Update MD037/no-space-in-emphasis to ignore emphasis markers in code spans (fixes #278).

This commit is contained in:
David Anson 2020-04-25 15:10:07 -07:00
parent bdc0246b34
commit f5a71521d4
5 changed files with 195 additions and 76 deletions

View file

@ -40,6 +40,7 @@
"multiline-comment-style": ["error", "separate-lines"], "multiline-comment-style": ["error", "separate-lines"],
"multiline-ternary": "off", "multiline-ternary": "off",
"newline-per-chained-call": "off", "newline-per-chained-call": "off",
"no-continue": "off",
"no-empty-function": "off", "no-empty-function": "off",
"no-extra-parens": "off", "no-extra-parens": "off",
"no-implicit-coercion": "off", "no-implicit-coercion": "off",

View file

@ -25,6 +25,9 @@ module.exports.bareUrlRe = /(?:http|ftp)s?:\/\/[^\s\]"']*/ig;
module.exports.listItemMarkerRe = /^([\s>]*)(?:[*+-]|\d+[.)])\s+/; module.exports.listItemMarkerRe = /^([\s>]*)(?:[*+-]|\d+[.)])\s+/;
module.exports.orderedListItemMarkerRe = /^[\s>]*0*(\d+)[.)]/; module.exports.orderedListItemMarkerRe = /^[\s>]*0*(\d+)[.)]/;
// Regular expression for emphasis markers
const emphasisMarkersRe = /[_*]+/g;
// readFile options for reading with the UTF-8 encoding // readFile options for reading with the UTF-8 encoding
module.exports.utf8Encoding = { "encoding": "utf8" }; module.exports.utf8Encoding = { "encoding": "utf8" };
@ -330,84 +333,90 @@ module.exports.forEachHeading = function forEachHeading(params, handler) {
}); });
}; };
// Calls the provided function for each inline code span's content /**
module.exports.forEachInlineCodeSpan = * Calls the provided function for each inline code span's content.
function forEachInlineCodeSpan(input, handler) { *
let currentLine = 0; * @param {string} input Markdown content.
let currentColumn = 0; * @param {Function} handler Callback function.
let index = 0; * @returns {void}
while (index < input.length) { */
let startIndex = -1; function forEachInlineCodeSpan(input, handler) {
let startLine = -1; let currentLine = 0;
let startColumn = -1; let currentColumn = 0;
let tickCount = 0; let index = 0;
let currentTicks = 0; while (index < input.length) {
let state = "normal"; let startIndex = -1;
// Deliberate <= so trailing 0 completes the last span (ex: "text `code`") let startLine = -1;
for (; index <= input.length; index++) { let startColumn = -1;
const char = input[index]; let tickCount = 0;
// Ignore backticks in link destination let currentTicks = 0;
if ((char === "[") && (state === "normal")) { let state = "normal";
state = "linkTextOpen"; // Deliberate <= so trailing 0 completes the last span (ex: "text `code`")
} else if ((char === "]") && (state === "linkTextOpen")) { for (; index <= input.length; index++) {
state = "linkTextClosed"; const char = input[index];
} else if ((char === "(") && (state === "linkTextClosed")) { // Ignore backticks in link destination
state = "linkDestinationOpen"; if ((char === "[") && (state === "normal")) {
} else if ( state = "linkTextOpen";
((char === "(") && (state === "linkDestinationOpen")) || } else if ((char === "]") && (state === "linkTextOpen")) {
((char === ")") && (state === "linkDestinationOpen")) || state = "linkTextClosed";
(state === "linkTextClosed")) { } else if ((char === "(") && (state === "linkTextClosed")) {
state = "normal"; state = "linkDestinationOpen";
} } else if (
// Parse backtick open/close ((char === "(") && (state === "linkDestinationOpen")) ||
if ((char === "`") && (state !== "linkDestinationOpen")) { ((char === ")") && (state === "linkDestinationOpen")) ||
// Count backticks at start or end of code span (state === "linkTextClosed")) {
currentTicks++; state = "normal";
if ((startIndex === -1) || (startColumn === -1)) {
startIndex = index + 1;
}
} else {
if ((startIndex >= 0) &&
(startColumn >= 0) &&
(tickCount === currentTicks)) {
// Found end backticks; invoke callback for code span
handler(
input.substring(startIndex, index - currentTicks),
startLine, startColumn, tickCount);
startIndex = -1;
startColumn = -1;
} else if ((startIndex >= 0) && (startColumn === -1)) {
// Found start backticks
tickCount = currentTicks;
startLine = currentLine;
startColumn = currentColumn;
}
// Not in backticks
currentTicks = 0;
}
if (char === "\n") {
// On next line
currentLine++;
currentColumn = 0;
} else if ((char === "\\") &&
((startIndex === -1) || (startColumn === -1)) &&
(input[index + 1] !== "\n")) {
// Escape character outside code, skip next
index++;
currentColumn += 2;
} else {
// On next column
currentColumn++;
}
} }
if (startIndex >= 0) { // Parse backtick open/close
// Restart loop after unmatched start backticks (ex: "`text``code``") if ((char === "`") && (state !== "linkDestinationOpen")) {
index = startIndex; // Count backticks at start or end of code span
currentLine = startLine; currentTicks++;
currentColumn = startColumn; if ((startIndex === -1) || (startColumn === -1)) {
startIndex = index + 1;
}
} else {
if ((startIndex >= 0) &&
(startColumn >= 0) &&
(tickCount === currentTicks)) {
// Found end backticks; invoke callback for code span
handler(
input.substring(startIndex, index - currentTicks),
startLine, startColumn, tickCount);
startIndex = -1;
startColumn = -1;
} else if ((startIndex >= 0) && (startColumn === -1)) {
// Found start backticks
tickCount = currentTicks;
startLine = currentLine;
startColumn = currentColumn;
}
// Not in backticks
currentTicks = 0;
}
if (char === "\n") {
// On next line
currentLine++;
currentColumn = 0;
} else if ((char === "\\") &&
((startIndex === -1) || (startColumn === -1)) &&
(input[index + 1] !== "\n")) {
// Escape character outside code, skip next
index++;
currentColumn += 2;
} else {
// On next column
currentColumn++;
} }
} }
}; if (startIndex >= 0) {
// Restart loop after unmatched start backticks (ex: "`text``code``")
index = startIndex;
currentLine = startLine;
currentColumn = startColumn;
}
}
}
module.exports.forEachInlineCodeSpan = forEachInlineCodeSpan;
/** /**
* Adds a generic error object via the onError callback. * Adds a generic error object via the onError callback.
@ -484,6 +493,41 @@ module.exports.frontMatterHasTitle =
frontMatterLines.some((line) => frontMatterTitleRe.test(line)); frontMatterLines.some((line) => frontMatterTitleRe.test(line));
}; };
/**
* Returns a list of emphasis markers in code spans.
*
* @param {Object} params RuleParams instance.
* @returns {number[][]} List of markers.
*/
function emphasisMarkersInCodeSpans(params) {
const { lines } = params;
const byLine = new Array(lines.length);
filterTokens(params, "inline", (token) => {
const { children, lineNumber, map } = token;
if (children.some((child) => child.type === "code_inline")) {
const tokenLines = lines.slice(map[0], map[1]);
forEachInlineCodeSpan(
tokenLines.join("\n"),
(code, lineIndex, column, tickCount) => {
const codeLines = code.split(newLineRe);
codeLines.forEach((codeLine, codeLineIndex) => {
let match = null;
while ((match = emphasisMarkersRe.exec(codeLine))) {
const byLineIndex = lineNumber - 1 + lineIndex + codeLineIndex;
const inLine = byLine[byLineIndex] || [];
const codeLineOffset = codeLineIndex ? 0 : column - 1 + tickCount;
inLine.push(codeLineOffset + match.index);
byLine[byLineIndex] = inLine;
}
});
}
);
}
});
return byLine;
}
module.exports.emphasisMarkersInCodeSpans = emphasisMarkersInCodeSpans;
/** /**
* Gets the most common line ending, falling back to the platform default. * Gets the most common line ending, falling back to the platform default.
* *

View file

@ -2,7 +2,8 @@
"use strict"; "use strict";
const { addErrorContext, forEachLine, isBlankLine } = require("../helpers"); const { addErrorContext, emphasisMarkersInCodeSpans, forEachLine,
includesSorted, isBlankLine } = require("../helpers");
const { lineMetadata } = require("./cache"); const { lineMetadata } = require("./cache");
const emphasisRe = /(^|[^\\])(?:(\*\*?\*?)|(__?_?))/g; const emphasisRe = /(^|[^\\])(?:(\*\*?\*?)|(__?_?))/g;
@ -63,6 +64,7 @@ module.exports = {
return null; return null;
} }
// Initialize // Initialize
const ignoreMarkersByLine = emphasisMarkersInCodeSpans(params);
resetRunTracking(); resetRunTracking();
forEachLine( forEachLine(
lineMetadata(), lineMetadata(),
@ -83,7 +85,12 @@ module.exports = {
let match = null; let match = null;
// Match all emphasis-looking runs in the line... // Match all emphasis-looking runs in the line...
while ((match = emphasisRe.exec(line))) { while ((match = emphasisRe.exec(line))) {
const ignoreMarkersForLine = ignoreMarkersByLine[lineIndex] || [];
const matchIndex = match.index + match[1].length; const matchIndex = match.index + match[1].length;
if (includesSorted(ignoreMarkersForLine, matchIndex)) {
// Ignore emphasis markers inside code spans
continue;
}
const matchLength = match[0].length - match[1].length; const matchLength = match[0].length - match[1].length;
if (emphasisIndex === -1) { if (emphasisIndex === -1) {
// New run // New run

View file

@ -73,3 +73,50 @@ emphasis _ text {MD037}
Text ** bold {MD037} Text ** bold {MD037}
bold ** text {MD037} bold ** text {MD037}
Emphasis `inside
of * code *
blocks` is okay.
Emphasis `* inside`
code
`blocks *` is okay.
Emphasis `inside *`
code
`* blocks` is okay.
Emphasis `inside
_ code _
blocks` is okay.
Emphasis `_ inside`
code
`blocks _` is okay.
Emphasis `inside _`
code
`_ blocks` is okay.
Mixed `code_span`
scenarios
are _also_ okay.
Mixed `code*span`
scenarios
are *also* okay.
This paragraph
contains *a* mix
of `*` emphasis
scenarios and *should*
not trigger `*` any
violations at *all*.
This paragraph
contains `a * slightly
more complicated
multi-line emphasis
scenario * that
should * not trigger
violations * either`.

View file

@ -154,3 +154,23 @@ Text *emph***strong ** text {MD037}
```markdown ```markdown
Violations * are * allowed in code blocks where emphasis does not apply. Violations * are * allowed in code blocks where emphasis does not apply.
``` ```
Emphasis `inside * code * blocks` is okay.
Emphasis `* inside` code `blocks *` is okay.
Emphasis `inside *` code `* blocks` is okay.
Emphasis `inside _ code _ blocks` is okay.
Emphasis `_ inside` code `blocks _` is okay.
Emphasis `inside _` code `_ blocks` is okay.
Mixed `code_span` scenarios are _also_ okay.
Mixed `code*span` scenarios are *also* okay.
Mixed `code*span` scenarios are _also_ okay.
Mixed `code_span` scenarios are *also* okay.