Menu
▾
▴
RE: [Junit-devel] Patch suggestion for ComparsionFailure
|
From: Kent B. <ke...@th...> - 2002-11-19 17:17:11
|
You're absolutely right that the current solution isn't it. Fixed-sized
(even preference-based) regions offend my aesthetics. If we can't do
better than that, okay, but we should at least try.
Kent
> -----Original Message-----
> From: jun...@li...
> [mailto:jun...@li...] On Behalf Of
> Erik Mattsson
> Sent: Monday, November 18, 2002 1:55 PM
> To: jun...@li...
> Subject: [Junit-devel] Patch suggestion for ComparsionFailure
>
>
>
> Hi
>
> The new ComparsionFailuer class to handle string comparsions
> is great but I thought it could be improved a little bit. Im
> thinking about the "..." handling, it is annoying that the
> comparsion is only displaying the difference between the
> expected and the actual string. This makes it hard to check
> why some assertEquals fails such as the examples below.
> My patch displays some common strings before and after the
> difference,
> either by going back/forward 15 chars or to the nearest
> space. This would
> make it easier to spot the difference and perhaps find out
> why the testcase
> failed before looking in the code.
> It is not a big improvement but it might avoid some confusion
> when comparing strings or objects (by toString()). Any
> feedback or suggestions??
>
> //
> // EXAMPLES
> // assertion
> assertEquals("abcdef", "abdef" )
> // old result, which shows no help at all
> expected:<...c...> but was:<......>
> // with patch it could be:
> expected:<abcdef> but was:<bcdef>
>
>
> //assertion
> assertEquals( "b=16, c=15, d=15, e=5", "b=16, c=16, d=15,
> e=5" ); // old output - which 5 is wrong? expected:<...5...>
> but was:<...6...> // with patch expected:<... c=15, ...> but
> was:<... c=16, ...>
>
>
> //assertion
> assertEquals(
> "IPTC=[caption='caption';contact='contact';]",
> "IPTC=[caption='caption';city='gbg';contact='contact';]"
> ); //old output expected:<......> but was:<...ity='gbg';c...>
> //with patch expected:<...ion='caption';contact='contact...> but was:
> <...ion='caption';city='gbg';contact='contact...>
>
>
> // stupid short example to show the problem with comparing
> // floats in a toString() method.
> assertEquals( Float.toString( 0.124235f), Float.toString(
> 0.124435f ) ); // old result - which one of the 2's are
> wrong? expected:<...2...> but was:<...4...> // with patch:
> expected:<0.124235> but was:<0.124435>
>
>
> Please look through my patch suggestion and give me some
> feedback on it...
>
>
> Here is my patch suggestion: (dunno if I got the diff correct) 9d8
> < public final static int COMMON_FIX_LENGTH = 15;
> 55,79d53
> < int pos = fExpected.lastIndexOf( ' ', i );
> < if ( ( pos == -1 ) || ( ( i - pos ) >
> COMMON_FIX_LENGTH ) ) {
> < i -= COMMON_FIX_LENGTH;
> < } else {
> < i = pos;
> < }
> <
> < pos = fExpected.indexOf( ' ', j );
> < if ( ( pos == -1 ) || ( ( pos - j ) >
> COMMON_FIX_LENGTH ) ) {
> < j += COMMON_FIX_LENGTH;
> < } else {
> < j = pos;
> < }
> <
> < pos = fActual.indexOf( ' ', k );
> < if ( ( pos == -1 ) || ( ( pos - k ) >
> COMMON_FIX_LENGTH ) ) {
> < k += COMMON_FIX_LENGTH;
> < } else {
> < k = pos;
> < }
> <
> < i = ( i < 0 ? 0 : i );
> < j = ( j >= fExpected.length() ?
> fExpected.length() - 1 : j );
> < k = ( k >= fActual.length() ?
> fActual.length()
> - 1 : k );
> <
>
>
> Regards Erik Mattsson
>
>
>
>
> -------------------------------------------------------
> This sf.net email is sponsored by: To learn the basics of securing
> your web site with SSL, click here to get a FREE TRIAL of a Thawte
> Server Certificate: http://www.gothawte.com/rd524.html
> _______________________________________________
> Junit-devel mailing list
> Jun...@li...
> https://lists.sourceforge.net/lists/listinfo/j> unit-devel
>
|