@sasjs/lint 5.3.1 → 5.4.0

package/utils/getDataSectionsDetail.d.ts

@@ -3,6 +3,20 @@ interface DataSectionsDetail {
     start: number;
     end: number;
 }
+/**
+ * A data section: the statement that opens it, and the record list it holds.
+ *
+ * The opener is matched against the line's CODE, with literals and comments
+ * blanked, so a `;cards;` inside a quoted string or a comment opens nothing -
+ * and against statement boundaries rather than the line start, so
+ * `input a b; cards;` on one line still opens one. The records themselves are
+ * not code: a quote in a record is data, so the scan state is reset at the
+ * opener and the records are matched as plain lines.
+ *
+ * @param {string} text - the text to read.
+ * @param {LintConfig} config - the configuration in force.
+ * @returns {DataSectionsDetail[]} the line ranges each section covers.
+ */
 export declare const getDataSectionsDetail: (text: string, config: LintConfig) => DataSectionsDetail[];
 export declare const checkIsDataLine: (dataSections: DataSectionsDetail[], lineIndex: number) => boolean;
 export {};

package/utils/getDataSectionsDetail.js

@@ -1,14 +1,29 @@
 "use strict";
 Object.defineProperty(exports, "__esModule", { value: true });
 exports.checkIsDataLine = exports.getDataSectionsDetail = void 0;
+const scanLine_1 = require("./scanLine");
 const splitText_1 = require("./splitText");
+/**
+ * A data section: the statement that opens it, and the record list it holds.
+ *
+ * The opener is matched against the line's CODE, with literals and comments
+ * blanked, so a `;cards;` inside a quoted string or a comment opens nothing -
+ * and against statement boundaries rather than the line start, so
+ * `input a b; cards;` on one line still opens one. The records themselves are
+ * not code: a quote in a record is data, so the scan state is reset at the
+ * opener and the records are matched as plain lines.
+ *
+ * @param {string} text - the text to read.
+ * @param {LintConfig} config - the configuration in force.
+ * @returns {DataSectionsDetail[]} the line ranges each section covers.
+ */
 const getDataSectionsDetail = (text, config) => {
     const dataSections = [];
     const lines = (0, splitText_1.splitText)(text, config);
-    const dataSectionStartRegex1 = new RegExp('^(datalines;)|(cards;)|(parmcards;)');
+    const opener = /(?:^|;)\s*(datalines|cards|parmcards)(4?);/i;
     const dataSectionEndRegex1 = new RegExp(';');
-    const dataSectionStartRegex2 = new RegExp('^(datalines4;)|(cards4;)|(parmcards4;)');
     const dataSectionEndRegex2 = new RegExp(';;;;');
+    let state = (0, scanLine_1.cleanState)();
     let dataSectionStarted = false;
     let dataSectionStartIndex = -1;
     let dataSectionEndRegex = dataSectionEndRegex1;
@@ -17,19 +32,20 @@ const getDataSectionsDetail = (text, config) => {
             if (dataSectionEndRegex.test(line)) {
                 dataSections.push({ start: dataSectionStartIndex, end: index });
                 dataSectionStarted = false;
+                state = (0, scanLine_1.cleanState)();
             }
+            return;
         }
-        else {
-            if (dataSectionStartRegex1.test(line)) {
-                dataSectionStarted = true;
-                dataSectionStartIndex = index;
-                dataSectionEndRegex = dataSectionEndRegex1;
-            }
-            else if (dataSectionStartRegex2.test(line)) {
-                dataSectionStarted = true;
-                dataSectionStartIndex = index;
-                dataSectionEndRegex = dataSectionEndRegex2;
-            }
+        const scanned = (0, scanLine_1.scanLine)(line, state);
+        state = scanned.endState;
+        const match = opener.exec(scanned.code);
+        if (match) {
+            dataSectionStarted = true;
+            dataSectionStartIndex = index;
+            dataSectionEndRegex = match[2]
+                ? dataSectionEndRegex2
+                : dataSectionEndRegex1;
+            state = (0, scanLine_1.cleanState)();
         }
     });
     return dataSections;

package/utils/getDataSectionsDetail.js.map

@@ -1 +1 @@
-{"version":3,"file":"getDataSectionsDetail.js","sourceRoot":"","sources":["../../src/utils/getDataSectionsDetail.ts"],"names":[],"mappings":";;;AACA,2CAAuC;AAOhC,MAAM,qBAAqB,GAAG,CAAC,IAAY,EAAE,MAAkB,EAAE,EAAE;IACxE,MAAM,YAAY,GAAyB,EAAE,CAAA;IAC7C,MAAM,KAAK,GAAG,IAAA,qBAAS,EAAC,IAAI,EAAE,MAAM,CAAC,CAAA;IAErC,MAAM,sBAAsB,GAAG,IAAI,MAAM,CACvC,qCAAqC,CACtC,CAAA;IACD,MAAM,oBAAoB,GAAG,IAAI,MAAM,CAAC,GAAG,CAAC,CAAA;IAC5C,MAAM,sBAAsB,GAAG,IAAI,MAAM,CACvC,wCAAwC,CACzC,CAAA;IACD,MAAM,oBAAoB,GAAG,IAAI,MAAM,CAAC,MAAM,CAAC,CAAA;IAE/C,IAAI,kBAAkB,GAAG,KAAK,CAAA;IAC9B,IAAI,qBAAqB,GAAG,CAAC,CAAC,CAAA;IAC9B,IAAI,mBAAmB,GAAG,oBAAoB,CAAA;IAE9C,KAAK,CAAC,OAAO,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE;QAC5B,IAAI,kBAAkB,EAAE,CAAC;YACvB,IAAI,mBAAmB,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;gBACnC,YAAY,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,qBAAqB,EAAE,GAAG,EAAE,KAAK,EAAE,CAAC,CAAA;gBAC/D,kBAAkB,GAAG,KAAK,CAAA;YAC5B,CAAC;QACH,CAAC;aAAM,CAAC;YACN,IAAI,sBAAsB,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;gBACtC,kBAAkB,GAAG,IAAI,CAAA;gBACzB,qBAAqB,GAAG,KAAK,CAAA;gBAC7B,mBAAmB,GAAG,oBAAoB,CAAA;YAC5C,CAAC;iBAAM,IAAI,sBAAsB,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;gBAC7C,kBAAkB,GAAG,IAAI,CAAA;gBACzB,qBAAqB,GAAG,KAAK,CAAA;gBAC7B,mBAAmB,GAAG,oBAAoB,CAAA;YAC5C,CAAC;QACH,CAAC;IACH,CAAC,CAAC,CAAA;IAEF,OAAO,YAAY,CAAA;AACrB,CAAC,CAAA;AArCY,QAAA,qBAAqB,yBAqCjC;AAEM,MAAM,eAAe,GAAG,CAC7B,YAAkC,EAClC,SAAiB,EACjB,EAAE;IACF,KAAK,MAAM,WAAW,IAAI,YAAY,EAAE,CAAC;QACvC,IAAI,SAAS,IAAI,WAAW,CAAC,KAAK,IAAI,SAAS,IAAI,WAAW,CAAC,GAAG;YAChE,OAAO,IAAI,CAAA;IACf,CAAC;IAED,OAAO,KAAK,CAAA;AACd,CAAC,CAAA;AAVY,QAAA,eAAe,mBAU3B"}
+{"version":3,"file":"getDataSectionsDetail.js","sourceRoot":"","sources":["../../src/utils/getDataSectionsDetail.ts"],"names":[],"mappings":";;;AACA,yCAAiD;AACjD,2CAAuC;AAOvC;;;;;;;;;;;;;GAaG;AACI,MAAM,qBAAqB,GAAG,CAAC,IAAY,EAAE,MAAkB,EAAE,EAAE;IACxE,MAAM,YAAY,GAAyB,EAAE,CAAA;IAC7C,MAAM,KAAK,GAAG,IAAA,qBAAS,EAAC,IAAI,EAAE,MAAM,CAAC,CAAA;IAErC,MAAM,MAAM,GAAG,6CAA6C,CAAA;IAC5D,MAAM,oBAAoB,GAAG,IAAI,MAAM,CAAC,GAAG,CAAC,CAAA;IAC5C,MAAM,oBAAoB,GAAG,IAAI,MAAM,CAAC,MAAM,CAAC,CAAA;IAE/C,IAAI,KAAK,GAAG,IAAA,qBAAU,GAAE,CAAA;IACxB,IAAI,kBAAkB,GAAG,KAAK,CAAA;IAC9B,IAAI,qBAAqB,GAAG,CAAC,CAAC,CAAA;IAC9B,IAAI,mBAAmB,GAAG,oBAAoB,CAAA;IAE9C,KAAK,CAAC,OAAO,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE;QAC5B,IAAI,kBAAkB,EAAE,CAAC;YACvB,IAAI,mBAAmB,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;gBACnC,YAAY,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,qBAAqB,EAAE,GAAG,EAAE,KAAK,EAAE,CAAC,CAAA;gBAC/D,kBAAkB,GAAG,KAAK,CAAA;gBAC1B,KAAK,GAAG,IAAA,qBAAU,GAAE,CAAA;YACtB,CAAC;YACD,OAAM;QACR,CAAC;QAED,MAAM,OAAO,GAAG,IAAA,mBAAQ,EAAC,IAAI,EAAE,KAAK,CAAC,CAAA;QACrC,KAAK,GAAG,OAAO,CAAC,QAAQ,CAAA;QAExB,MAAM,KAAK,GAAG,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,CAAA;QAEvC,IAAI,KAAK,EAAE,CAAC;YACV,kBAAkB,GAAG,IAAI,CAAA;YACzB,qBAAqB,GAAG,KAAK,CAAA;YAC7B,mBAAmB,GAAG,KAAK,CAAC,CAAC,CAAC;gBAC5B,CAAC,CAAC,oBAAoB;gBACtB,CAAC,CAAC,oBAAoB,CAAA;YACxB,KAAK,GAAG,IAAA,qBAAU,GAAE,CAAA;QACtB,CAAC;IACH,CAAC,CAAC,CAAA;IAEF,OAAO,YAAY,CAAA;AACrB,CAAC,CAAA;AAvCY,QAAA,qBAAqB,yBAuCjC;AAEM,MAAM,eAAe,GAAG,CAC7B,YAAkC,EAClC,SAAiB,EACjB,EAAE;IACF,KAAK,MAAM,WAAW,IAAI,YAAY,EAAE,CAAC;QACvC,IAAI,SAAS,IAAI,WAAW,CAAC,KAAK,IAAI,SAAS,IAAI,WAAW,CAAC,GAAG;YAChE,OAAO,IAAI,CAAA;IACf,CAAC;IAED,OAAO,KAAK,CAAA;AACd,CAAC,CAAA;AAVY,QAAA,eAAe,mBAU3B"}

package/utils/scanLine.d.ts

@@ -0,0 +1,59 @@
+/**
+ * Reads one line of a SAS program, carrying the state a break decision needs
+ * from one physical line to the next.
+ *
+ * A quoted literal, a block comment and a masking macro's argument text all
+ * continue onto the next line, so a caller that decides where a line may be
+ * broken - or where a statement really ends - needs to know which of them is
+ * open. The scan is one pass with those three states, because each affects the
+ * reading of the others: `%str(%')` is an apostrophe, and a scanner that read it
+ * as the start of a literal would carry a phantom string into every line below.
+ */
+export interface ScanState {
+    /** The quote character a literal is still open with, or `''`. */
+    quote: string;
+    /** Whether a block comment is still open. */
+    inComment: boolean;
+    /** Depth of the masking call the scan is inside. */
+    maskedDepth: number;
+}
+export declare const cleanState: () => ScanState;
+/** A block comment's span on one line, as absolute column indices. */
+export interface CommentSpan {
+    /** The column at which the comment opens. */
+    start: number;
+    /** The column of the closing marker's asterisk, or the line length. */
+    end: number;
+}
+/** A physical line, as the scanner reads it. */
+export interface ScannedLine {
+    /**
+     * The columns at which a newline may be inserted: every space that is not
+     * inside a quoted literal, inside a masking macro's parentheses, or inside a
+     * block comment.
+     */
+    positions: number[];
+    /** The line with every literal and comment character blanked to a space. */
+    code: string;
+    /** Whether the line carries any block comment, open or closed. */
+    hasComment: boolean;
+    /** Each block comment the line opens, with the span it covers. */
+    comments: CommentSpan[];
+    endState: ScanState;
+}
+/**
+ * The scan of a line that is not code, and so opens and closes nothing.
+ *
+ * A `datalines` record is data: a comment marker or an apostrophe in one is a
+ * character in a value, not the start of a comment or a literal. Reading such a
+ * line with the code scanner would carry that state into every line below it.
+ */
+export declare const noScan: () => ScannedLine;
+/**
+ * Reads one line, starting from the state the previous line ended in.
+ *
+ * A break is never taken inside a literal, because a SAS literal really does
+ * continue onto the next line (`"ab` / `cd"` is one value) and an inserted
+ * newline would become part of the string.
+ */
+export declare const scanLine: (line: string, state: ScanState) => ScannedLine;

package/utils/scanLine.js

@@ -0,0 +1,115 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.scanLine = exports.noScan = exports.cleanState = void 0;
+const cleanState = () => ({
+    quote: '',
+    inComment: false,
+    maskedDepth: 0
+});
+exports.cleanState = cleanState;
+/**
+ * The calls whose argument text is carried as DATA rather than as tokens.
+ *
+ * The masking macros - `%str(`, `%nrstr(`, `%bquote(`, `%nrbquote(`,
+ * `%superq(`, `%nrsuperq(`, `%quote(`, `%nrquote(`, `%unquote(` - return their
+ * argument as a value, so a newline inserted inside one becomes part of it.
+ * `%sysfunc(` and `%qsysfunc(` are the same problem one level down: the argument
+ * text is handed to a DATA step function, and for the text functions the
+ * whitespace is data. Measured on a live engine, breaking inside
+ * `%sysfunc(fput(&fid,%str(&name): %str(&value)))` wrote
+ * `Content-Type:   application/csv` where the macro meant one space - so no
+ * space inside either form is a break point.
+ */
+const maskStart = /%n?r?(?:str|bquote|superq|quote|unquote)\s*\(|%q?sysfunc\s*\(/iy;
+/**
+ * The scan of a line that is not code, and so opens and closes nothing.
+ *
+ * A `datalines` record is data: a comment marker or an apostrophe in one is a
+ * character in a value, not the start of a comment or a literal. Reading such a
+ * line with the code scanner would carry that state into every line below it.
+ */
+const noScan = () => ({
+    positions: [],
+    code: '',
+    hasComment: false,
+    comments: [],
+    endState: (0, exports.cleanState)()
+});
+exports.noScan = noScan;
+/**
+ * Reads one line, starting from the state the previous line ended in.
+ *
+ * A break is never taken inside a literal, because a SAS literal really does
+ * continue onto the next line (`"ab` / `cd"` is one value) and an inserted
+ * newline would become part of the string.
+ */
+const scanLine = (line, state) => {
+    const positions = [];
+    const code = new Array(line.length).fill(' ');
+    const comments = [];
+    let { quote, inComment, maskedDepth } = state;
+    let hasComment = inComment;
+    let commentStart = inComment ? 0 : -1;
+    for (let index = 0; index < line.length; index++) {
+        const character = line[index];
+        if (inComment) {
+            if (character === '*' && line[index + 1] === '/') {
+                inComment = false;
+                comments.push({ start: commentStart, end: index });
+                index++;
+                continue;
+            }
+            // A space inside a block comment is prose, not a break point: the
+            // contract above states it, and no caller wants one - the wrap refuses a
+            // comment-bearing line and `reflowComments` owns its text.
+            continue;
+        }
+        // Masked text is data, so it is read before anything that OPENS a literal
+        // or a comment: `%str(%')` is how a single quote is written, and reading
+        // that apostrophe as the start of a literal would carry a phantom string
+        // into every line below it.
+        if (maskedDepth) {
+            if (character === '(')
+                maskedDepth++;
+            else if (character === ')')
+                maskedDepth--;
+            continue;
+        }
+        if (quote) {
+            if (character === quote)
+                quote = '';
+            continue;
+        }
+        if (character === "'" || character === '"') {
+            quote = character;
+            continue;
+        }
+        if (character === '/' && line[index + 1] === '*') {
+            inComment = true;
+            hasComment = true;
+            commentStart = index;
+            index++;
+            continue;
+        }
+        maskStart.lastIndex = index;
+        if (maskStart.test(line)) {
+            maskedDepth = 1;
+            index = maskStart.lastIndex - 1;
+            continue;
+        }
+        if (character === ' ')
+            positions.push(index);
+        code[index] = character;
+    }
+    if (inComment)
+        comments.push({ start: commentStart, end: line.length });
+    return {
+        positions,
+        code: code.join(''),
+        hasComment,
+        comments,
+        endState: { quote, inComment, maskedDepth }
+    };
+};
+exports.scanLine = scanLine;
+//# sourceMappingURL=scanLine.js.map

package/utils/scanLine.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"scanLine.js","sourceRoot":"","sources":["../../src/utils/scanLine.ts"],"names":[],"mappings":";;;AAoBO,MAAM,UAAU,GAAG,GAAc,EAAE,CAAC,CAAC;IAC1C,KAAK,EAAE,EAAE;IACT,SAAS,EAAE,KAAK;IAChB,WAAW,EAAE,CAAC;CACf,CAAC,CAAA;AAJW,QAAA,UAAU,cAIrB;AAEF;;;;;;;;;;;;GAYG;AACH,MAAM,SAAS,GACb,iEAAiE,CAAA;AA2BnE;;;;;;GAMG;AACI,MAAM,MAAM,GAAG,GAAgB,EAAE,CAAC,CAAC;IACxC,SAAS,EAAE,EAAE;IACb,IAAI,EAAE,EAAE;IACR,UAAU,EAAE,KAAK;IACjB,QAAQ,EAAE,EAAE;IACZ,QAAQ,EAAE,IAAA,kBAAU,GAAE;CACvB,CAAC,CAAA;AANW,QAAA,MAAM,UAMjB;AAEF;;;;;;GAMG;AACI,MAAM,QAAQ,GAAG,CAAC,IAAY,EAAE,KAAgB,EAAe,EAAE;IACtE,MAAM,SAAS,GAAa,EAAE,CAAA;IAC9B,MAAM,IAAI,GAAa,IAAI,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;IACvD,MAAM,QAAQ,GAAkB,EAAE,CAAA;IAClC,IAAI,EAAE,KAAK,EAAE,SAAS,EAAE,WAAW,EAAE,GAAG,KAAK,CAAA;IAC7C,IAAI,UAAU,GAAG,SAAS,CAAA;IAC1B,IAAI,YAAY,GAAG,SAAS,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAA;IAErC,KAAK,IAAI,KAAK,GAAG,CAAC,EAAE,KAAK,GAAG,IAAI,CAAC,MAAM,EAAE,KAAK,EAAE,EAAE,CAAC;QACjD,MAAM,SAAS,GAAG,IAAI,CAAC,KAAK,CAAC,CAAA;QAE7B,IAAI,SAAS,EAAE,CAAC;YACd,IAAI,SAAS,KAAK,GAAG,IAAI,IAAI,CAAC,KAAK,GAAG,CAAC,CAAC,KAAK,GAAG,EAAE,CAAC;gBACjD,SAAS,GAAG,KAAK,CAAA;gBACjB,QAAQ,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,YAAY,EAAE,GAAG,EAAE,KAAK,EAAE,CAAC,CAAA;gBAClD,KAAK,EAAE,CAAA;gBACP,SAAQ;YACV,CAAC;YACD,kEAAkE;YAClE,yEAAyE;YACzE,2DAA2D;YAC3D,SAAQ;QACV,CAAC;QAED,0EAA0E;QAC1E,yEAAyE;QACzE,yEAAyE;QACzE,4BAA4B;QAC5B,IAAI,WAAW,EAAE,CAAC;YAChB,IAAI,SAAS,KAAK,GAAG;gBAAE,WAAW,EAAE,CAAA;iBAC/B,IAAI,SAAS,KAAK,GAAG;gBAAE,WAAW,EAAE,CAAA;YACzC,SAAQ;QACV,CAAC;QAED,IAAI,KAAK,EAAE,CAAC;YACV,IAAI,SAAS,KAAK,KAAK;gBAAE,KAAK,GAAG,EAAE,CAAA;YACnC,SAAQ;QACV,CAAC;QAED,IAAI,SAAS,KAAK,GAAG,IAAI,SAAS,KAAK,GAAG,EAAE,CAAC;YAC3C,KAAK,GAAG,SAAS,CAAA;YACjB,SAAQ;QACV,CAAC;QAED,IAAI,SAAS,KAAK,GAAG,IAAI,IAAI,CAAC,KAAK,GAAG,CAAC,CAAC,KAAK,GAAG,EAAE,CAAC;YACjD,SAAS,GAAG,IAAI,CAAA;YAChB,UAAU,GAAG,IAAI,CAAA;YACjB,YAAY,GAAG,KAAK,CAAA;YACpB,KAAK,EAAE,CAAA;YACP,SAAQ;QACV,CAAC;QAED,SAAS,CAAC,SAAS,GAAG,KAAK,CAAA;QAC3B,IAAI,SAAS,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;YACzB,WAAW,GAAG,CAAC,CAAA;YACf,KAAK,GAAG,SAAS,CAAC,SAAS,GAAG,CAAC,CAAA;YAC/B,SAAQ;QACV,CAAC;QAED,IAAI,SAAS,KAAK,GAAG;YAAE,SAAS,CAAC,IAAI,CAAC,KAAK,CAAC,CAAA;QAC5C,IAAI,CAAC,KAAK,CAAC,GAAG,SAAS,CAAA;IACzB,CAAC;IAED,IAAI,SAAS;QAAE,QAAQ,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,YAAY,EAAE,GAAG,EAAE,IAAI,CAAC,MAAM,EAAE,CAAC,CAAA;IAEvE,OAAO;QACL,SAAS;QACT,IAAI,EAAE,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;QACnB,UAAU;QACV,QAAQ;QACR,QAAQ,EAAE,EAAE,KAAK,EAAE,SAAS,EAAE,WAAW,EAAE;KAC5C,CAAA;AACH,CAAC,CAAA;AAxEY,QAAA,QAAQ,YAwEpB"}