Page 1 of 1

HTTP Server?

Posted: Tue Dec 11, 2018 1:21 pm
by NC3Z
Under CAT Controls > User Interface Is this working??

Re: HTTP Server?

Posted: Tue Dec 11, 2018 8:43 pm
by w-u-2-o
It has some very rudimentary and fragmented functionality and is not by any means a finished capability. I've tried it and it shows great promise for the future. I have no idea who the dev's are/were on this, but perhaps we should start sending them free pizza and beer ;)

Re: HTTP Server?

Posted: Tue Dec 11, 2018 9:05 pm
by NC3Z
I have tried enabling it but no joy, even ports scanning shows no port responding. No big deal, just wondering.

Re: HTTP Server?

Posted: Tue Dec 11, 2018 9:30 pm
by w-u-2-o
Gary: open the port as http://127.0.0.1:80 (or whatever your port number is) on the same machine. If memory serves, you have to enable it, then restart Thetis.

Re: HTTP Server?

Posted: Tue Dec 11, 2018 9:39 pm
by NC3Z
I already tried all those tricks with no joy. Network port scanner can't find anything either.
But no issue, just was curious. Other issues need to be addressed.

Re: HTTP Server?

Posted: Mon Jan 07, 2019 8:34 pm
by Jay_N2MGA
I know this thread is old but I was able to get the http server working by turning off Thetis, selecting the "On: Advanced" check box shown on the Cat Control/User Interface setup screen, turning on Thetis then browsing to the local IP 127.0.0.1:8081 port # address. Selecting "On: Basic" did not work for me.

cheers!
jay
n2mga

Re: HTTP Server?

Posted: Mon Jan 07, 2019 10:21 pm
by NC3Z
Thanks for the hint, indeed it worked on Advanced, and not on Basic.

Re: HTTP Server?

Posted: Tue Jan 19, 2021 11:33 pm
by PD3LK
almost 4 years later .... is the developer still waiting for pizza's :?:

Re: HTTP Server?

Posted: Wed Jan 20, 2021 12:28 am
by w-u-2-o
Welcome to the wonderful world of ham radio open source development, Leon. It is such a small and highly specialized world that it is a rare thing indeed to find any developers at all. We have been extremely lucky to obtain what we have.

IMHO I'd rather see a dev. tackle a true client/server split of Thetis than muck around with some web server implementation.

Re: HTTP Server?

Posted: Wed Jan 20, 2021 8:59 am
by laurencebarker
I've worked up an outline design with Doug and Warren for a client/server split in Thetis. It is a monstrous amount of effort - probably a person-year of full time work. Done part time in evenings, potentially several years. There is something like 600 function calls that need to be replaced by IP messages. Warren has recommended that more of the Thetis functionality (scopes and WAV file record and replay) should be moved into the ChannelMaster layer first: that will significantly improve the complexity because it puts almost all the data traffic into one direction. That will have to be the first step.

If it could be achieved, the data rate between client and server parts would be far lower than between the FPGA and PC. And it's only an IP address away from having the two halves on separate continents.. Just... monstrous effort needed.

There is a plan; it has proof of concept and risk reducing stages first, to establish that there would be gold at the end of the rainbow. I have a full time job, so I'm not going to be able to complete this alone for a while. Four years seems quick! There's lots of opportunity for others to join in though....

Re: HTTP Server?

Posted: Wed Jan 20, 2021 1:29 pm
by KA5KKT
laurencebarker wrote:I've worked up an outline design with Doug and Warren for a client/server split in Thetis. It is a monstrous amount of effort - probably a person-year of full time work. Done part time in evenings, potentially several years. There is something like 600 function calls that need to be replaced by IP messages. Warren has recommended that more of the Thetis functionality (scopes and WAV file record and replay) should be moved into the ChannelMaster layer first: that will significantly improve the complexity because it puts almost all the data traffic into one direction. That will have to be the first step.

If it could be achieved, the data rate between client and server parts would be far lower than between the FPGA and PC. And it's only an IP address away from having the two halves on separate continents.. Just... monstrous effort needed.

There is a plan; it has proof of concept and risk reducing stages first, to establish that there would be gold at the end of the rainbow. I have a full time job, so I'm not going to be able to complete this alone for a while. Four years seems quick! There's lots of opportunity for others to join in though....


Thank you for your efforts!

Re: HTTP Server?

Posted: Wed Jan 20, 2021 3:27 pm
by w-u-2-o
laurencebarker wrote:I've worked up an outline design with Doug and Warren for a client/server split in Thetis.
That's great, Laurence!
It is a monstrous amount of effort - probably a person-year of full time work. Done part time in evenings, potentially several years. There is something like 600 function calls that need to be replaced by IP messages.
Only if you do it right ;) I'm no developer, but my system engineering experience suggests an easier (but less elegant, of course) way would be to exploit Thetis itself as the server, as follows:

1. Don't try to use Thetis as the UI. Make a new, fresh UI. That way you avoid all the horrible spaghetti code.
2. Exploit the fact that there are already CAT messages for nearly everything, so merely stand-up an TCP pipe between the client and Thetis-server to carry that traffic from the client UI. You may need to add a few more CAT messages for some things.
3. Now that Ritchie has brought us into the 21st century with DirectX, pick off the spectral display data and move that. You'll have to send window size back to Thetis-server, of course.
4. Replace the VAC code in Thetis-server with some modern library that moves audio over IP (not UDP).

That leaves the really tricky bit, which is dealing with the server side setup menus. It would really suck to have to deal with all those messages. And I suspect it would not be easy to merely yank that code out of the server side and put it in the client side. Perhaps it would be possible to use something like X-windows to serve up the setup window on the client side. Make it an exercise in remote access instead of the real deal.
Warren has recommended that more of the Thetis functionality (scopes and WAV file record and replay) should be moved into the ChannelMaster layer first: that will significantly improve the complexity because it puts almost all the data traffic into one direction.
That is certainly the elegant method. After that it would only need to be a thin layer over ChannelMaster on the server side. It really suggests starting from scratch might be easier given the right changes to ChannelMaster.
There is a plan; it has proof of concept and risk reducing stages first, to establish that there would be gold at the end of the rainbow. I have a full time job, so I'm not going to be able to complete this alone for a while. Four years seems quick! There's lots of opportunity for others to join in though....
I wish I could help in some substantive way. Alas, I am not a coder or developer. If you need a guinea pig I'm your guy, though! I've got plenty of machines laying around here that would be happy to become a radio server.

Re: HTTP Server?

Posted: Wed Jan 20, 2021 8:32 pm
by PD3LK
Thank for explaining and your splendid work Laurance, never expected that this would be so much work.
That would mean that you almost have to rebuild everything from scratch. If you, or someone else, wants to this in the future then do a little marketing first, how many ham's are waiting for this remote possebility over WAN/Internet?

Then consider to stop the further development of Thetis (it's very good as it is) and start a new and seperate server - client(s) based project.
I like that idea, wish i had any programming skils, tried it but my head is a sieve. My only project ever was a (if i may say so, very good) RTTY decoder for the C64 but that was long ago...

Re: HTTP Server?

Posted: Mon Jun 21, 2021 3:18 pm
by iw7dmh
Hi,

it seems that it is more than a plan.
A headless RPI next to any Anan rig would implement the client/server model you talked about before.
It is also the same design used in ExpertSDR Remote System.
I hope it will be released asap (y)

Best 73'
Enzo, iw7dmh

https://www.youtube.com/watch?v=UEgcAghrFtU&t=208s