Open Source Realtime Strategy Game Engine
https://springrts.com/phpbb/
bugged very descriptive of you.AntiAllez wrote:laggy and bugged
Any mod that was ported the rest stubbornly stayed in the old spring. Eh it doesn't matter, players like you only count something that has players all the time.AntiAllez wrote:seriously, which mods working on it? BA TechA ... the rest is still on 91, 94 or awaiting 96
a very common phrase of your species in this communitysmoth wrote: Eh it doesn't matter, players like you only count something that has players all the time.
You might even say that Spring 96 is also Spring 94.1 based!jamerlan wrote:Sad but r-t-s-m-p is spring 94.1 based.Jools wrote:You should benchmark vs spring 95 or 96...
Yes, this is what I meant. It's just a number, but since most people now play spring 95 and soon 96, what's the point of benchmarking against an old version?luckywaldo7 wrote:You might even say that Spring 96 is also Spring 94.1 based!jamerlan wrote:Sad but r-t-s-m-p is spring 94.1 based.Jools wrote:You should benchmark vs spring 95 or 96...
Why you love to bend your own rules? that makes me sadsmoth wrote:ok, I guess we are done.
Code: Select all
Major:
- kill memory leaks crippling large games (#4152)
- kill micro-jitter in simulation framerate (#4171)
- kill unclean exits causing various forms of corruption (#4053)
- "nugget" is hereby officially promoted as the new umbrella term for widget or gadget
Misc:
- default disable grass for ATI cards (crashes for some people)
- made stacktrace-translation script recognize more crash/hang patterns
- improve Shard AI for BA
- fix some MSVC compile errors
- TimeProfiler (alt+b) shows now max `lag` of each item
- massively speedup caching with many .sdd archives
- fixed some FreeBSD compile issues (thx AMDmi3)
- dump (engine) shaders to infolog when they fail to compile
- more extensive logging during shutdown
- fix a number of potential issues if a thread crashes on engine shutdown
- increase precision of several timer variables to preserve intermediate results
- Lua: track more memory-allocater statistics for display in debug-mode
- Lua: limit maximum amount of memory allocated globally and per handle
- GameServer: always echo back client sync-responses every 60 frames (see #4140)
- GameServer: removed code that blocks pause / speed change commands from players with high CPU-use in median speedctrl policy
- GameServer: sleep less between updates so it does not risk falling behind client message consumption rate
- LuaHandle: make time and steps spent on garbage collection configurable
CollectGarbage callin is restricted to run for at most
MaxLuaGarbageCollectionTime (user-configurable) millisecs
MaxLuaGarbageCollectionSteps (user-configurable) iterations
- GlobalConfig: add new config-option 'UseNetMessageSmoothingBuffer' (see #4053)
- LogOutput: make enabled 'section' log-levels configurable by users
- MoveDef: add key 'allowTerrainCollisions' for handling units with super-sized footprints (see #4217)
- SolidObject: make {Unit,Feature}{Pre}Damaged events receive the 'attacker' ID when object is crushed
- QuadField: add raytraced projectiles to three cells instead of one
- GameInfo: add map hardness label/value to 'i' overlay
Sim:
- Rifle weapons can hit features too now and apply impulse to the target
Lua:
- add gadgetHandler:AllowWeaponInterceptTarget
- LuaSyncedRead:
make Get{Unit,Factory}Commands (optionally) return just the queue's size
(ie. without copying the entire command queue into Lua VM memory, *VERY*
important for performance when a nugget is not interested in the commands
themselves)
"if #Spring.GetCommandQueue(unitID, -1 | nil ) == 0 then return end" --> BAD, CREATES COPY OF QUEUE
"if Spring.GetCommandQueue(unitID, -1 | nil, false) == 0 then return end" --> GOOD, DO THIS IF POSSIBLE
Bugfixes:
- fix #4215 (stack-buildable units on non-flat terrain)
- fix #4209 (antinukes firing at nukes targeted outside their area)
- fix #4205 (air-transports unable to unload hovercraft on water)
- fix #4204 (float3.* clang compilation failure)
- fix #4203 (air-transport unable to load units floating on water)
- fix #4202 (S3O textures loaded multiple times instead of being cached)
- fix #4200 (secret)
- fix #4198 (amphibious units without waterweapons chasing hovercraft whilst underwater)
- fix #4197 (units moving right up to their target instead of attacking from range)
- fix #4186 (log-system crash on exit)
- fix #4173 ("Error: [Watchdog::ClearTimer] Invalid thread N")
- fix #4171 (see 'Major' section)
- fix #4163 (corrupted unit environment reflections)
- fix #4161 (Spring.Restart failing when path contains spaces)
- fix #4155 (reverse-built gunships maintain their vertical speed)
- fix #4154 (calling LUS 'AttachUnit(piece, unitID)' turns following 'Move(piece, ...)' calls into no-ops)
- fix #4152 (see 'Major' section)
- fix #4149 (rendering corruption in DynamicWater mode)
- fix #4144 (partly submerged units not hittable by non-waterweapons)
- fix #4145 (depthcharges fired by ships having trouble to reach their targets)
- fix #4138 (same as #4171)
- fix #4137 (units randomly switching team-color, old bug with new cause)
- fix #4135 (ATI driver crash with *disabled* shadows)
- fix #4133 (cloaked units hidden even when set to alwaysVisible)
- fix #4131 (units moving to position of manualfire command if outside weapon range)
- fix #4129 (air-constructors flailing around construction targets)
- fix #4128 (units stuck in 'skidding' state: repeatedly bouncing on ground while unable to move)
- fix #4091 (gadget:MousePress() and other unsynced gadget functions not called with LoadingMT=0)
- fix #4053 (see 'Major' section)
- fix #3749 ('initCloaked' UnitDef tag overriding user cloak commands for being-built units)
- fixed wreckage heaps could be resurrected
- LaserCannon: (once again) fix range overshooting
- always clamp length of blended wind-vector to [map.minwind, map.maxwind]
- the 'AllowDeferred*Rendering' config options are now safe to enableInstead of bugs, you listed bug-fixes and then your "proof" of bugs was the insinuation thatAntiAllez wrote:Im pretty sure you noticed that list -that just showing that things what they/you believe to have solved. It is currently a horror with ATI/AMD.