On 2006-10-10, at 15:40 EDT, Henry Minsky wrote:

>> > Tests: 688 Failures: 6 Errors: 6
>> > TestFailure: Object >>> testSerialize failed: Equals:  expected
>> > «string(70)#24| "<bug2263><record a=\"fl&apos;ll\"/><record
>> > b=\"dkd&quot;kddk\"/></bug2263>"» got «string(168)#25|
>> > "<bug2263><record a=\"fl&apos;ll\"
>> > xmlns=\"http://www.laszlosystems.com/2003/05/lzx\"/><record
>> > b=\"dkd&quot;kddk\"
>> > xmlns=\"http://www.laszlosystems.com/2003/05/lzx\"/></bug2263>"»
>>
>> This is a bogus test.  Serialization does not guarantee to visit the
>> nodes in any particular order.
>>
>
> Serialization should produce arbitrary order for attributes, but the
> nodes themselves need
> to occur in a fixed order, for XML semantics.
>
> I actually implemented a "equal" (in the lisp sense) for the data
> tests, in test/lfc/data/xmlequals.lzx, I should put it  someplace
> common to include into the lztest and lzunit libraries as well.

That's what I meant.

I'm ok with you adding an xmlequality predicate, just let's not go  
down the `equalp` path...

_______________________________________________
Laszlo-dev mailing list
[email protected]
http://www.openlaszlo.org/mailman/listinfo/laszlo-dev

Reply via email to