Oleg,

Am stating my personal pov why this is something i *personally* may
not be interested in. Nothing more, nothing less. If anyone else wants
to run with it, that's fine with me.

thanks,
-- dims

On Sat, Oct 4, 2008 at 2:01 PM, Oleg Kalnichevski <[EMAIL PROTECTED]> wrote:
> On Thu, Oct 02, 2008 at 12:22:17PM -0400, Davanum Srinivas wrote:
>> Oleg,
>>
>> One problem from my personal point of view is that for JAXWS support
>> we need javamail :( to be compliant to the spec.
>>
>> -- dims
>>
>
> Davanum,
>
> I understand your position, yet I find it regrettable that JAX-WS
> specific requirements dictate architectural decisions for Axiom which is
> meant a generic XML processing framework. I also find it regrettable
> that Axiom is currently unable to parse and build MIME messages without
> buffering the first MIME part (SOAP message) in memory. It may not be a
> big problem for a great deal of applications, but for some it certainly
> is.
>
> Oleg
>
>
>
>
>> On Thu, Oct 2, 2008 at 11:48 AM, Oleg Kalnichevski <[EMAIL PROTECTED]> wrote:
>> > On Sun, Sep 28, 2008 at 10:54:33AM -0500, Thilina Gunarathne wrote:
>> >> Hi Oleg,Apologies for been this late to reply...
>> >>
>> >> There is a little catch that you might need to consider.. Axis2 MIME 
>> >> parsing
>> >> works similar to StAX. Axis2 MIME parser does differed parsing of the MIME
>> >> messages. It parses MIME parts only when needed and requested by the AXIOM
>> >> model. Axiom requests the contents of the attachments only when somebody
>> >> reads the contents of the corresponding OMText object (not when it creates
>> >> the object model). So altogether when using MTOM+ Axiom it's a double
>> >> layer differed parsing, differed parsing of XML & differed parsing of
>> >> Attachments. But to be honest I'm not sure whether there are any people
>> >> taking advantage of this.. One reason for us not use the JavaMail was it's
>> >> inability to do differed MIME parsing..
>> >> Also Axis2 supports streaming of attachments. That means somebody can
>> >> directly stream the attachments stream from the input to the output. This
>> >> comes handy when proxying or mediating. This case might become unusable (I
>> >> can't really remember how it works now :(.. ) if we read and parse the 
>> >> whole
>> >> MIME message.. Once again I'm not sure whether people really use it
>> >> (Synapse??)..
>> >>
>> >> All and all I believe the ability to differed parse MIME attachments will 
>> >> be
>> >> a good feature for MIME4J too..
>> >>
>> >
>> > Thilina,
>> >
>> > mime4j is perfectly capable of deferred (or on demand) parsing of MIME
>> > messages. One can retrieve just the first mime part and then defer the
>> > processing of all subsequent parts only when / if accessed.
>> >
>> > Presently the patch I submitted (WSCOMMONS-387) does not provide support
>> > for the deferred parsing, but it should be relatively trivial to add it.
>> > However, as no one showed any interest in WSCOMMONS-387 so far I am
>> > somewhat reluctant to put any more work into it.
>> >
>> > Oleg
>> >
>> >
>> >> thanks,
>> >> Thilina
>> >>
>> >> On Wed, Sep 17, 2008 at 1:06 PM, Oleg Kalnichevski <[EMAIL PROTECTED]> 
>> >> wrote:
>> >>
>> >> >
>> >> > Thilina,
>> >> >
>> >> > Mime4j parser works similarly to SAX. It is an event based API. One can
>> >> > provide a custom handler implementing a particular custom processing
>> >> > logic in response to appearance of a certain MIME element in the data
>> >> > stream, such as a MIME header or a content part.
>> >> >
>> >> > Oleg
>> >> >
>> >> > > thanks,
>> >> > > Thilina
>> >> > >
>> >> > >
>> >> > > > Oleg Kalnichevski wrote:
>> >> > > >
>> >> > > >> Folks,
>> >> > > >>
>> >> > > >> Would there be any interest in a SwA implementation based on Apache
>> >> > > >> mime4j [1]? Mime4j can handle very complex MIME messages, is
>> >> > reasonably
>> >> > > >> fast, and, most importantly, can stream complex MIME messages in 
>> >> > > >> and
>> >> > out
>> >> > > >> with a predictable memory footprint (using just a small internal
>> >> > buffer
>> >> > > >> of a constant length)
>> >> > > >> I _personally_ find Java Activation API pretty suboptimal and would
>> >> > like
>> >> > > >> Axiom to provide an alternative API based on a fully streamable 
>> >> > > >> model.
>> >> > > >> Let me know what you think.
>> >> > > >>
>> >> > > >> Oleg
>> >> > > >>
>> >> > > >> [1] http://james.apache.org/mime4j/index.html
>> >> > > >>
>> >> > > >>
>> >> > > >>
>> >> > > > --
>> >> > > > Sanjiva Weerawarana, Ph.D.
>> >> > > > Founder & Director; Lanka Software Foundation;
>> >> > http://www.opensource.lk/
>> >> > > > Founder, Chairman & CEO; WSO2, Inc.; http://www.wso2.com/
>> >> > > > Member; Apache Software Foundation; http://www.apache.org/
>> >> > > > Visiting Lecturer; University of Moratuwa; http://www.cse.mrt.ac.lk/
>> >> > > >
>> >> > > > Blog: http://sanjiva.weerawarana.org/
>> >> > > >
>> >> > >
>> >> > >
>> >> > >
>> >> >
>> >> >
>> >>
>> >>
>> >> --
>> >> Thilina Gunarathne  - http://thilinag.blogspot.com
>> >
>>
>>
>>
>> --
>> Davanum Srinivas :: http://davanum.wordpress.com
>



-- 
Davanum Srinivas :: http://davanum.wordpress.com

Reply via email to