[Prev][Next][Index][Thread]
Re: a question about converter
-
To: Henrik Frystyk Nielsen <frystyk@w3.org>
-
Subject: Re: a question about converter
-
From: Jiang Tao <jiangt@ceci.mit.edu>
-
Date: Mon, 01 Apr 1996 15:03:36 -0500
-
Cc: www-lib@w3.org
-
From jiangt@ceci.mit.edu Mon Apr 1 15: 03:45 1996
-
Message-Id: <31603698.61133CF4@ceci.mit.edu>
-
Organization: CECI
-
References: <315C5B73.237C228A@ceci.mit.edu> <199603300214.VAA05412@anansi.w3.org>
-
Sender: jiangt@ceci.mit.edu
-
X-Mailer: Mozilla 2.0 (X11; I; SunOS 4.1.4 sun4c)
Henrik Frystyk Nielsen wrote:
> parser. As the MIME header gets parsed, the MIME parser finds out the "true"
> media type of the MIME body and then calls the stream stack again. What are
> the symptoms? Do you have a trace of the line mode browser?
Thank you, I have solved the problem. I registered three converters:
HTTPStatus_new, MIMEParser and HTXParse. I find that HTTPStatus_new will handle
the data from http server first which always have the type "text/x-http",
then call stream stack, and pass data to MIMEParse, which will analyse
the MIME header, get the content type (such as "text/html, image/gif etc"),
and remove the MIME header from the data and call stream stack again,
then pass it to HTXParse where I can process the data according to
its content type.
So suppose we registered many converters including HTTPStatus_new
and MIMEParse, then HTTPStatus_new and MIMEParse will always be called
first, then the best match converter will be called according to the
input format, output format and quality factor.
Is my understanding right?
--
Tao Jiang Visiting Scientist, CECI, MIT
email: jiangt@ceci.mit.edu URL: http://ceci.mit.edu/staff/jiangt/
Follow-Ups: