Mailing-List: contact cxf-dev-help@incubator.apache.org; run by ezmlm
Precedence: bulk
Reply-To: cxf-dev@incubator.apache.org
Received-SPF: pass (herse.apache.org: domain of daniel.kulp@iona.com
 designates 65.223.216.181 as permitted sender)
From: Daniel Kulp <daniel.kulp@iona.com>
Organization: IONA
To: cxf-dev@incubator.apache.org
Subject: Re: Can we distribute the sun mail and activation jar files in our
 distribution?
Date: Tue, 20 Feb 2007 18:15:25 -0500
User-Agent: KMail/1.9.5
References: <45DAE8E9.4070804@gmail.com>
 <200702201729.22555.daniel.kulp@iona.com>
 <7b774c950702201452i1d3847d1h5fc296a2fd07a9f4@mail.gmail.com>
In-Reply-To: <7b774c950702201452i1d3847d1h5fc296a2fd07a9f4@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain;
  charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
Message-Id: <200702201815.25978.daniel.kulp@iona.com>


Dan,

On Tuesday 20 February 2007 17:52, Dan Diephouse wrote:
> Any one know if any implementations send ^M in attachment headers? I could
> see this causing issues if so.
> - Dan

The issue is actually the reverse.   The spec says the lines should be CR/LF 
terminated.   If something sends something that is only LF terminated, Sun's 
implementation will parse it fine.   Geronimo's won't.  Thus, you need to 
find implementations that DON'T put the CR(^M)'s.   

That said, some of our own code in the Attachment processing doesn't 
work "correctly" if it's not CR/LF.   It just didn't cause a test failure.   
There's at least one place where we "break" on CR, then just blindly 
ins.read() the next char assuming it's the LF.   (FYI: that's exactly how the 
Geronimo specs version of mail does it as well)  I think Sun's impl must 
ignore the CR (skip over it) and breaks on the LF (or breaks on the LF 
and "trim()"s the result).

Dan


>
> On 2/20/07, Daniel Kulp <daniel.kulp@iona.com> wrote:
> > Rick,
> >
> > On Tuesday 20 February 2007 10:35, Daniel Kulp wrote:
> > > On Tuesday 20 February 2007 10:15, Rick McGuire wrote:
> > > > Geronimo has implementations of activation-1.1 and mail-1.4.  I was
> >
> > the
> >
> > > > author of both.
> > >
> > > Oh.   Cool.   Last I had checked (early Dec), they were in trunk, but
> >
> > not
> >
> > > yet available on the maven repositories.  (not released)   I see they
> >
> > are
> >
> > > there now.   And annotation.   Good.  I'll see if CXF can build/run
> > > with them.
> >
> > OK.  I managed to get CXF to use the geronimo specs versions of
> > annotation,
> > servlet, mail, and activation.    I had some problems with Mail.   The
> > Geronimo Specs version is definitely not a 100% drop in replacement for
> > Sun's
> > 1.4 versions.   The InternetHeaders object from Sun is much more tolerant
> > of "out of spec" streams (example: unix EOL style) than Geronimo.   We
> > had a
> > couple tests that failed, but it's because of bad test data.   As soon as
> > I
> > did a unix2dos on the test data, the tests passed.
> >
> > Anyway, I'm committing the changes to CXF now.
> >
> > --
> > J. Daniel Kulp
> > Principal Engineer
> > IONA
> > P: 781-902-8727    C: 508-380-7194
> > daniel.kulp@iona.com

-- 
J. Daniel Kulp
Principal Engineer
IONA
P: 781-902-8727    C: 508-380-7194
daniel.kulp@iona.com