W3C home > Mailing lists > Public > public-xsd-databinding@w3.org > March 2008

Re: Outputting default values (Was: Re: Data binding results update for XBinder)

From: Ed Day <edday2006@gmail.com>
Date: Wed, 12 Mar 2008 08:58:58 -0400
Message-ID: <c92fe5920803120558y70d92d43vc6f45b8cea9607b2@mail.gmail.com>
To: "Pete Cordell" <petexmldev@tech-know-ware.com>
Cc: public-xsd-databinding@w3.org

Hi Pete,

On Wed, Mar 12, 2008 at 7:02 AM, Pete Cordell
<petexmldev@tech-know-ware.com> wrote:
> I see from Ed's results that they 'failed' on LocalElementDefault because
>  they output the default value when it wasn't present on the input.  I'm just
>  wondering whether this should really be considered a failure?  If anything
>  it proves that they captured the default value rather than capturing
>  nothing.
>
>  Also, I see that that stupid float number is still in there.  I think Ed
>  must have done something to get this to round trip as I can't see how C/C++
>  generates a float of the form 1267.43233E12 (I may be wrong).

The way we got it to round-trip was to introduce a configuration
option that allowed the use of a string variable for these numbers.
That is all you can do.  EXI has a similar issue with round-tripping
from binary back to text format in the context of XML digital
signatures.  To solve it, the format has a "preserve.lexical" option
that specifies the use of strings for numbers in these situations.

> In my opinion
>  the choice of this number encourages you to make such exceptions.  For
>  example, when we do the tests we will be mapping xs:float to C's 'double'
>  rather than the default 'float' to address this.  Maybe there should be a
>  section of the report that captures any special setup that has been used for
>  the tests.
>
>  Regards,
>
>  Pete Cordell
>  Codalogic
>  For XML C++ data binding visit http://www.codalogic.com/lmx/
>
>  ----- Original Message -----
>  From: "Ed Day" <edday2006@gmail.com>
>  To: <public-xsd-databinding@w3.org>
>  Sent: Tuesday, March 11, 2008 5:23 PM
>  Subject: Data binding results update for XBinder
>
>
>  >
>  > Hello everyone,
>  >
>  > We would like to update our original submission of XBinder results
>  > that were made last summer.  We have had two releases since then and
>  > now score considerably better on the tests.
>  >
>  > Our latest test report can be found at the following location:
>  >
>  > http://www.obj-sys.com/w3cdatabinding/report_xbinder_c_1.4.html
>  >
>  > The full test suite implementation is available here:
>  >
>  > http://www.obj-sys.com/w3cdatabinding/xbinder14_w3cdb.zip
>  >
>  > Please let us know if anything is missing that is necessary for
>  > submission.  We look forward to your prompt action on this request and
>  > would like to thank the group for all the fine work you have done in
>  > bringing data binding issues to light.
>  >
>  > Regards,
>  > --
>  > Ed Day
>  > Objective Systems, Inc.
>  > http://www.obj-sys.com
>  >
>  >
>  >
>
>
>



-- 
Ed Day
Objective Systems, Inc.
http://www.obj-sys.com
Received on Wednesday, 12 March 2008 12:59:11 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Saturday, 18 December 2010 18:20:38 GMT