View Issue Details
| ID | Project | Category | View Status | Date Submitted | Last Update |
|---|---|---|---|---|---|
| 0000446 | LDMud | Implementation | public | 2006-02-28 19:57 | 2021-04-15 17:21 |
| Reporter | Assigned To | fufu | |||
| Priority | normal | Severity | major | Reproducibility | random |
| Status | closed | Resolution | fixed | ||
| Summary | 0000446: Socket implementation is not binary proof | ||||
| Description | If during a charmode read it happens that the firs character in the buffer is 0x00, the driver gets stuck there because the input handling functions still consider the buffer a C-String. It needs to be changed that it acts more like a m-string. This plays into the general theme that the socket implementation is rather convoluted and probably should be re-implemented like Genesis does it. | ||||
| Tags | No tags attached. | ||||
| External Data (URL) | |||||
|
|
I guess, this is one more reason to use something like the libtelnet mentioned in 0000810 in the long run. |
|
|
I think this is fixed in 3.6.3 (commit: b24832bb671 "Added interactive configuration option for the encoding."), \0 character arrive in the input_to callback now. In my tests, 3.5.3 and also 3.3.720 skip over \0 characters, but the driver does not get stuck anymore. |
| Date Modified | Username | Field | Change |
|---|---|---|---|
| 2006-02-28 19:57 |
|
New Issue | |
| 2012-12-06 19:17 | zesstra | Note Added: 0002170 | |
| 2012-12-06 19:17 | zesstra | Relationship added | related to 0000810 |
| 2021-04-15 17:19 | fufu | Note Added: 0002585 | |
| 2021-04-15 17:21 | fufu | Assigned To | => fufu |
| 2021-04-15 17:21 | fufu | Status | new => closed |
| 2021-04-15 17:21 | fufu | Resolution | open => fixed |
