|
Revision 794, 1.3 KB
(checked in by shawnw, 21 months ago)
|
|
Assorted NetBSD and general fixes
|
| Line | |
|---|
| 1 | Bugs that aren't our fault, but might bite people: |
|---|
| 2 | |
|---|
| 3 | Modern: |
|---|
| 4 | |
|---|
| 5 | * You might experience crashes because the default executable stack |
|---|
| 6 | size is too small. An 8 megabyte stack is more than ample. On unix, it |
|---|
| 7 | can be raised with 'ulimit -s 8192'. On Windows, you must use the editbin |
|---|
| 8 | progam that comes with MS development environments to raise the stack on a |
|---|
| 9 | per-executable basis, via 'editbin /STACK:8388608 netmush.exe' |
|---|
| 10 | |
|---|
| 11 | * Some IPv6-enabled systems are not configured out of the box to tunnel |
|---|
| 12 | IPv4 connection attempts on the same port to the IPv6 socket that is |
|---|
| 13 | listening. If you can connect to a local game via ::1 but not |
|---|
| 14 | 127.0.0.1 this is likely the issue. On unix systems, a user with |
|---|
| 15 | root privileges can do 'sysctl -w net.inet6.ip6.v6only=0', after |
|---|
| 16 | which you'll have to do a complete shutdown and restart of the mush |
|---|
| 17 | for the change to take effect. The file /etc/sysctl.conf will have to |
|---|
| 18 | be changed too so the change persists across reboots. |
|---|
| 19 | |
|---|
| 20 | |
|---|
| 21 | Ancient: |
|---|
| 22 | * Ralph Melton reports that compiling with gcc 2.5.8 under SunOS 4.1.1 |
|---|
| 23 | using -O optimization and forking dumps causes the dump process to |
|---|
| 24 | crash. Removing -O fixes the problem; so might using a more recent gcc. |
|---|
| 25 | * Javelin has confirmed that compiling with gcc 2.7.2.3 under Linux 2.2.16 |
|---|
| 26 | using -O optimization causes ansi(rh,a) to crash. Removing -O fixes |
|---|
| 27 | the problem; so might using a more recent gcc |
|---|
