fix: do not match too many characters when removing sourceMappingURL … #286
Conversation
…comment a sourceMappingURL comment may be created by another loader and this removes it again, but too many characters are removed, thus uncommenting a different line that contains invalid javascript Closes #285
|
|
Codecov Report
@@ Coverage Diff @@
## master #286 +/- ##
=======================================
Coverage 74.48% 74.48%
=======================================
Files 6 6
Lines 145 145
Branches 52 52
=======================================
Hits 108 108
Misses 31 31
Partials 6 6
Continue to review full report at Codecov.
|
src/utils.js
Outdated
| @@ -87,7 +87,7 @@ ${ | |||
| } | |||
|
|
|||
| // Matches only the last occurrence of sourceMappingURL | |||
| const innerRegex = /\s*[#@]\s*sourceMappingURL\s*=\s*([^\s'"]*)\s*/; | |||
| const innerRegex = /\s*[#@]\s*sourceMappingURL\s*=\s*((?:[^\s'"\\]|\\[^n])*(?:\\n)?)\s*/; | |||
There was a problem hiding this comment.
^n? Looks it is invalid
There was a problem hiding this comment.
The idea was to match everything except all whitespace, quotes and \n literally (not the line feed character), but not a backslash on its own. It works like a negative lookahead but consuming characters. Another possible solution would be .*?(?=[\s'"]|\\n)(?:\\n)?, which actually looks a lot more readable (instead of (?:[^\s'"\\]|\\[^n])*(?:\\n)?)
And I've just noticed that with the current regex the match wouldn't stop when it encounters a whitespace or quote after a backslash, but it would have done before. Also the test will fail, if it is in a multiline comment. I will work on a fix. Thanks for not letting this slip past!
…lashes when it shouldn't The regex is also much more readable
Fix failing test when `sourceURL=webpack-internal:///...` is inside a multiline comment
|
I don't know why the lint job fails. I've tested it at the |
|
@schl3ck All fine, don't worry, I will look at this in near future |
…comment
a sourceMappingURL comment may be created by another loader and this removes it again, but too many
characters are removed, thus uncommenting a different line that contains invalid javascript
Closes #285
This PR contains a:
Motivation / Use-Case
See #285