Chunked data is basically what you use if you don't know the content length. It's in RFC2616. You basically transfer a length, followed by the data, and then a new length, new data, and so on until you've transferred it all. And then you send an end-of-data marker. No content length needed. Johnny On 2012-04-13 18:17, Paul_Koning at Dell.com wrote: RFC 2616 is the authority on all this. Judging by a quick scan: what do you mean by "chunk data"? If you don't specify a transfer-coding, then yes, I think 8 bit clean is assumed. But since you mentioned "chunk" do you mean that you're specifying chunked encoding? I didn't read all that but there's a lot of content about that in the RFC. You'll need a content-length header, unless the connection is closed after the jpg file is transferred. See section 4.4 of the RFC. paul -----Original Message----- From: owner-hecnet at Update.UU.SE [mailto:owner-hecnet at Update.UU.SE] On Behalf Of Johnny Billquist Sent: Friday, April 13, 2012 12:07 PM To: hecnet at Update.UU.SE Subject: Re: [HECnet] Web-server under RSX-11M-PLUS On 2012-04-13 17:01, Sampsa Laine wrote: I also noticed a disturbing lack of funny cat pictures on the demo site... That is related to my most immediate question. Exactly how does pictures get served in http. I tried implementing it, but I must have something wrong, so I'm trying to figure it out right now. Questions are: I assume the TCP stream is to be 8-bit clean. That is - no special interpretation of the byte 255 (as is done in telnet). I assume that pictures should be ok with "Content-type: image/jpeg" I assume chunk data should work for this I assume that all the bytes of the image should just come in the stream with nothing more involved. Can anyone comment on those assumptions? I have a jpeg image on mim, which I'm using as my experiment. I have checked that the contents match that of the original file which I copied, so I think the actual image is served right... Johnny Sampsa On 13 Apr 2012, at 17:58, Mark Wickens wrote: Johnny, That's really fantastic! Just one question, where are the dancing dogs? ;) Regards, Mark On 13/04/12 13:41, Johnny Billquist wrote: This is perhaps slightly offtopic, but fun enough that I like to announce it anyway... === Ok. I just thought I'd stir some interest and just general noise by announcing that I've written a small web-server running under RSX. It's written in BASIC+2, and uses the TCP/IP stack for RSX that I've also written. There are probably a bunch of bugs and issues still around, so I'm happy to take any bug reports, comments or whatever. The url is http://madame.update.uu.se/, and if anyone is curious about the code, it's at mim.update.uu.se (same machine, other IP), under MIM::DU:[HTTPD]WWW.B2S (also on HECnet) If people have any interest in this stuff, or something else/more, I'm interested in hearing about it. The TCP/IP stack will eventually (soon) be available for others to download and use, and apart from the web server, I've also written a telnet client, and a few small services under TCP, as well as some tools for administration. I have some polishing to do, I need to finish a DNS resolved, and I'd like to also finish FTP and a telnet server, but I might be open to distributing things before I've finished all those things, especially if someone is interested in helping writing stuff. I have interfaces completed for BASIC+2, PDP-11 C, Macro-11. FORTRAN 77 should also work, but I haven't tried it yet. This all runs under RSX-11M-PLUS V4.6, but I think it should be possible to get running under almost any M+ version, but there might be some hacking needed for some versions. It will not work under 11M, and I never expect it to. One or two drivers as well as one or two tools really are big enough that I need to use the split I/D space feature in M+. Rewriting stuff to not need that is way too much work. Johnny -- Johnny Billquist || "I'm on a bus || on a psychedelic trip email: bqt at softjar.se || Reading murder books pdp is alive! || tryin' to stay hip" - B. Idol