chore: minor change as per reviews

packages/unified-latex-prettier/libs/printer/printer.ts

@@ -84,7 +84,7 @@ export function printLatexAst(
         case "whitespace":
             return line;
         case "hash_number":
-            return "#" + node.number.toString(10);
+            return `#${node.number}`;
         default:
             console.warn("Printing unknown type", node);
             return printRaw(node);

packages/unified-latex-util-argspec/libs/argspec-parser.ts

@@ -34,14 +34,6 @@ export function printRaw(
 function printRawInner(node: ArgSpec.Node) {
     const decorators = getDecorators(node);
     let spec = decorators;
-    function appendDefaultArg() {
-        if (node.defaultArg) {
-            spec = appendTokenOrGroup(spec, node.defaultArg);
-        }
-    }
-    function appendCollection(collection: string[]) {
-        spec += printTokenOrCollection(collection);
-    }
     const type = node.type;
     switch (type) {
         case "body":
@@ -58,7 +50,9 @@ function printRawInner(node: ArgSpec.Node) {
                 spec += node.defaultArg ? "D" : "d";
                 spec += node.openBrace + node.closeBrace;
             }
-            appendDefaultArg();
+            if (node.defaultArg) {
+                spec = appendTokenOrGroup(spec, node.defaultArg);
+            }
             return spec;
         case "mandatory":
             // {...} is the default enclosure for mandatory arguments
@@ -68,20 +62,22 @@ function printRawInner(node: ArgSpec.Node) {
                 spec += node.defaultArg ? "R" : "r";
                 spec += node.openBrace + node.closeBrace;
             }
-            appendDefaultArg();
+            if (node.defaultArg) {
+                spec = appendTokenOrGroup(spec, node.defaultArg);
+            }
             return spec;
         case "embellishment":
             spec += node.defaultArgs ? "E" : "e";
-            appendCollection(node.tokens);
+            spec += printTokenOrCollection(node.tokens);
             if (node.defaultArgs) {
-                appendCollection(node.defaultArgs);
+                spec += printTokenOrCollection(node.defaultArgs);
             }
             return spec;
         case "verbatim":
             return spec + "v" + node.openBrace;
         case "until": {
             spec += "u";
-            appendCollection(node.stopTokens);
+            spec += printTokenOrCollection(node.stopTokens);
             return spec;
         }
         default:
@@ -95,8 +91,22 @@ function printRawInner(node: ArgSpec.Node) {
  * This function will reconstruct a representative in an inverse image of token_or_group
  * for a given array of strings, and append it to a given string.
  * In order to avoid parsing ambiguity, we force enclose the representative with braces in some case.
- * For instance, if the given string ends with a control word such as "\asdf", and if the representative is a
- * whitespace where we are in a circumstance where no whitespaces are allowed.
+ *
+ * Examples)
+ * Appending a `token_or_group` representing an embellishment tokens, with existingString `e`.
+ * token of several chars "ab" --> {ab} will be appended, so the result will be `e{ab}`.
+ * token of single char "^" ---> ^ will be appended, so the result will be `e^`.
+ * token of single char " " ---> { } will be appended, so the result will be `e{ }`.
+ *
+ * Appending a `token_or_group` representing a bracespec of a delimited argument spec.
+ * Already constructed a string until the opening brace, existingString = "r\open".
+ * If the bracespec is a single alphabetric char "a", then we need to append it with braces, so the result will be "r\open{a}".
+ *
+ * Appending an embellishment token, with existingString = "e{a\token".
+ * If a next token is again a control world, say \anotherToken, we can append it as-is. "e{a\token\anotherToken".
+ * If a next token is a single char, say "b", we may either enclose it with braces or separate it with a space,
+ * because in this circumstance, space can be used to separate tokens (as conveyed by allowWhitespace = true).
+ * Then the result will be "e{a\token b".
  */
 function appendTokenOrGroup(
     existingString: string,

packages/unified-latex-util-arguments/libs/gobble-single-argument.ts

@@ -151,9 +151,6 @@ export function gobbleSingleArgument(
         }
         case "until": {
             const stopTokens = argSpec.stopTokens.map(parseToken);
-            // TODO: in order to match xparse's behavior, multiple spaces at the start
-            // or in a middle should be collapsed to a single whitespace token,
-            // and spaces at the end should be ignored.
             let nextStartPos = startPos;
             let bracePos: [number, number] | undefined;
             while (nextStartPos < nodes.length) {
@@ -236,7 +233,6 @@ function cloneStringNode(node: Ast.String, content: string): Ast.String {
     return Object.assign({}, node, { content });
 }
 
-type Braces = string | Ast.Macro | Ast.Whitespace;
 /**
  * Find the position of the open brace and the closing brace.
  * Returns undefined if the brace isn't found.
@@ -246,8 +242,8 @@ type Braces = string | Ast.Macro | Ast.Whitespace;
 function findBracePositions(
     nodes: Ast.Node[],
     startPos: number,
-    openMark?: Braces,
-    closeMark?: Braces,
+    openMark?: string | Ast.Macro | Ast.Whitespace,
+    closeMark?: string | Ast.Macro | Ast.Whitespace,
     endPos?: number
 ): [number, number] | undefined {
     let openMarkPos: number | undefined = startPos;
@@ -285,9 +281,15 @@ function findBracePositions(
     return [openMarkPos, closeMarkPos];
 }
 
+/**
+ * Find the position of the delimiter in `nodes`. A delimiter can either be a single character,
+ * or a single control word or a symbol (represented as Ast.Macro). Returns `undefined`
+ * if it cannot be found. If a search found a character delimiter in a middle of a string,
+ * this function may mutate `nodes` to split the string.
+ */
 function findDelimiter(
     nodes: Ast.Node[],
-    token: Braces,
+    token: string | Ast.Macro | Ast.Whitespace,
     startPos: number,
     endPos?: number
 ): number | undefined {
@@ -338,7 +340,7 @@ function findDelimiter(
 
 function parseToken(
     str: string | undefined
-): string | Ast.Whitespace | Ast.Macro {
+): string | Ast.Macro | Ast.Whitespace {
     if (!str) {
         return "";
     }

packages/unified-latex-util-arguments/tests/gobble-single-argument.test.ts

@@ -601,8 +601,8 @@ describe("unified-latex-util-arguments", () => {
         expect(nodes).toEqual([{ content: "yx", type: "string" }]);
     });
     it("can gobble an 'until' argument with multiple stop tokens", () => {
-        let argspec = parseArgspec("u{a \\bcd}")[0];
-        value = "asdf asydfxya{x}sa \\bcd2df";
+        let argspec = parseArgspec("u{| \\stop}")[0];
+        value = "|this_bar_is_not_a_stop|{token}this| \\stop is.";
         file = processLatexToAstViaUnified().processSync({ value });
         let nodes = trimRenderInfo((file.result as any).content) as Ast.Node[];
         expect(gobbleSingleArgument(nodes, argspec)).toEqual({
@@ -611,28 +611,24 @@ describe("unified-latex-util-arguments", () => {
                 content: [
                     // Due to a current implementation of gobbleSingleArgument,
                     // we may introduce extra string split during the search.
-                    { type: "string", content: "a" },
-                    { type: "string", content: "sdf" },
-                    { type: "whitespace" },
-                    { type: "string", content: "a" },
-                    { type: "string", content: "sydfxy" },
-                    { type: "string", content: "a" },
+                    { type: "string", content: "|" },
+                    { type: "string", content: "this_bar_is_not_a_stop" },
+                    { type: "string", content: "|" },
                     {
                         type: "group",
-                        content: [{ type: "string", content: "x" }],
+                        content: [{ type: "string", content: "token" }],
                     },
-                    { type: "string", content: "s" },
+                    { type: "string", content: "this" },
                 ],
                 openMark: "",
-                closeMark: "a \\bcd",
+                closeMark: "| \\stop",
             },
-            nodesRemoved: 11,
+            nodesRemoved: 8,
         });
         expect(nodes).toEqual([
-            {
-                type: "string",
-                content: "2df",
-            },
+            { type: "whitespace" },
+            { type: "string", content: "is" },
+            { type: "string", content: "." },
         ]);
     });
     it("gobbleSingleArgument gobbles non-punctuation delimited arguments", () => {
@@ -825,7 +821,7 @@ describe("unified-latex-util-arguments", () => {
             nodesRemoved: 2,
         });
     });
-    it("can gobble optional argument with default argument", () => {
+    it("can skip optional argument with default argument", () => {
         const expectNoMatch = (ast: Ast.Node[]) => {
             expect(
                 gobbleSingleArgument(ast, parseArgspec("O{default}")[0])

packages/unified-latex-util-macros/libs/newcommand.ts

@@ -252,24 +252,24 @@ export function createMacroExpander(
                 // some options that is provided to whatever function that called this to
                 // the below parse call. Note that `parse` is done in several passes, and we
                 // may be able to cache result of a first few passes that aren't context-dependent.
-                const subst = parse(defaultArg).content;
-                const hasSubstitutions = attachHashNumbers(subst);
+                const sub = parse(defaultArg).content;
+                const hasSubstitutions = attachHashNumbers(sub);
 
                 if (!hasSubstitutions) {
-                    return subst;
+                    return sub;
                 }
 
                 stack.push(hashNum);
                 try {
-                    expandArgs(subst);
+                    expandArgs(sub);
 
                     if (lastSelfReference !== hashNum) {
-                        return subst;
+                        return sub;
                     }
                     // At this point, we have encountered #n while expanding #n.
                     // Check if we got exactly #n by expanding #n,
                     // in which case we should return the -NoValue-.
-                    if (`#${hashNum}` === printRaw(subst)) {
+                    if (`#${hashNum}` === printRaw(sub)) {
                         // We are good, clear the last self-reference variable
                         lastSelfReference = null;
                         return emptyArg();

packages/unified-latex-util-macros/tests/macro-expansion.test.ts

@@ -1,11 +1,10 @@
+import util from "util";
 import { m } from "@unified-latex/unified-latex-builder";
 import * as Ast from "@unified-latex/unified-latex-types";
-import util from "util";
-import { attachMacroArgsInArray } from "../../unified-latex-util-arguments/libs/attach-arguments";
-
 import * as latexParser from "@unified-latex/unified-latex-util-parse";
 import { printRaw } from "@unified-latex/unified-latex-util-print-raw";
 import { trimRenderInfo } from "@unified-latex/unified-latex-util-render-info";
+import { attachMacroArgsInArray } from "../../unified-latex-util-arguments/libs/attach-arguments";
 import { expandMacros } from "../libs/expand-macros";
 import { createMacroExpander } from "../libs/newcommand";
 
@@ -169,29 +168,29 @@ describe("unified-latex-utils-macros", () => {
     });
 
     it("Can substitute default arguments that cross-references each other", () => {
-        let substitutionBody = latexParser.parse("#1/#2").content;
+        let substitutionBody = latexParser.parse("#1.#2").content;
         // This macro defines the args that will be substituted
         let macro = parseXxxMacro("\\xxx", "O{#2} O{#1}");
 
         const expander = createMacroExpander(substitutionBody, "O{#2} O{#1}");
-        expect(printRaw(expander(macro))).toEqual("/");
+        expect(printRaw(expander(macro))).toEqual(".");
     });
 
     it("Can substitute default arguments referencing other default arguments that cross-references each other", () => {
-        let substitutionBody = latexParser.parse("#1/#2/#3/#4").content;
+        let substitutionBody = latexParser.parse("#1.#2.#3.#4").content;
         // This macro defines the args that will be substituted
         let macro = parseXxxMacro("\\xxx", "O{a#2#4} O{#3} O{#2} O{b#2}");
 
         const expander = createMacroExpander(
             substitutionBody,
             "O{a#2#4} O{#3} O{#2} O{b#2}"
         );
-        expect(printRaw(expander(macro))).toEqual("ab///b");
+        expect(printRaw(expander(macro))).toEqual("ab...b");
     });
 
     it("Can substitute default arguments with complex dependency graph", () => {
         let substitutionBody = latexParser.parse(
-            "#1/#2/#3/#4/#5/#6/#7/#8/#9"
+            "#1.#2.#3.#4.#5.#6.#7.#8.#9"
         ).content;
         // This macro defines the args that will be substituted
         let macro = parseXxxMacro(
@@ -203,7 +202,7 @@ describe("unified-latex-utils-macros", () => {
             substitutionBody,
             "D<>{#2} E{^_}{{#1}{#4}} O{#5} E{<>,}{{#3}{a#3b#1c#7}{#9d#4}} O{#6e} D\\a\\b{f#8}"
         );
-        expect(printRaw(expander(macro))).toEqual("/////abc9d/9d/abc9de/9");
+        expect(printRaw(expander(macro))).toEqual(".....abc9d.9d.abc9de.9");
     });
 
     it("Can expand macro", () => {

packages/unified-latex-util-pegjs/grammars/xparse-argspec.pegjs

@@ -89,6 +89,8 @@ required
     / "r" braceSpec:brace_spec { return createNode("mandatory", braceSpec); }
 
 // An "until" argument gobbles tokens until the specified stop token(s). Until token allows whitespace.
+// TODO: in order to match xparse's behavior, multiple spaces at the start or in a middle
+// should be collapsed to a single whitespace token, and spaces at the end should be ignored.
 until
     = "u" stopTokens:(x:token { return [x] } / '{' @(token_or_whitespace+) '}') {
             return createNode("until", { stopTokens });