URL_Fetcher

I looked at the sources of AcornHTTP and was amazed to see various references to chunked transmit. This may be just what I want for IPP printing. I thought a few experiments were in order.

The first problem is that I have been unable to get URL_GetURL to target any port other than 80 for http. I’ve tried suffixing the URL with :631, :631/, /:631 and /:631/. All result in it trying to connect to port 80.

Does anyone know how to make it work? Has anyone tried it with a non-default port?

The URL I’m using is http://LinuxBox/printers/Epson-Stylus-Photo-RX620 with any of the suffices above. The host is resolved (to 192.168.16.47), and Wireshark running on LinuxBox shows attempts to connect to 80/TCP.

URL_Fetcher is version 0.58 (11 Feb 2023) and AcornHTTP is 1.09 (27 May 2023).

The URL I’m using is http://LinuxBox

LinuxBox shows attempts to connect to 80/TCP

What does https://linuxbox/ etc show in Wireshark, then check what does http://linuxbox:443 show?
Just wondering whether AcornHTTP is assuming HTTP is always port 80 and https is always 443

I just tweaked a simple URL fetcher written in BASIC to fetch http://192.168.1.10:23/ and it responded with a bunch of ANSI codes (the BBS) rather than the weather (on port 80).

Thank you both. Shortening the URL to just http://LinuxBox:631 does indeed open port 631 and send to it. Not only that, but the Linux box then renders a printout! Despite having been given an incomplete URL. Presumably this only works because port 631 is only used for IPP, and the Linux box’s printing system has chosen to send it to the default (or only) printer.

I’ve been running a call to URL_ParseURL too, which shows that http://LinuxBox:631 parses the port number out, but adding even one level of URL, e.g. http://LinuxBox/printers:631, fails to parse the port number and attempts to connect to port 80.

So the URL parsing appears to be deficient.

I don’t know if sending to an incomplete URL would work in all cases – I suspect not – but at least this gives me a test bed to try chunked transmit using AcornHTTP.

So the URL parsing appears to be deficient.

How? The port should follow the hostname, not be tagged onto the end.

http://LinuxBox/printers:631

Should be http://LinuxBox:631/printers

See RFC 1738 part 3.1

The port should follow the hostname, not be tagged onto the end.

Well, I never knew that. You live and you learn.

And it works: http://LinuxBox:631/printers/Epson-Stylus-Photo-RX620

Thanks again for your help.

Shortening the URL to just http://LinuxBox:631

http://Linuxbox.domain:631 would be nicer

Presumably this only works because port 631 is only used for IPP, and the Linux box’s printing system has chosen to send it to the default (or only) printer.

Yup. Default print queue.
Be aware that certain things will give you problems in certain circumstances:

1. Spaces in the queue share name¹
2. Share names longer than 10 characters

The port should follow the hostname, not be tagged onto the end.

If you’re picky, that is the end, of the host identifier; the stuff after / is a parameter defining the specific queue you expect to be used.

¹ A terminally dumb feature of having Microsoft involved in anything – spaces are delimiters, end.