Open eirikbakke opened 2 years ago
Also note that the Javadoc and parameter name for CharInputReader.enableNormalizeLineEndings(escaping) seems to reverse the actual behavior of the method as assumed by callers and implemented in AbstractCharInputReader. In fact, in the latter overridden method, the parameter has been renamed to normalizeLineEndings, which seems like a more correct name.
The following CSV file, with "\r" style line endings...
...should parse as [[colA, colB, colC], [a, A, x], [b, B, k]]. However, when lineSeparatorDetectionEnabled=true and normalizeLineEndingsWithinQuotes=false, I instead get [[colA, colB, colC], [a, A, "x"\rb, B, k]].
Here is a complete test case, which fails with Univocity 2.9.1 on Windows 11 and Java 17:
Thank you for your work on the excellent Univocity library! I am using it for Ultorg and am in the process of writing unit tests, which is how I found the bug above...