Commit Graph

453 Commits

Author SHA1 Message Date
D3athrow
196aa9c366 keep rogue process from hanging scheduler. 2015-08-16 17:13:26 -05:00
D3athrow
c3a9905811 Like wizards n shit, zap mothafucka 2015-08-15 22:45:37 -05:00
PJB3005
9d7ff729b6 Atom locking 2015-08-15 15:06:04 +02:00
PJB3005
9aaa3dd038 Ports Mloc's latest optimizations 2015-08-14 01:52:30 +00:00
Unfit
8e63fec2a8 remove sillyness
reject_bad_name( replaces sanitize(stripped_input(
plating added to the list of bad names
2015-08-13 05:45:18 +00:00
unid15
6c4d0add75 SHUTTleS 2015-08-06 16:59:13 +00:00
PJB3005
3bc5d86263 LIGHTING IS BACK.
AND A SHITLOAD FASTER.

SANIC.
2015-07-21 14:16:27 +02:00
D3athrow
cf333c09e7 FFUUUUUUUUUUUUUUUUUUUUUUUUUUUUCCCKKK 2015-07-20 16:04:58 -05:00
D3athrow
d59f6a07b0 comment out panicwriter and saytesting 2015-07-20 16:04:48 -05:00
D3athrow
f545aed710 fix dem crashums 2015-07-19 11:22:55 -05:00
D3athrow
7a3ff4efc3 THE LIGHTING IS DEAD, LONG LIVE DAL 2015-07-18 23:05:18 -05:00
D3athrow
41d7df01fd just fuck my shit up 2015-07-18 02:30:51 -05:00
d3athrow
471efb29a1 Merge pull request #5281 from PJB3005/just_what
Removes setLoc(), a retarded piece of code
2015-07-15 11:14:21 -05:00
clusterfack
523e1eaf24 Merge pull request #5342 from clusterfack/FUCKYOUSAYLAGGYSAYCODE
SHOVE IT IN ONII-SAMA
2015-07-14 18:28:42 -05:00
clusterfack
53b4182eef Optimizes Say
I dunno make say code faster or something
2015-07-14 16:20:52 -05:00
clusterfack
e4084756dc Add midnight rollover to station time
Fixes #5329
2015-07-13 14:27:10 -05:00
PJB3005
decf441d75 Why was N3X15 one of our head coders again? 2015-07-07 15:58:10 +02:00
clusterfack
82acb92f1a Runtimes
Fix two old as shit runtimes, one that occurs on every local server involving making a random name. The other coming from my recent removal of call() from client/move, istype is being called upon the object instead of the variable which isn't blocking other types.
2015-07-05 21:09:25 -05:00
d3athrow
32bd239298 Merge pull request #5226 from clusterfack/thingy
Various small optimizations
2015-07-05 12:01:52 -05:00
clusterfack
ee1d271894 Various small optimizations
Makes mobs gibbed or dusted not force del and also makes the atom/movable/overlays they use for their death animation not force del.

Also removes the call() from client/move because WHY THE FUCK was it there.
2015-07-05 02:06:19 -05:00
D3athrow
5f2992c55f warblgarbl 2015-07-04 16:49:07 -05:00
D3athrow
257b09d586 FUCKing runtimes 2015-07-03 19:01:34 -05:00
dylanstrategie
7dd39a9c52 Port experimental base_turf system from Baystation - Third time in a row
No, I don't know how to manually fix conflicts on a PR, and this doesn't
make me wanna figure out how it's done

Fuck you too for not even trying when you were working on Endgame @N3X15

Also thanks for fucking up my PR by stealth-merging a fix @Comic

Because Baystation has finally gotten around to it, and because the code
looked simple, I decided to get my hands on porting a system that allows
up to define base_turfs. As a bonus, it also defines them per Z-level.
So if you wanted, you could have a space station, and a Z-level on a
map-spanning asteroid, and then an away mission on land, all in the same
file

Note that you can modify it mid-round via a DEBUG flag verb, it will
however only affect destroyed turfs (it won't change space, which I
suppose is intentional)

Tested were destroying walls and floors and even dropping a bomb in the
middle of Medbay

Not tested were more finicky conditions like the Singularity or
buildmode. I simply replaced all ChangeTurf() instructions that pointed
to space, because that is how Baystation implemented it

Only problem with this is that it doesn't work on a per-map basis, so
unless some wizardry is done on that code, it'll be hard to get anything
interesting going with it

Discussions go below. Since this is pure code refactor, I doubt there'll
be much up to debate
- Old PR over at #5192
- Older PR over at #5167

Example picture (with the current code, there would be no lattices over
the sand) :

![dune](https://cloud.githubusercontent.com/assets/6137403/8409264/9d1b0e5e-1e76-11e5-9a53-13a972268c87.png)
2015-07-02 01:30:15 +02:00
clusterfack
1a177a26b3 Fix resetvariables runtime 2015-06-29 11:37:50 -05:00
clusterfack
03383f1fc0 Merge pull request #5155 from clusterfack/dpooldatums
SINGULARITY STOP LAGGING
2015-06-29 05:44:06 -05:00
clusterfack
cb58c5bf8d Merge pull request #5156 from PJB3005/bluespess
Fixes #4787
2015-06-28 07:10:46 -05:00
PJB3005
a9a979e3e6 Fixes #4787 2015-06-27 18:25:14 +02:00
clusterfack
d8035c7820 Fixed Optimization
Makes material datums not lag by not calling del() when a datum pool is overrun, instead just cutting the oldest one loose from the pool.

Made the pull from earlier called 'optimizations' work properly, this method is almost identical except it works fully and is only minimally less efficient.

Instead of looping through the exclusion lists every time, it only loops through the variables. Instead of calling initial every time, it merely copies a value from an associative list.
2015-06-27 07:09:16 -05:00
PJB3005
5c75d4b51f Fixes #2182 2015-06-27 12:51:11 +02:00
D3athrow
f71d09ba60 Merge pull request #5006 from clusterfack/optimizemin
Optimizations (reverted from commit 4075a9a855)
2015-06-17 18:29:53 -05:00
d3athrow
4075a9a855 Merge pull request #5006 from clusterfack/optimizemin
Optimizations
2015-06-17 17:33:31 -05:00
clusterfack
b17f96c2c6 Optimizations
Optimizes reset variables, initial costs about 1e-5 to 1e-6. Whereas an operation that just sets a variable in an associate list using another variable in an associative list has a negligible cost even at extremely high calls. Therefore all the variables are initialized only once in a single associative type list holding all the vars.

Optimizes recursive_hear_check as used within say() code. Currently it checks all atom/movables within view which is a minimum of 225 lighting overlays, and checks each of these atom/movables recursively and then uses the list OR operator to add them in due to possible overlap. Instead recursively checks turfs only once and adds them to the found list due to eliminating the possibility of overlap.
2015-06-16 23:50:39 -05:00
D3athrow
cea4e4e00c break everything again, scheck and other loop sleeps now sleep for a proper 10th of a second instead of just 1 tick at whatever FPS. Throw some scheck like calls in the zas ticker. 2015-06-15 15:13:24 -05:00
D3athrow
eac5dd185e grand theft goofball 2015-06-14 22:14:58 -05:00
D3athrow
6f8f1ff08f 110% plagiarism 2015-06-13 20:18:43 -05:00
PJB3005
fa77d9f2fc GOTTA GO FAST! 2015-06-11 12:56:12 +02:00
PJB3005
9eaa5b6c76 Lighting optimizations.
Less lists, more optimized for() loops, more turf ref usage, better turf ref handling when the resolution is 1, pooling for lighting overlays
2015-06-07 22:57:11 +02:00
clusterfack
7f3c0f88b1 Merge pull request #4875 from PJB3005/mlocbest
Pomf why did you merge that already.
2015-06-07 14:26:58 -05:00
PJB3005
fa8ef10978 Pomf why did you merge that already 2015-06-07 15:46:54 +02:00
D3athrow
99a6984c68 sanity 2015-06-06 23:27:30 -05:00
PJB3005
ae7ab25539 BAY LIGHTING PEOPLE 2015-06-06 22:15:21 +02:00
D3athrow
67efd887bd Implement moderately cheaper get_turf thanks to lummox, http://www.byond.com/forum/?post=1860571#comment15256505
Comparison: https://a.pomf.se/faikjo.txt
2015-05-28 16:03:38 -05:00
DeityLink
8f9d6103d9 snooowwwwflakes 2015-05-23 20:51:14 +02:00
d3athrow
3879aadbfc Merge pull request #4573 from Exxion/fireaxe
Fireaxe cabinet construction fixes
2015-05-22 12:08:44 -05:00
Exxion
b139d45cc7 Yep, it worked 2015-05-21 08:30:23 -04:00
Exxion
672d991828 Finally 2015-05-20 21:10:40 -04:00
gbasood
d9cf2d8c8b ASSERT 2015-05-20 10:13:42 -07:00
gbasood
2af1da03c8 Fixes runtime on returnToPool being given a null argument. 2015-05-20 09:14:49 -07:00
d3athrow
dd4395863c Merge pull request #4511 from PJB3005/EXACTLY0words
Removes the most USELESS thing ever.
2015-05-18 17:37:25 -05:00
d3athrow
4afb537d27 Merge pull request #4116 from wwjnc/oh-god-help-me
More feature freeze fixes please kill me
2015-05-17 14:20:48 -05:00