COLOR_RESET}"
+      npx peggy "$outfile" -o "$outjs" --format es --dts
     fi
   done
 }
diff --git a/developer/src/kmc-ldml/package.json b/developer/src/kmc-ldml/package.json
index 08708b64cc..5375e2673a 100644
--- a/developer/src/kmc-ldml/package.json
+++ b/developer/src/kmc-ldml/package.json
@@ -29,7 +29,6 @@
     "@keymanapp/keyman-version": "*",
     "@keymanapp/kmc-kmn": "*",
     "@keymanapp/ldml-keyboard-constants": "*",
-    "peggy": "^4.2.0",
     "semver": "^7.5.4"
   },
   "devDependencies": {
@@ -44,6 +43,7 @@
     "chalk": "^2.4.2",
     "common-tags": "^1.8.2",
     "mocha": "^8.4.0",
+    "peggy": "^4.2.0",
     "typescript": "^5.4.5"
   },
   "mocha": {
diff --git a/developer/src/kmc-ldml/src/compiler/ldml-compiler-messages.ts b/developer/src/kmc-ldml/src/compiler/ldml-compiler-messages.ts
index 3577a2e110..7a78126423 100644
--- a/developer/src/kmc-ldml/src/compiler/ldml-compiler-messages.ts
+++ b/developer/src/kmc-ldml/src/compiler/ldml-compiler-messages.ts
@@ -258,4 +258,9 @@ export class LdmlCompilerMessages {
     \`<transform from="($[fromSet])" to="$[1:toSet]"/>\`.
   `);
 
+  static ERROR_UnparseableTransformTo   = SevErrorTransform | 0x06;
+  static Error_UnparseableTransformTo   = (o: { to: string, message: string }) =>
+  m(this.ERROR_UnparseableTransformTo,    `Invalid transform to="${def(o.to)}": "${def(o.message)}"`);
+
+
 }
diff --git a/developer/src/kmc-ldml/src/compiler/tran.ts b/developer/src/kmc-ldml/src/compiler/tran.ts
index 1a91e8d382..ea55ec2ae8 100644
--- a/developer/src/kmc-ldml/src/compiler/tran.ts
+++ b/developer/src/kmc-ldml/src/compiler/tran.ts
@@ -17,6 +17,7 @@ import LKTransforms = LDMLKeyboard.LKTransforms;
 import { verifyValidAndUnique } from "../util/util.js";
 import { LdmlCompilerMessages } from "./ldml-compiler-messages.js";
 import { Substitutions, SubstitutionUse } from "./substitution-tracker.js";
+import { transform_from_parse, transform_to_parse } from "../util/abnf/abnf.js";
 
 type TransformCompilerType = 'simple' | 'backspace';
 
@@ -144,8 +145,24 @@ export abstract class TransformCompiler<T extends TransformCompilerType, TranBas
     let cookedFrom = transform.from;
 
     // check for incorrect \uXXXX escapes. Do this before substituting markers or sets.
+    // We run this first because it's more helpful than the ABNF.
     cookedFrom = this.checkEscapes(cookedFrom); // check for \uXXXX escapes before normalizing
 
+    // run the parser here next
+    try {
+      if (cookedFrom != null) {
+        transform_from_parse(cookedFrom);
+      }
+    } catch (e) {
+      this.callbacks.reportMessage(LdmlCompilerMessages.Error_UnparseableTransformFrom({ from: cookedFrom, message: e.toString() }));
+    }
+
+    try {
+      transform_to_parse(transform.to || '');
+    } catch (e) {
+      this.callbacks.reportMessage(LdmlCompilerMessages.Error_UnparseableTransformTo({ to: transform.to || '', message: e.toString() }));
+    }
+
     cookedFrom = sections.vars.substituteStrings(cookedFrom, sections, true);
     const mapFrom = LdmlKeyboardTypes.VariableParser.CAPTURE_SET_REFERENCE.exec(cookedFrom);
     const mapTo = LdmlKeyboardTypes.VariableParser.MAPPED_SET_REFERENCE.exec(transform.to || '');
diff --git a/developer/src/kmc-ldml/src/util/abnf/abnf.ts b/developer/src/kmc-ldml/src/util/abnf/abnf.ts
index e69de29bb2..30cbaaf4f3 100644
--- a/developer/src/kmc-ldml/src/util/abnf/abnf.ts
+++ b/developer/src/kmc-ldml/src/util/abnf/abnf.ts
@@ -0,0 +1,6 @@
+/**
+ * Re-export the generated parsers.
+ */
+
+export { parse as transform_to_parse } from './transform-to-required.js';
+export { parse as transform_from_parse } from './transform-from-required.js';
diff --git a/developer/src/kmc-ldml/tsconfig.json b/developer/src/kmc-ldml/tsconfig.json
index 36d915f074..9f17f7898f 100644
--- a/developer/src/kmc-ldml/tsconfig.json
+++ b/developer/src/kmc-ldml/tsconfig.json
@@ -7,7 +7,8 @@
     "baseUrl": ".",
   },
   "include": [
-    "src/**/*.ts"
+    "src/**/*.ts",
+    "src/util/abnf/*.js",
   ],
   "references": [
     { "path": "../../../common/web/keyman-version" },
diff --git a/package-lock.json b/package-lock.json
index 012a3ea53b..4e340f8af4 100644
--- a/package-lock.json
+++ b/package-lock.json
@@ -1281,7 +1281,6 @@
         "@keymanapp/keyman-version": "*",
         "@keymanapp/kmc-kmn": "*",
         "@keymanapp/ldml-keyboard-constants": "*",
-        "peggy": "^4.2.0",
         "semver": "^7.5.4"
       },
       "devDependencies": {
@@ -1296,6 +1295,7 @@
         "chalk": "^2.4.2",
         "common-tags": "^1.8.2",
         "mocha": "^8.4.0",
+        "peggy": "^4.2.0",
         "typescript": "^5.4.5"
       }
     },
@@ -3664,6 +3664,7 @@
       "version": "1.3.5",
       "resolved": "https://registry.npmjs.org/@peggyjs/from-mem/-/from-mem-1.3.5.tgz",
       "integrity": "sha512-oRyzXE7nirAn+5yYjCdWQHg3EG2XXcYRoYNOK8Quqnmm+9FyK/2YWVunwudlYl++M3xY+gIAdf0vAYS+p0nKfQ==",
+      "dev": true,
       "dependencies": {
         "semver": "7.6.3"
       },
@@ -3675,6 +3676,7 @@
       "version": "7.6.3",
       "resolved": "https://registry.npmjs.org/semver/-/semver-7.6.3.tgz",
       "integrity": "sha512-oVekP1cKtI+CTDvHWYFUcMtsK/00wmAEfyqKfNdARm8u1wNVhSgaX7A8d4UuIlUI5e84iEwOhs7ZPYRmzU9U6A==",
+      "dev": true,
       "bin": {
         "semver": "bin/semver.js"
       },
@@ -12433,6 +12435,7 @@
       "version": "4.2.0",
       "resolved": "https://registry.npmjs.org/peggy/-/peggy-4.2.0.tgz",
       "integrity": "sha512-ZjzyJYY8NqW8JOZr2PbS/J0UH/hnfGALxSDsBUVQg5Y/I+ZaPuGeBJ7EclUX2RvWjhlsi4pnuL1C/K/3u+cDeg==",
+      "dev": true,
       "license": "MIT",
       "dependencies": {
         "@peggyjs/from-mem": "1.3.5",
@@ -12450,6 +12453,7 @@
       "version": "12.1.0",
       "resolved": "https://registry.npmjs.org/commander/-/commander-12.1.0.tgz",
       "integrity": "sha512-Vw8qHK3bZM9y/P10u3Vib8o/DdkvA2OtPtZvD871QKjy74Wj1WSKFILMPRPSdUSx5RFK1arlJzEtA4PkFgnbuA==",
+      "dev": true,
       "engines": {
         "node": ">=18"
       }
@@ -13615,6 +13619,7 @@
       "version": "0.8.0",
       "resolved": "https://registry.npmjs.org/source-map-generator/-/source-map-generator-0.8.0.tgz",
       "integrity": "sha512-psgxdGMwl5MZM9S3FWee4EgsEaIjahYV5AzGnwUvPhWeITz/j6rKpysQHlQ4USdxvINlb8lKfWGIXwfkrgtqkA==",
+      "dev": true,
       "engines": {
         "node": ">= 10"
       }