Re: Rejecting invalid/unsupported p:serialization

"Toman_Vojtech@emc.com" <Toman_Vojtech@emc.com> writes:
> I have a question about reporting err:C0001 (unsupported serialization
> method) and err:XD0020 (invalid combination of serialization
> parameters).
>
> The test suite contains a test (err-c0001-001.xml) that does the
> following:
>
> <p:pipeline> <p:serialization port="result" method="nonexistant"/>
> <p:identity/> </p:pipeline>
>
> The test is expected to fail with err:XC0001.
>
> The spec says the following about p:serialization (section 5.6):
>
> "If the pipeline processor serializes the output on the specified
> port, it must use the serialization options specified. If the
> processor is not serializing (if, for example, the pipeline has been
> called from another pipeline), then the p:serialization must be
> ignored. The processor may reject statically a pipeline that requests
> serialization options that it cannot provide."
>
> Our implementation does not serialize - except when you use the
> command-line interface, in which case we of course apply the provided
> serialization options to the output ports.
>
> Because both err:XC0001 and err:XD0020 are dynamic errors, we don't
> reject invalid/unsupported p:serializations statically. Only when the
> processors tries to use them, we report an error.
>
> My question is:
>
> - Is the test correct?
> - Or: does the test suite assume, for the purpose of easy testing of
> serialization options, that serialization options are always applied
> to the top-level pipeline (as if it was ran from the command-line)?
> - Or: are we doing something wrong in our implementation?

Or, maybe this is just an area where the spec isn't clear. I think my
implementation checks the serialization options when it encounters the
p:serialization element, whether or not the results are ever
serialized.

I don't think we currently say whether that is required or even
allowed.

                                        Be seeing you,
                                          norm

-- 
Norman Walsh <ndw@nwalsh.com> | A man who is 'of sound mind' is one who
http://nwalsh.com/            | keeps the inner madman under lock and
                              | key.--Paul Valéry