[mashup-dev] ScraperService or ScraperHostObject?

[thilina at wso2.com]

> 1. WH allows you to include other config files using a include tag
> (Cannot use it in ScraperService scenario)
> 2. Cannot read and write to the file system directly through WH using
> its file tag.
Are we planning on using the FileHostObject for the above or are we
allowing users to specify arbitrary files. We might need to be careful to
keep the file system restriction we imposed in the file system host
object.
> 3. A WH config can define variables usinf var-def tag and during
> excution of the config it will assign various parts of the data to these
> variables (Depending on the way the config is written). The user can
> then query the result using the defined variables (Even our host object
> does not support this scenario as yet. It can be implemented trivially
> though). This cannot be done in the Scraper service case too.
Well.. Are we going to map these to java script variables?

IMHO it's better if we can identify the use-cases which we need to support
in the mashup server. Then supporting those requirements using the web
harvest features rather than trying to provide all the web harvest
features :).. This is just my personal preference.
>
> I'm more inclined to the HostObject case but its not only xpath that we
> will be loosing. We will be loosing xquery and xslt too.
>
> I'm not exactly sure of the consequences we will have if we have saxon
> in the global class path. Which would help us get 100% of WH using a
> hostobject but this is going to effect all other XSLT processing stuff
> that we have (?stub, ?try-it).
Axis2 WSDL generation gets screwed when saxon is present in the classpath..

thanks,
Thilina
>
> Note: Saxon issues a warning when processing XSLT 1.0 docs saying that
> an XSLT 1.0 doc was processed using a XSLT 2.0 processor.
>
> Thanks,
> Keith.
>> That way it feels like a host object but the object is simply a stub
>> to the service.
>>
>> Sanjiva.
>>
>> Jonathan Marsh wrote:
>>> If XPath doesn't work in the host object, it doesn't give us much
>>> choice.
>>> But in all other respects I think a host object is preferable.
>>>
>>> Jonathan Marsh - http://www.wso2.com -
>>> http://auburnmarshes.spaces.live.com
>>>
>>>> -----Original Message-----
>>>> From: mashup-dev-bounces at wso2.org [mailto:mashup-dev-bounces at wso2.org]
>>>> On Behalf Of Keith Chapman
>>>> Sent: Friday, August 31, 2007 4:34 AM
>>>> To: mashup-dev at wso2.org
>>>> Subject: [mashup-dev] ScraperService or ScraperHostObject?
>>>>
>>>> Hi all,
>>>>
>>>> I managed to get the scraper service working. Which means that we have
>>>> both a host object and a service at the moment. Both mechanisms have
>>>> there advantages and disadvantages.
>>>>
>>>> Having as a host object.
>>>> ---------------------------------------
>>>>
>>>> 1. Convenient to use
>>>> 2. Cannot support xpath (Unless we have saxon in our class path)
>>>> 3. Can directly scrape a page and sage parts of it on to the file
>>>> system
>>>> through the webharvest config file itself.
>>>>
>>>>
>>>> Having a service.
>>>> ----------------------------
>>>>
>>>> 1. Can use xpath (Cause we can include saxon in the lib of the service
>>>> (does not effect the global classpath)).
>>>> 2. Cannot use file elements in config.
>>>>
>>>>
>>>> We cannot have both of these implementations too (Unless we have saxon
>>>> in the global classpath). They cant coexist together cause the first
>>>> requires webharvest  to be on the global classpath and the second
>>>> method
>>>> requires it to be inside the lib of the service.
>>>>
>>>> Are we going to decide to use one of these?
>>>>
>>>> What do you guys think?
>>>>
>>>> Thanks,
>>>> Keith.
>>>>
>>>> _______________________________________________
>>>> Mashup-dev mailing list
>>>> Mashup-dev at wso2.org
>>>> http://www.wso2.org/cgi-bin/mailman/listinfo/mashup-dev
>>>
>>>
>>> _______________________________________________
>>> Mashup-dev mailing list
>>> Mashup-dev at wso2.org
>>> http://www.wso2.org/cgi-bin/mailman/listinfo/mashup-dev
>>>
>>
>
>
> _______________________________________________
> Mashup-dev mailing list
> Mashup-dev at wso2.org
> http://www.wso2.org/cgi-bin/mailman/listinfo/mashup-dev
>