RE: [dev] slock: anti OOM killer - proper priv dropping - etc.

2013-07-28 Thread Robert Schneider
> If Robert provides the patch with ifdef __linux__ specificness, I > might consider adding it. Here's the ifdef'ed patch. -- ziggomatic 1_dontkillme_v3.patch Description: Binary data

RE: [dev] slock: anti OOM killer - proper priv dropping - etc.

2013-07-26 Thread Robert Schneider
>> hence shoul be in an ifdef linux or something. On the other hand the code is just "Does the proc file exist? No? Then skip this." But yes it could be wrapped in #ifdef __linux__ sections. > Or use setpriority() to make it portable. I don't see how scheduler priority has much influence on OOM

RE: [dev] slock: anti OOM killer - proper priv dropping - etc.

2013-07-23 Thread Robert Schneider
> In dontkillme(), wouldn't it be desirable to close the file descriptor? Shit, good catch. [1_dontkillme_v2.patch] -- ziggomatic 1_dontkillme_v2.patch Description: Binary data

[dev] slock: anti OOM killer - proper priv dropping - etc.

2013-07-22 Thread Robert Schneider
0_smallstuff.patch: Fix missing newlines and so on. 1_dontkillme.patch: Disable Linux's out-of-memory killer for slock if the system has the configuration knob. That douche hunts small game! Oh how super fun to come back to an unlocked desktop. 2_dropprivs.patch: Only try dropping privileges if p