| 27 Jul 2024 |
|  chris_tannery joined the room. | 19:59:10 |
| 28 Jul 2024 |
|  elit007 joined the room. | 17:34:41 |
 M0VUB - Shane | 
Download IMG_8898.jpg | 20:17:44 |
| M0VUB - Shane | Compiler error on
Debian bullseye (raspian) | 20:17:47 |
| M0VUB - Shane | I’ll post the errors on the mrefd compile too in a min | 20:18:03 |
 kc1awv | I can't remember if the OpenDHT compile error is an issue with the version of the compiler or not. But the error itself is specific to the source code. | 20:18:54 |
| kc1awv | I think if you're using a recent Ubuntu version, the OpenDHT package available in apt is sufficient | 20:19:26 |
| M0VUB - Shane | I’ll do some more digging | 20:20:15 |
| M0VUB - Shane | Well! We are in business with mrefd latest compile. Running two instances of M172DMR for two DMR talkgroups on two modules. M17-DMR reflector is back online… couldn’t get the OpenDHT to compile though unfortunately, | 21:06:56 |
| 29 Jul 2024 |
 tom.early | The OpenDHT library with its dependencies can be installed on Debian 12 (bookworm) and Ubuntu 24.04 (Noble Numbat) without compiling: sudo apt install libopendht-dev | 03:41:00 |
 G4KLX | Which is why the C++ I write tends to be conservative and not use the latest features. | 07:37:56 |
| tom.early | And that's why were still using host files.
I'd love to have a couple of beers with you some time and discuss your reluctance to C++11 or, for that matter C++14. Both are over a decade old. | 15:56:44 |
| tom.early | * And that's why we're still using host files.
I'd love to have a couple of beers with you some time and discuss your reluctance to C++11 or, for that matter C++14. Both are over a decade old. | 16:06:50 |
| G4KLX | Yes, and you should try supporting software on multiple platforms. Modern C++ is a nightmare. What do you mean about host files? | 20:20:39 |
| tom.early | Client software should not be getting connection parameters from a static file of questionable age and accuracy. They should be getting that info directly from the desired target. | 22:20:57 |
| kc1awv | So, I agree on that. But the question is, how do the clients get the initial information for the target? I can understand the idea of bootstrap servers, but why not just have the bootstrap servers contain the information for all the targets anyway? And then we come full circle, I suppose | 22:23:37 |
| tom.early | List go stale. An M17 reflector changes IP and the sponsor forgets to update his m17refl.json info. | 22:35:26 |
| tom.early | * I'm not sure I know what a "bootstrap server" is and google is not helping much, but where would a bootstrap server get it's info on available targets?
Lists go stale. An M17 reflector changes IP and the sponsor forgets to update his m17refl.json info. | 22:49:34 |
| tom.early | Have you ever had to support computers with a 20-bit word size? I'm well aware of the problems. | 22:56:53 |
|  @thompson_1:matrix.org left the room. | 23:01:58 |
| 30 Jul 2024 |
| kc1awv | Yeah, we’ve got a chicken/egg situation. I’m thinking of adding in a requirement that records must be updated after a certain set of time, or they become unlisted. They won’t be deleted, just unlisted with a checkbox to click to relist it if it times out. | 00:38:15 |
| kc1awv | Unless we can come up with a solution like Luc did for xlxd | 00:38:35 |
 Tony VK3JED | That would be a PITA for me. Even more risky than a stale IP. 😄 | 03:59:42 |
| G4KLX | The fact that my use of host files is queried shows that you don’t understand the reasons which were set out at the time of their creation. | 06:45:36 |
| G4KLX | The 20-bit CPU is a red herring. Even within a particular architecture there are many different ages of compilers and modern C++ isn’t universally supported to the same levels. It’s a pain, so by keeping it simple makes my life easier. | 06:47:39 |
| G4KLX | There are over 100,000 MMDVM deployments, so simplicity and reproducibility is key, not to mention reliability. I’ve worked in safety critical development where simple and robust is encouraged. | 06:49:18 |
| G4KLX | BTW why was DHT used when we have the perfectly usable ircDDB available, doing its job since 2010. | 06:51:26 |
|  fritsflapears joined the room. | 13:01:04 |
| tom.early | I am sorry about making an off-topic response about an off-topic comment. I'll stop this tread.
I would urge mrefd owners to support DHT if possible. I have already shown that a DHT-enabled client can find an mrefd instance even if it is not registered on the M17 Project website. You don't need to know the IP address, you just need the reflector name. Yes, there is only one client that uses DHT, but I think there will be another.
DHT is also very useful to you as an administrator especially if your reflector is interlinked with other mrefd reflectors, or, want to join an existing group.The technology works and it is very easy to include in your reflector build, especially if you are on Debian 12 or Ubuntu 24.04. | 20:01:43 |
| 31 Jul 2024 |
|  maxg7uoz joined the room. | 14:25:58 |
