| sadmac2 | Keybuk: I think I have an idea for the linked list issue | 16:13 |
|---|---|---|
| Keybuk | oh? | 16:16 |
| sadmac2 | Keybuk: for cursor objects, why not set the prev pointer to something weird? | 16:16 |
| sadmac2 | Keybuk: and then have them keep the real prev value in a 3rd element | 16:16 |
| sadmac2 | so in the insert/delete/etc methods if (prev == (void *)0x2) //the real prev is off the end of the structure, and this is a cursor | 16:17 |
| sadmac2 | we could even set it NULL, since I think no valid list structure should ever have null pointers | 16:18 |
| Keybuk | and have add_after/add know to check for it? | 16:22 |
| sadmac2 | those, and possibly a few others | 16:23 |
| sadmac2 | it could touch quite a few functions given how drastic a change it is | 16:23 |
| sadmac2 | or rather how drastic a departure from the list structure it is | 16:23 |
| sadmac2 | but all the information you need to operate on the list is always available | 16:24 |
| sadmac2 | the other thing we need (and we need this for any cursor-based solution) is an NIH_LIST_PREV and NIH_LIST_NEXT set of macros, since just grabbing the pointers is unreliable from within the linking application | 16:26 |
| === notting_ is now known as notting | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!