help with web interface

Hi

I have a data logging type program running under RISC OS

I would like if possible to send and receive simple data between my program

The most versatile interface I imagine would web based? . as this
could then be accessed on different platforms and different hardware.

I would like to be able to do simple tasks like:-

Password protected access
Read the current time.
Set current time.
Send data
receive date

I have quite a lot of programming experience with RISC OS but very little
knowledge of web based programs I have looked at webjames could this do what I am looking
for?

Anyway I am looking for some guidance before I start delving into solutions, Can RISC OS do
what I want and what do think I would need?

At the moment my program is a single tasking graphic program but it could be made to multi task.

Regards

Terry

What you want is not difficult the most difficult bit is understanding sockets.

Basically you need to choose a port write a server to accept incoming connections on that port and parse commands from the client and act on them.

You then need to write the client.

The ‘write a server’ ‘write a client’ but sounds complicated but it isn’t as bad as it sounds and each can probably be done in a few of screenfuls of code. The terminology is scarier than the implimentation.

When connected to the server the client would send an authentication command to verify user and disable all other commands until authentication succeeds – the authentication is nothing to do with sockets its a protocol that you make up.

The protocol – that you make up – may go something like

Client sends ‘auth user password’
server replies ‘100 ok’ or ‘201 fail’
client sends ‘readtime’
server replies ‘201 fail’ or ’100 ok 12:15 Wed 01-01-2016
… and so on

‘readtime’ would fail if ‘auth’ failed

The numbers would be error codes – don’t use them if you don’t want them

You can simplify the multitasking part by running in a taskwindow

What language do you want to write this in?

Hi

It can be in C or BASIC I don’t mind, what I would like is something
that works with a web browser that’s why I was looking at webjames as
that supports CGI (which I am reading up on at the moment :-) etc

Regards

Terry

That’s not something I’ve tried. I’ll leave any further suggestions to CGI experts

I have a data logging type program running under RISC OS
I would like if possible to send and receive simple data between my program

You need to clarify ‘between my program’ and what else?

At the moment my program is a single tasking graphic program but it could be made to multi task.

I would suspect that it would need to be multi-tasking, as it will (presumably) need to be do the logging and communicate at the ‘same’ (human) time?

Hi

The communication would be between my program and
a web browser eg a web page would be served to set
or read the internal clock and then some other web page
could supply the data that has be collected.

Another web page could be used to set up configuration data
I think I am looking for something similar to how my ADSL box
interfaces with a web browser.

The program can be changed to multi task that is not a problem
I just need some guidance with regards to the web based side of things.

Regards

Terry

I think I am looking for something similar to how my ADSL box
interfaces with a web browser.

Try looking in NAT or port forwarding in your router’s configuration. Open the port (can be 80 for http), point it at your machine. Everything connecting to your router’s public IP will be pushed to your server.
Note you can’t use the public IP address within your network, your router will notice you are “inside” and probably give you its server instead. Use the 192.168.×.x address in this case.

You can get a domain name alias from NoIP.com. Useful if your ADSL router can automatically sync your public address with NoIP. Orange Livebox can. I know, mine does.

Be very very aware that you will get hundreds of connections pretending to be the BaiduBot and often making really weird requests. It may be better to rely upon WebJames to be the server (mine has been coping with it since my last reboot 42 days ago).
However with this in mind and with the fact that you cannot rely upon public communications links to be secure, be very VERY wary about placing the ability to alter the time and/or logging behind a CGI script.
I have some CGI code on my server to view log files and the like, but by intention there is nothing that can perform a destructive act.

Don’t take my word for it. Fire up WebJames, put in a dummy index and delete the default CGI scripts, open a port, give it a few days, then check your logs.
Oh, and you think that’s bad, you don’t wanna know how many bogus connection attempts my own custom server gets…

Web2.0 is not a friendly place. Work with that knowledge from the outset.

Hi Terry,

I see no reason why RISCOS can not do what you are looking for. The latest formal build is a very stable and versatile system (client or server).

Writing a web interface is not as scary as it sounds and Webjames will let you write CGI scripts in Basic. Don’t delete the default CGI scripts. Modify copies of them to learn how they work.

I use !Wget on the Client side of things. It is also very easy to drive from RISCOS (and many other systems). Sending messages/data from one RiscOS system to another is very simple using the combination of !Wget on a client to !Webjames on a server.

The issue about using a public IP address from within a local network depends on how good your router is. Many will do it fine – Search for IP loopback. (I use Draytek)

The number of rogue connections you will get depends on many factors (including how good your ISP is). I have servers running Apache on Linux and Webjames on RISCOS. I get very few rogue connections on tho port 80 apache and none on Webjames with higher port numbers on a number of RasPi systems. In any case, most hackers would not know what to do with a RiscOS/Webjames system so will just go elsewhere.

This thread got me to thinking about writing a simple server (in C). Something I’d been meaning to do for a long time.

http://www.heyrick.co.uk/blog/index.php?diary=20160104

The latest formal build is a very stable and versatile system (client or server).

A self-built RISC OS from May(ish) sources (IIRC!) is running happily after 42 days. I’m sure it can do more, but as a developer I reboot from time to time. I’m not doing much low level development work because I’m just wondering how long everything will keep running…
…until there’s a power cut. ;-)

Hmm… Has anybody recorded the longest known uptime of a RISC OS machine?

Don’t delete the default CGI scripts. Modify copies of them to learn how they work.

That’s fair enough, but don’t put the server on-line with any of the default things in place.

The issue about using a public IP address from within a local network depends on how good your router is. Many will do it fine – Search for IP loopback. (I use Draytek)

Advanced routers, custom (non-ISP) ones, and anything with the likes of OpenWRT perhaps.

I’m not so sure a bog standard ISP provided router will do fancy stuff like that.
;-) On the other hand, my router aliases devicename.home, so kiichigo.home will connect to the Pi. Kiichigo is raspberry in Japanese, not terribly original… Saves remembering a bunch of IP addresses.

The number of rogue connections you will get depends on many factors (including how good your ISP is).

I think it comes down to the criteria of how the ISP implements the blocking. Because there is a trade-off between stopping rubbish and blocking too much. I remember reading that one ISP (American?) blocks the entirety of China. It appears from my examinations that the dross comes from hijacked machines. I don’t think I’ve been poked by the same IP address twice (and I’ve had 593 attempts since Saturday evening). The addresses are mostly China, but I’ve seen Israel, Turkey, Spain, and Moldova. There are probably many more but I just pick addresses at random to look up, most of them get dropped on the floor as life is too short to care where ~600 machines are located.

I get very few rogue connections

I wonder how your ISP blocks, and if it inconveniences legitimate visitors?

In any case, most hackers would not know what to do with a RiscOS/Webjames system so will just go elsewhere.

True, they will. With my server (on a telnet port), they connect and sit idle for a minute until the server times them out and kicks them off. I think things would be more interesting if the server looked like a Unix login prompt. But it doesn’t.

However it is useful to be aware of this stuff, because the “hackers” (in quotes, as it is a script, not anything real) are stupid and will keep connecting for as long as there is a reply. My server will notice quick-fire connections and will blacklist that IP address. After a set number of reconnects, it will simply abort the connection as soon as it is noticed. This stops the bot. If it gets a reply it will keep on going for maybe as many as 40 or 50 attempts. Slam the door on it, it goes away.
My blacklist is a circular list of 16 IP addresses. I don’t need more, as I said, the same machine doesn’t call back.

It isn’t a big deal. It doesn’t harm anything, and it is useful for stress testing the server. ~150 connections a day is…about 149 more than I’d expect otherwise…so I’m not bothered if the botnet wants to poke the server, it can cope.

Hi

Thanks for the replies I think I will continue with Webjames at the moment.
Security is not such a big problem because the unit is probably not going
to be connected to the outside world, only internal.

Anyway back to the books

Regards

Just looked at my stats. It looks like I’ve had about 360 connection attempts today. Must be a rainy day in Shengzen…

Hmm… Has anybody recorded the longest known uptime of a RISC OS machine?

Well no recorded time, but I am aware of at least 10 A3/4xx Class machines which have only been powered down a handful of times, with the odd update to the software.

Otherwise they have been on and working since the day they were installed RISC OS 3.11.

Michael: I’m intrigued, what are they doing?
1992-2016 24 years!

Chris:

Testing!

The actual machines started off with Arthur and then RISC OS 2 then 3…

I need to get my dad to write his story to do a meeting at ROUGOL :)