RE: [Junit-devel] assertEquals: String compare with context diff

[Erik M. <em...@ge...>]

Alex,

Good stuff!  When on the road I get asked for just this bit of
functionality a lot.  You can go ahead and use SF to submit your
patch (http://sourceforge.net/tracker/?group_id=15278&atid=315278)
You'll notice a lot of patches there.  You greatly increase your
odds of getting your patch in by supplying tests and making your
code look like the JUnit code.

int at= strdiffat(actual, expected);

vs.

int at = strdiffat(actual, expected);

I sort of like the assertEqualsDiff, but that is probably due to
my fear that we would break someone's tests.  I would feel a lot
better about modifying assertEquals if Kent or Erich blessed it.
The purpletech license looks okay, and the Util class has a JUnit
TestCase :)  Not sure how we should handle the external
dependency... I see two options, put it in the JUnit CVS or use
an Ant task to download it (I like the second).

Erik Meade
http://junit.org
em...@ob...

> -----Original Message-----
> From: jun...@li...
> [mailto:jun...@li...]On Behalf Of
> gu...@st...
> Sent: Monday, February 18, 2002 10:56 AM
> To: jun...@li...
> Subject: [Junit-devel] assertEquals: String compare with context diff
>
>
> The method assertEquals(String, String) is very useful, but breaks
> down with large strings.  The problem is that, if the strings differ
> in only a few characters, or start to differ more than a few lines in,
> the error message is almost completely useless.
>
> expected:<HUGE STRING> but was:<NEARLY IDENTICAL HUGE STRING>
>
> When you add in line breaks it gets impossible to "eyeball" the error
> message and know what's going wrong.  It also breaks up the format of
> the output report with miles of strings, possibly including line
> breaks.
>
> I've written some code that fixes this.  The output for a failed test
> with message "HTML Report" is:
>
> HTML Report: strings differ at character 22
> Expected: ...>Poll:</b> dummy<br>\n<b>Question:</b> Vote for your favor...
>   Actual: ...>Poll:</b> dummy<br>\r\n<b>Question:</b> Vote for
> your favo...
>
> Note that you can tell right off the bat that the problem is a
> difference in line endings (since my code escapes those).  This
> particular bug was actually impossible for me to track down using the
> normal failure message, since line endings are invisible.
>
> I'd like to contribute the code to JUnit.  I'm just not sure where you
> want to put it.
>
> I see three options:
>
> 1. Change the behavior of the existing Assert.assertEquals(String, String)
>
> 	(I don't think this would break anybody's tests, and I doubt
> 	anyone has written code that depends on the actual format of
> 	the string error messages.  I could be wrong, of course.)
>
> 2. Add a new assertEqualsDiff to class Assert
>
> 	(This would mean that the benefit of the new messages would
> 	not be given to existing tests without massive test code
> 	rewrites.)
>
> 3. Make a new DiffTestCase that overrides assertEquals
>
> 	(Same problem as above -- existing test code would need to be
> 	rewritten.  It's also a sloppy substitute for mixins...)
>
> If you guys want the code, let me know, and I'll download the JUnit
> source and give you a patch.
>
>  - A
>
> P.S.  I've attached the current version below.  I don't think this is
> the final version, though -- I want it to be smarter about diffs close
> to the start and end of the strings.
>
> --
> Alex Chaffee                       mailto:al...@jg...
> jGuru - Java News and FAQs         http://www.jguru.com/alex/
> Creator of Gamelan                 http://www.gamelan.com/
> Founder of Purple Technology       http://www.purpletech.com/
> Curator of Stinky Art Collective   http://www.stinky.com/
>
>
>     void assertEqualsDiff(String message, String expected, String actual)
>     {
> 	if (!expected.equals(actual)) {
> 	    if (expected.length() < 20 || actual.length() < 20) {
> 		assertEquals(message, expected, actual);
> 		return;	// we'll get an exception, but just in case
> 	    }
> 	    int at = strdiffat(actual, expected);
> 	    String expectedContext =
> 		Utils.abbreviate(expected.substring(at - 20), 60);
> 	    String actualContext =
> 		Utils.abbreviate(actual.substring(at - 20), 60);
> 	    assertTrue(message + ": strings differ at character " +
> at + "\n" +
> 		       "Expected: ..." +
> Utils.jsEscape(expectedContext) + "\n" +
> 		       "  Actual: ..." +
> Utils.jsEscape(actualContext) + "\n",
> 		       false);
> 	}
>     }
>
>     /**
>      * returns the index where the strings diverge<p>
>      * strdiff("i am a machine", "i am a robot") -> 7<p>
>      * @return -1 if they are the same
>      *
>      **/
>     public static int strdiffat(String s1, String s2) {
> 	int i;
> 	for (i=0; i<s1.length() && i<s2.length(); ++i) {
> 	    if (s1.charAt(i) != s2.charAt(i)) {
> 		break;
> 	    }
> 	}
> 	if (i<s2.length() || i<s1.length())
> 	    return i;
> 	return -1;
>     }
>
> for abbreviate() and jsEscape() see
http://www.purpletech.com/code/src/com/purpletech/util/Utils.java

_______________________________________________
Junit-devel mailing list
Jun...@li...
https://lists.sourceforge.net/lists/listinfo/junit-devel