Merge pull request #588 from Tencent/bugfix/emphasis-performance-issue

fix(emphasis): fix poor performance when matching across lines
Tencent · Sep 14, 2023 · fc01d76 · fc01d76
2 parents e574f80 + 6246d87
commit fc01d76
Show file tree

Hide file tree

Showing 2 changed files with 14 additions and 12 deletions.
diff --git a/src/core/hooks/Emphasis.js b/src/core/hooks/Emphasis.js
@@ -14,12 +14,7 @@
  * limitations under the License.
  */
 import SyntaxBase from '@/core/SyntaxBase';
-import {
-  compileRegExp,
-  DO_NOT_STARTS_AND_END_WITH_SPACES_MULTILINE_ALLOW_EMPTY,
-  ALLOW_WHITESPACE_MULTILINE,
-  UNDERSCORE_EMPHASIS_BOUNDARY,
-} from '@/utils/regexp';
+import { compileRegExp, ALLOW_WHITESPACE_MULTILINE, UNDERSCORE_EMPHASIS_BOUNDARY } from '@/utils/regexp';
 
 export default class Emphasis extends SyntaxBase {
   static HOOK_NAME = 'fontEmphasis';
@@ -94,19 +89,20 @@ export default class Emphasis extends SyntaxBase {
    */
   rule({ config } = { config: undefined }) {
     const allowWhitespace = config ? !!config.allowWhitespace : false;
-    const REGEX = allowWhitespace
-      ? ALLOW_WHITESPACE_MULTILINE
-      : DO_NOT_STARTS_AND_END_WITH_SPACES_MULTILINE_ALLOW_EMPTY;
+    const emRegexp = (allowWhitespace, symbol) => {
+      const char = `[^${symbol}\\s]`;
+      return allowWhitespace ? ALLOW_WHITESPACE_MULTILINE : `(${char}|${char}(.*?(\n${char}.*)*)${char})`;
+    };
     const asterisk = {
-      begin: '(^|[^\\\\])(\\*+)', // ?<leading>, ?<symbol>
-      content: `(${REGEX})`, // ?<text>
+      begin: '(^|[^\\\\])([*]+)', // ?<leading>, ?<symbol>
+      content: `(${emRegexp(allowWhitespace, '*')})`, // ?<text>
       end: '\\2',
     };
 
     // UNDERSCORE_EMPHASIS_BORDER：允许除下划线以外的「标点符号」和空格出现，使用[^\w\S \t]或[\W\s]会有性能问题
     const underscore = {
       begin: `(^|${UNDERSCORE_EMPHASIS_BOUNDARY})(_+)`, // ?<leading>, ?<symbol>
-      content: `(${REGEX})`, // ?<text>
+      content: `(${emRegexp(allowWhitespace, '_')})`, // ?<text>
       end: `\\2(?=${UNDERSCORE_EMPHASIS_BOUNDARY}|$)`,
     };
 

diff --git a/src/utils/regexp.js b/src/utils/regexp.js
@@ -37,6 +37,12 @@ export const ALLOW_WHITESPACE_MULTILINE = '(?:.*?)(?:(?:\\n.*?)*?)';
 export const DO_NOT_STARTS_AND_END_WITH_SPACES = '(?:\\S|(?:\\S.*?\\S))';
 export const DO_NOT_STARTS_AND_END_WITH_SPACES_MULTILINE =
   '(?:(?:\\S|(?:\\S[^\\n]*?\\S))(?:\\n(?:\\S|(?:\\S[^\\n]*?\\S)))*?)';
+
+/**
+ * @deprecated
+ *
+ * 存在严重性能问题，应弃用
+ */
 export const DO_NOT_STARTS_AND_END_WITH_SPACES_MULTILINE_ALLOW_EMPTY = '(?:(?:\\S|(?:\\S.*?\\S))(?:[ \\t]*\\n.*?)*?)';
 
 export const NOT_ALL_WHITE_SPACES_INLINE = '(?:[^\\n]*?\\S[^\\n]*?)';