| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
|
|
|
| |
It wasn't all that easy.
|
|
|
|
|
|
|
|
|
|
| |
On my own machine (4096 max FD's), scanning localhost for all services,
there was no change to speak of. On another one, though, with 65536
max FD's scanning localhost and another machine simultaneously for
a specific service, there was a dramatic speedup allowing much faster
scans, with userspace load dropping to about a quarter of a second.
As a bonus, it makes the code simpler.
|
| |
|
|
|
|
| |
Setting the fd to -1 broke poller_remove_from_dispatch().
|
| |
|
|
|
|
| |
Valgrind seems to be quite happy with it now.
|
| |
|
| |
|
|
|
|
| |
No idea how this happened.
|
|
|
|
| |
`void *' is more appropriate here.
|
|
|
|
| |
help2man should be enough so far.
|
|
|
|
| |
Fucking hell. We're still having some event scheduling issues.
|
| |
|
|
|
|
|
|
|
|
| |
It's about time we stopped fucking ourselves in the butt.
The scanning should be much faster now.
Thanks to libuv for inspiration.
|
|
|
|
| |
There are still some problems but at least it does something now.
|
| |
|
|
|
|
|
|
|
| |
- renamed *_func to *_fn
- some initial code for the indicator (needs curses)
- moved option handler to utils
- more work on unit generation & processing
|
| |
|
|
|
|
| |
Make them compile.
|
|
|
|
| |
Don't worry, there will be better commit messages eventually.
|
|
Nothing's working yet, only some supporting code has been put in place.
|