View Issue Details
| ID | Project | Category | View Status | Date Submitted | Last Update |
|---|---|---|---|---|---|
| 0000425 | LDMud 3.3 | Networking | public | 2005-12-13 11:55 | 2022-10-06 19:22 |
| Reporter | lynx | Assigned To | Gnomi | ||
| Priority | normal | Severity | minor | Reproducibility | always |
| Status | closed | Resolution | unable to reproduce | ||
| Product Version | 3.3.712 | ||||
| Summary | 0000425: support for dyndns'ed machines | ||||
| Description | it has worked until approx 620 that i could pass my dyndns name to the driver and it would figure out its ip number itself, but it no longer works since "we" added the patch for multihomed bind, which is actually useful in a server situation. so in my situation the driver would bail out from connecting with a "bind during net_connect" perror. i'm not sure if this is within the plan but for myself i solved the issue by putting an if around the bind ensuring if the user actually did pass an ip address. if not, then the bind is not expected anyway. the result looks like this (make it nicer if != 0 is incorrect): /* On multihomed machines it is important to bind the socket to * the proper IP address. */ if (host_ip_addr_template.sin_addr.s_addr != 0) { ret = bind(d, (struct sockaddr *) &host_ip_addr_template, sizeof(host_ip_addr_template)); if (ret == -1) { perror("bind during net_connect"); rc = errno; break; } } i hope this is useful, or you have a different masterplan on how to handle this. | ||||
| Tags | No tags attached. | ||||
|
|
This change has proven to be a good one in the last 3 years. And some people may just have all reasons to run a MUD behind dynamic DNS. |
|
|
I'll look at this. I think that the idea is sound, but I hope there's a better check than comparing s_addr with 0. |
|
|
(I'm relinquishing this because I have no longer an opinion on what the right behavior is. Sorry for dropping the ball 12 years ago...) |
|
|
I don't understand why bind fails with given an INADDR_ANY address (it might be unnecessary to bind on all addresses, but shouldn't fail, as there are usually free ports on all addresses). So I would like to know the error message that perror() gave. I'm closing this bug, as it is really old and I'm not expecting an answer. But if this problem persists, please reopen it and provide more information. |
| Date Modified | Username | Field | Change |
|---|---|---|---|
| 2005-12-13 11:55 | lynx | New Issue | |
| 2009-05-22 03:16 | lynx | Note Added: 0001133 | |
| 2009-05-23 15:29 | fufu | Relationship added | related to 0000616 |
| 2009-05-23 15:31 | fufu | Note Added: 0001147 | |
| 2009-05-23 15:31 | fufu | Assigned To | => fufu |
| 2009-05-23 15:31 | fufu | Status | new => assigned |
| 2021-04-08 22:29 | fufu | Assigned To | fufu => |
| 2021-04-08 22:29 | fufu | Status | assigned => new |
| 2021-04-08 22:32 | fufu | Note Added: 0002576 | |
| 2021-04-16 17:27 | zesstra | Relationship added | related to 0000826 |
| 2022-10-06 19:22 | Gnomi | Assigned To | => Gnomi |
| 2022-10-06 19:22 | Gnomi | Status | new => closed |
| 2022-10-06 19:22 | Gnomi | Resolution | open => unable to reproduce |
| 2022-10-06 19:22 | Gnomi | Note Added: 0002691 |
