\r in toString messes up tabled scenario

[nikowitt]

Hi Jan,

    @Test
    @DataProvider(splitBy = __, value = {
            "Test" + __ + "<p>Test</p>",
            "<p>Test</p>" + __ + "<p>Test</p>",
            "<p>Test 1\nTest 2</p>" + __ + "<p>Test 1<br>Test 2</p>",
            "<p>Test 1\nTest 2\r\nTest 3</p>" + __ + "<p>Test 1<br>Test 2\r<br>Test 3</p>",
    })
    @CaseDescription("convertNonHtmlStringToTinyMceString(\"$0\")")
    public void convert_non_html_string_to_tiny_mce_html(String input, String output) throws Exception {
        assert_that().assert_expression(JAFStringUtils.convertNonHtmlStringToTinyMceString(input), is(output));
    }

results in:

Test Class: com.sobis.jaf.tests.tools.JAFStringUtilsTest

 Scenario: convert non html string to tiny mce html

   Assert that <output> <matcher>

  Cases:

    | # | Description                                        | output                             | matcher                                  | Status  |
    +---+----------------------------------------------------+------------------------------------+------------------------------------------+---------+
    | 1 | convertNonHtmlStringToTinyMceString("Test")        | <p>Test</p>                        | is "<p>Test</p>"                         | Success |
    | 2 | convertNonHtmlStringToTinyMceString("<p>Test</p>") | <p>Test</p>                        | is "<p>Test</p>"                         | Success |
    | 3 | convertNonHtmlStringToTinyMceString("<p>Test 1     | <p>Test 1<br>Test 2</p>            | is "<p>Test 1<br>Test 2</p>"             | Success |
    |   | Test 2</p>")                                       |                                    |                                          |         |
    | 4 | convertNonHtmlStringToTinyMceString("<p>Test 1     | <p>Test 1<br>Test 2
<br>Test 3</p> | is "<p>Test 1<br>Test 2\r<br>Test 3</p>" | Success |
    |   | Test 2
                                            |                                    |                                          |         |
    |   | Test 3</p>")                                       |                                    |                                          |         |

For \n, it is already fixed, but Windows users frequently encouter \r where the same should be applied.

[janschaefer]

I agree :-)

[nikowitt]

Just tried again with the snapshot:

    | # | Description                                        | output                             | matcher                                  | Status  |
    +---+----------------------------------------------------+------------------------------------+------------------------------------------+---------+
    | 1 | convertNonHtmlStringToTinyMceString("Test")        | <p>Test</p>                        | is "<p>Test</p>"                         | Success |
    | 2 | convertNonHtmlStringToTinyMceString("<p>Test</p>") | <p>Test</p>                        | is "<p>Test</p>"                         | Success |
    | 3 | convertNonHtmlStringToTinyMceString("<p>Test 1     | <p>Test 1<br>Test 2</p>            | is "<p>Test 1<br>Test 2</p>"             | Success |
    |   | Test 2</p>")                                       |                                    |                                          |         |
    | 4 | convertNonHtmlStringToTinyMceString("<p>Test 1     | <p>Test 1<br>Test 2
<br>Test 3</p> | is "<p>Test 1<br>Test 2\r<br>Test 3</p>" | Success |
    |   | Test 2                                             |                                    |                                          |         |
    |   | Test 3</p>")                                       |                                    |                                          |         |

Better, but not still not perfect :)

[janschaefer]

Yes. I have not handled the case of a single carriage return. Do you think that this might actually happen in practice?

[nikowitt]

At least in the current state of the conversion method this might happen - of course I could also adapt this in the implementation to convert both \n and \r\n to breaks which is probably better.