Mailing-List: contact dev-help@cocoon.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@cocoon.apache.org
User-Agent: Microsoft-Entourage/10.1.1.2418
Date: Mon, 11 Aug 2003 15:06:59 +0100
Subject: Re: AbstractSerializerTestCase
From: Mark Leicester <mark.leicester@energyintellect.com>
To: <dev@cocoon.apache.org>
Message-ID: <BB5D6193.14F11%mark.leicester@energyintellect.com>
In-Reply-To: <Pine.LNX.4.44.0308111022150.14840-100000@vern.chem.tu-berlin.de>
Mime-version: 1.0
Content-type: text/plain; charset="US-ASCII"
Content-transfer-encoding: 7bit

On 11/08/2003 9:29, "Stephan Michels" <stephan@apache.org> wrote:

> On Sat, 9 Aug 2003, Mark Leicester wrote:
> 
>> Do you have plans for serializer test cases, and if so does this code fit
>> the bill? I thought I should post it to you and the list for comment before
>> I submit it as a patch.
 
> Code is always welcome. I havn't written a Serializer testcase until now,
> because I find it difficult to test binary data. In my POV you must
> generate the expected output with a different tool than the component you
> want to test. In the case of binary data, it is not that easy.
Quite right. The chickens and the eggs.

In my case (with the XMidiGenerator and XMidiSerializer) I am trying to
confirm that a "round trip" has been successful, so for me this test is
ideal. After generating XML *from* my binary file I subsequently serialize
to a binary file which I hope is exactly the same as the original binary
file.

Also, I find that once I have finally succeeded in getting some code to work
it's good to have a test to confirm that if I make changes to it later, my
code is *still* working! So indeed, even if the expected output may have
been generated with the component I want to test, at least I can tell if the
component continues to behave consistently after I make changes.

> But anyway, I applied that patch.
Thanks!

Regards,
Mark