xmlgraphics-fop-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Art Welch <ar...@EASTPOINT.COM>
Subject RE: Streaming PDF
Date Thu, 29 Mar 2001 19:31:10 GMT
Yes, true. I just have this strange nagging feeling that I am missing
something (probably because I only read a couple of sections of the
specification document before writing it). It seems too simple. If my
understanding is not correct, I would hate to lead others astray.

Usually when I have a strange feeling about a piece of code it means that
there is a bug or logic problem that my unconscious mind sees, but does not
want to convey to my conscious mind for some reason. Or maybe it was just
the fish...

Semi-conscious,
Art

-----Original Message-----
From: Kelly Campbell [mailto:camk@channelpoint.com]
Sent: Thursday, March 29, 2001 2:33 PM
To: 'fop-dev@xml.apache.org'
Subject: RE: Streaming PDF


That's what opensource is for right? Collaborative development. More eyes to
find the problems and fix them. :-)

> -----Original Message-----
> From: Art Welch [mailto:art_w@EASTPOINT.COM]
> Sent: Thursday, March 29, 2001 12:05 PM
> To: 'fop-dev@xml.apache.org'
> Subject: RE: Streaming PDF
> 
> 
> I will ponder this for a bit. There are a few things that 
> make me nervous
> about this code. A couple of times during development it mysteriously
> stopped working and then equally mysteriously started working 
> again. It has
> remained working ever since (but I have not really been 
> banging on it). I
> developed it after a cursory review of the specifications and 
> looking at
> some other implementations. Basically it was a "quick and 
> dirty" test to see
> if it would work. I am also not sure I understand the implications of
> chunked or HTTP 1.1...
> 
> I must admit that the implementations I found were either 
> commercial or
> infected with GPL. But I did not look very hard either. After 
> seeing the
> implementations (not much to them) and looking at the spec 
> (not much to it -
> at least as far as chunked) (this probably means that I am 
> missing something
> - hence my reluctance). I decided to roll my own.
> 
> Pensively,
> Art
> 
> -----Original Message-----
> From: Kelly Campbell [mailto:camk@channelpoint.com]
> Sent: Thursday, March 29, 2001 1:58 PM
> To: 'fop-dev@xml.apache.org'
> Subject: RE: Streaming PDF
> 
> 
> Art, I'd be happy to take what you've got and clean it up and 
> integrate it
> into the Fop servlet we include with FOP. I think it would be 
> a good thing
> to have. I was looking around for implementations we could 
> bundle and use,
> but didn't get very far initially.
> 
> -Kelly
> 
> > -----Original Message-----
> > From: Art Welch [mailto:art_w@EASTPOINT.COM]
> > Sent: Thursday, March 29, 2001 11:34 AM
> > To: 'fop-dev@xml.apache.org'
> > Subject: Streaming PDF
> > 
> > 
> > I saw streaming PDFs mentioned in a recent e-mail. I do not 
> > know if what I
> > am about to say will be relevant to the original topic, but I 
> > thought that I
> > would relate the information anyway.
> > 
> > I had been having a problem in my environment where PDFs generated
> > dynamically by FOP running in a servlet were not rendered 
> > properly in M$ IE
> > 5. It was discovered that large PDFs (> 8k) displayed 
> > properly, but small
> > ones did not display at all. Further discussion on the list 
> > suggested that
> > IE requires that either the content-length be set or the data 
> > sent HTTP
> > chunked. I do not like having to store a potentially quite 
> > large PDF just so
> > that I can find out how big it is to set the header. So I 
> > tried implementing
> > an HTTP Chunked OutputStream. This was not too difficult to 
> > do and it kind
> > of works, except that in my testing I found that even with 
> the chunked
> > output, small (< 8k) PDFs STILL DO NOT DISPLAY IN IE5! I then 
> > changed the
> > chunked output stream to use an 8k buffer to build the 
> > chunks. If the output
> > is closed before the buffer is filled the first time then I 
> > just set the
> > content-length header and send the contents of the buffer 
> > non-chunked. If
> > there is more output after the buffer has been filled then 
> > the data is sent
> > chunked. So far this seems to be working, but I have not tested it
> > rigorously. I also understand that since this requires HTTP 
> > 1.1 it may not
> > work in the general case.
> > 
> > I present this information in the hopes that it will be 
> > helpful to someone.
> > I do not think that FOP should attempt to include chunked 
> > output capability.
> > It should write to a stream - the stream implementation 
> > determined by the
> > external application. For anyone interested in doing the same 
> > thing, there
> > are a few chunked stream implementations out there - or you 
> > could write your
> > own. My implementation is kind of limited and not well 
> > tested, so I am not
> > willing to make it available at this point. Maybe I will 
> > reconsider after I
> > have done more testing - but at the moment I am preoccupied 
> > with migrating
> > my application to XalanJ2 from XalanJ1 and have not been able 
> > to look at FOP
> > or output related things for a while.
> > 
> > FYI,
> > Art
> > 
> > 
> ---------------------------------------------------------------------
> > To unsubscribe, e-mail: fop-dev-unsubscribe@xml.apache.org
> > For additional commands, email: fop-dev-help@xml.apache.org
> > 
> 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: fop-dev-unsubscribe@xml.apache.org
> For additional commands, email: fop-dev-help@xml.apache.org
> 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: fop-dev-unsubscribe@xml.apache.org
> For additional commands, email: fop-dev-help@xml.apache.org
> 

---------------------------------------------------------------------
To unsubscribe, e-mail: fop-dev-unsubscribe@xml.apache.org
For additional commands, email: fop-dev-help@xml.apache.org

---------------------------------------------------------------------
To unsubscribe, e-mail: fop-dev-unsubscribe@xml.apache.org
For additional commands, email: fop-dev-help@xml.apache.org


Mime
View raw message