| sewardj | de4a1d0 | 2002-03-22 01:27:54 +0000 | [diff] [blame] | 1 | |
| 2 | Doesn't run |
| 3 | ~~~~~~~~~~~ |
| 4 | Anything with the __NR_clone system call. No idea what to do about |
| 5 | threads yet. |
| 6 | |
| 7 | netscape bombs shortly after successful normal startup. |
| 8 | (???) |
| 9 | |
| 10 | |
| 11 | ToDo |
| 12 | ~~~~ |
| 13 | Rewrite memory manager, with variable sized red zones. |
| 14 | |
| 15 | |
| 16 | Desirable |
| 17 | ~~~~~~~~~ |
| 18 | Demangle C++ names. |
| 19 | |
| 20 | Stack: make return address into NoAccess ? |
| 21 | |
| 22 | Give client programs a way to set memory range permissions. |
| 23 | |
| 24 | Have an --allow-scummy-symbols option, to allow in some of the symbols |
| 25 | currently ignored in vg_symtab2.c. |
| 26 | |
| 27 | Figure out whether cwq (convert signed-32 to signed-64) as |
| 28 | translated into ucode gives correct check semantics. |
| 29 | (I think it does.) |
| 30 | |
| 31 | Check __NR_select; I don't really understand it. |
| 32 | For that matter, __NR_newselect looks wrong too. |
| 33 | |
| 34 | |