DarthArgus
8bec0aa8ce
Merge branch 'testing'
2016-12-11 02:30:30 +00:00
DarthArgus
5ee6a5b9de
go back to our original ...
2016-12-11 00:27:23 +00:00
DarthArgus
8e94519082
persister cleanup (still broke but logic is there for when it isn't); also replace all the SOE BS for forwarding stl port with native object classes because i thought that may be part of the problem...
2016-12-10 23:17:06 +00:00
DarthArgus
688ab3fc9e
unfuck the code a little
2016-12-10 04:24:14 +00:00
DarthArgus
76d08c50da
kill me.
2016-12-10 04:16:07 +00:00
DarthArgus
3a62cb8b86
whoops!
2016-12-10 00:52:08 +00:00
DarthArgus
e7080c4283
misnomer
2016-12-10 00:50:30 +00:00
DarthArgus
8d471e8057
add discrepancy tracking for delete counts, with a loop that should clean up orphans
2016-12-10 00:46:54 +00:00
DarthArgus
773ec16c59
mutex removal since they aren't helping
2016-12-09 23:56:57 +00:00
DarthArgus
fd73109f03
this is broken but proves where our race condition comes from
2016-12-09 22:30:25 +00:00
DarthArgus
8834b81283
in theory this is all we need to keep from stepping on our own toes; if we have a delete bug after this we need to look at that freaking map and it's constituents, specifically the NetworkId member
2016-12-09 22:10:54 +00:00
DarthArgus
64eda543f9
in theory this should let things run but warn when we leak
2016-12-09 21:24:22 +00:00
DarthArgus
87443dd17e
prevent death of server if we're waiting on a save to finish but still warn about it; only lock at the top of saveCompleted
2016-12-09 21:02:46 +00:00
DarthArgus
802af495dc
silence output by default in release mode for things we can't do much about on a live server... TODO: @cekis can we find a nice way of silencing or outputting java stack dumps into a file instead of the console?
2016-12-09 21:02:00 +00:00
DarthArgus
bf396ac2a2
add loader mutex; cleanup a couple elseif chains into switches
2016-12-09 15:22:23 +00:00
DarthArgus
324ba63049
Merge branch 'testing'
2016-12-09 04:04:20 +00:00
DarthArgus
6418f430f9
this should let us delete snapshots at the cost of latency
2016-12-09 03:37:21 +00:00
DarthArgus
a353c128c3
fix db bug; cleanup; better destructors
2016-12-08 17:23:03 +00:00
DarthArgus
eab175813d
remove another unused flag
2016-12-08 10:07:51 +00:00
DarthArgus
4a1159ad14
this builds and runs but our objvars are never loaded (or buffered?) properly
2016-12-06 17:45:04 +00:00
DarthArgus
b17393763a
madness
2016-12-06 16:58:56 +00:00
DarthArgus
c8b782b0d1
g3 for debug builds
2016-12-06 03:10:35 +00:00
DarthArgus
c254e498bc
@cekis , you'll want to toggle this to true for debugging, but it just
...
spams in release mode
2016-12-04 01:13:01 +00:00
DarthArgus
cbfc88d54b
cleanup works and reduces our memory footprint, and error messages
2016-12-04 00:56:25 +00:00
DarthArgus
12cf580937
fix all the warnings
2016-12-04 00:56:24 +00:00
DarthArgus
066eac440e
nit picking on unrelated to the leak things...
2016-12-04 00:56:24 +00:00
DarthArgus
9d8b7c0d3a
Support oracle 10.2.0.4 client libs
2016-12-04 00:56:24 +00:00
DarthArgus
f9c0a47a39
finish removing the SOE memory management
2016-12-04 00:56:11 +00:00
DarthArgus
9f9fed4693
memory manager cleanup and removal
2016-12-04 00:55:37 +00:00
DarthArgus
317d755640
Merge branch 'testing'
2016-12-03 08:28:47 +00:00
DarthArgus
e1410dc208
fix all the warnings
2016-12-03 08:28:21 +00:00
DarthArgus
5d8b7fe160
nit picking on unrelated to the leak things...
2016-12-02 23:49:19 +00:00
DarthArgus
c7e825e4a9
this works but we're still not always freeing up memory...ugh
2016-12-02 16:47:17 +00:00
DarthArgus
4638386552
Revert "c++11, love you... although this may end up duplicating some items on"
...
This reverts commit 8a4cd1aeca
2016-12-02 16:24:36 +00:00
DarthArgus
60735cfd96
support oracle 10.2.0.4 client libs, may leak less
2016-12-01 17:57:46 +00:00
DarthArgus
8a4cd1aeca
c++11, love you... although this may end up duplicating some items on
...
the tc... - by removing the constructors and destructors for these
structs and relying on init lists instead, we allow the compiler to
automatically determine how to allocate and deallocate the values and
references
2016-11-30 23:04:10 +00:00
DarthArgus
46dc62461d
Revert "don't pass objectId by reference"
...
This reverts commit 5f3847c4cd
2016-11-30 22:28:26 +00:00
DarthArgus
31bce71c33
this is a better way of nuking saved snapshots
2016-11-30 22:16:33 +00:00
DarthArgus
cd3c1f3fe8
readd the delete - this branch i'm not sure if we need or not, same goes
...
for smart ptr branch
2016-11-29 23:13:28 +00:00
DarthArgus
8ebe434e0b
clean...
2016-11-29 23:01:00 +00:00
DarthArgus
5f3847c4cd
don't pass objectId by reference
2016-11-29 23:00:36 +00:00
DarthArgus
aaab7cd4ef
finish removing the SOE memory management
2016-11-29 04:25:55 +00:00
DarthArgus
e6921be6fe
ok so this runs but is still lacking the deletion logic
2016-11-29 03:52:23 +00:00
DarthArgus
176555370b
more memory madness
2016-11-28 18:54:09 +00:00
DarthArgus
89cd10fbcc
tidies some things up and removes SOE memory manager overriding
2016-11-28 17:07:49 +00:00
DarthArgus
1867ee7cc9
maybe try this?
2016-11-20 07:52:52 -06:00
DarthArgus
23cf1cc584
well according to SO this looks correct https://stackoverflow.com/questions/991335/how-to-erase-delete-pointers-to-objects-stored-in-a-vector
2016-11-20 07:47:19 -06:00
DarthArgus
cceff1dc1a
oops, don't want to invalidate... and yes the completedSnapshot should in theory point to the same snapshot, but for whatever reason deleting it crashes the TC but not my local setup
2016-11-20 07:34:15 -06:00
DarthArgus
b2a5fc55ab
well, not quite
2016-11-20 07:32:49 -06:00
DarthArgus
74ebc3ec43
if this builds i think it will fix the mem issue
2016-11-20 06:25:19 -06:00
DarthArgus