@rdiachenko
The cause of CI failure is:
How the format method works here:

sourceFile = sourceFile.withPrefix(
                            Space.format(fixedHeader + System.lineSeparator()));

format method treats \r and \n as separate values, therefore it fails in windows where System.lineSeprator() return \r\n and hence an extra space is added in each comment.

Will raise a separate PR for fix,
but i am confused how it worked earlier.
please share your insight.

i tried replicating the issue with AbstractRecipeTest where we manually provided the config but it worked well. i compared the header being passed to the main method it was exactly same.

@Anmol202005

Let's step back. Why do we need + System.lineSeparator() at all? I think we should rely on what is in the header file and just concatinate it.

Additionally, could you resolve conflicts please

Let's step back. Why do we need + System.lineSeparator() at all? I think we should rely on what is in the header file and just concatinate it.

@rdiachenko its necessary so that the code after the header starts from its own line.

Although IMO in the existing code there are two issues(bugs) that balances out.

• use of System.lineseprator()
• use of Files.readString() , to read input and output file.

AbstractPathTestSupport uses the following method to read files:

protected static String readFile(String filename) throws IOException {
        return toLfLineEnding(Files.readString(Path.of(filename)));
    }
    
protected static String toLfLineEnding(String text) {
        return text.replaceAll("(?x)\\\\r(?=\\\\n)|\\r(?=\\n)", "");
    }

to convert Windows-style line endings (\r\n) to Unix-style line endings (\n).

https://github.com/checkstyle/checkstyle/blob/master/src/test/java/com/puppycrawl/tools/checkstyle/AbstractPathTestSupport.java#L121

@rdiachenko
Since the readFile() bug is already resolved in this PR, I prefer to address the header bug here as well, instead of making changes to AbstractRecipeTest in a separate PR. Moreover, since AbstractRecipeTest will be removed in this PR, it’s cleaner to handle all related changes together.

@rdiachenko Done.