Spring 0.80.2 (MAKE IT HAPPEN) released

[lurker]

Yes, let's make a petition for something that was never going to happen.

If you have that much time on your hands why not fix the installer's ability to get tasclient correctly.

[Jazcash]

I actually thought about swapping it with qtlobby in the installer. But I didn't do it because I could somehow see the shitstorm coming, so I just left it as it was.

its like chopping a leg from spring..

Yes, an infected leg that needs to be amputated before it spreads to the rest of the body.

[Hacked]

thx for the link jaz

[albator]

Hi Guys
it didnt (MAKE IT HAPPEN) for me
afther installing the new 0.80.2
i get: GetluaAIName cant be fond in DLL-bestand UnitSync.dll.
Hope some one can help me ?!

Thanks

Greet
Bio_

I had the same problem. I solved i by :

1) unrar http://springrts.com/dl/spring_0.80.2.7z

2) unrar http://tasclient.licho.eu/TASClientLatest.7z


I think it is because I use an old Xp btw

[TradeMark]

unrar? its 7zip not rar... un7zip!

[Pendrokar]

O M G

TASClient works!!! (the latest one)

Did it always take 250 MB RAM when making minimaps. Behh I dont care I've got 3.7 more

[albator]

unrar? its 7zip not rar... un7zip!

u understand me... unrar read 7zip

[TradeMark]

is there something wrong with this new spring-dedicated.exe ? its crashed 4 times on one day since new release... before that running nicely for a week.

Edit: crash++

[Auswaschbar]

is there something wrong with this new spring-dedicated.exe ? its crashed 4 times on one day since new release... before that running nicely for a week.

Edit: crash++

Update please.

[reivanen]

Besides the irrational random folder creation and "improved" hardcoded archivemover behavior, there seems to be still one annoying "feature" since last spring update.

What i am talking about is a delay between giving the launch command to spring.exe to where it actually begins to load.

before the update the loading started immediately when launching a game or starting spring.exe directly, now with 0,80.2 (havent tried previous 0,8x versions) it takes about 4 seconds of nothing before the exe starts to load and do its usual business. What could cause such a thing? NO version before the current has displayed similar behaviour.

[CarRepairer]

Besides the irrational random folder creation and "improved" hardcoded archivemover behavior, there seems to be still one annoying "feature" since last spring update.

What i am talking about is a delay between giving the launch command to spring.exe to where it actually begins to load.

before the update the loading started immediately when launching a game or starting spring.exe directly, now with 0,80.2 (havent tried previous 0,8x versions) it takes about 4 seconds of nothing before the exe starts to load and do its usual business. What could cause such a thing? NO version before the current has displayed similar behaviour.

There was a time when this happened to me with tasclient. It was much worse though, about 30-60 seconds. In other words, long enough for the host to assume I'm not going to connect. I had to switch to Springlobby at that time.

[lurker]

folder creation

bug

hardcoded archivemover

it's always been hardcoded, and installing to program files is bad

delay

weird

[TradeMark]

heh, i never ever even tried archivemover, i knew it will just fail -_-

[Ionic]

Some compilation errors on Lenny/Testing using scons + gcc4.3

SConfigure linking error
7zip library has been linked twice that causes compiler linking error.
Solution - comment out the second linking:
unitsync_7zip_files = usync_get_source('rts/lib/7z'); #usync_get_source('rts/lib/7z/Archive/7z') +

System/float3.h assertion errors
Unsafe normalization causes call to assertion and SIGABORT during game-play.
Solution commenting out unsafe normalization to disable __SUPPORT_SNAN__:
inline float3& ANormalize() {
//#if defined(__SUPPORT_SNAN__)
// assert(SqLength() > NORMALIZE_EPS);
// return UnsafeANormalize();
//#else
return SafeANormalize();
//#endif
}

inline float3& Normalize() {
//#if defined(__SUPPORT_SNAN__)
// assert(SqLength() > NORMALIZE_EPS);
// return UnsafeNormalize();
//#else
return SafeNormalize();
//#endif
}
IMHO I do not see how unsafe version would be faster here because checking is done in both cases.

lua/lmathlib.cpp compile error in SSE mode.
Namespace error during compilation can be solved by commenting out namespace prefixes 'math::'
e.g.
static int math_abs (lua_State *L) {
lua_pushnumber(L, /*math::*/fabs(luaL_checknumber(L, 1)));
return 1;
}

RAI/Builder.cpp
Just starting RAI bot causes process dump. It is caused by cb->GetMetalUsage() in line
if( cb->GetEnergy()*(cb->GetEnergyIncome()/cb->GetEnergyUsage()) < cb->GetMetal()*(cb->GetMetalIncome()/cb->GetMetalUsage()) )
Property cb->GetMetalUsage() has sometimes (at game beginning) 0 which causes simply division error and crash offline game.

Summarizing - still messy is 0.80.2 version. I had never such problems earlier (previous 0.79.x versions).
Please prepare source package in usable form before release. Some have to compile source code and like more optimized version. Now it makes me (almost) impossible to join lobby with such code. I do not like to play on-line through gdb.

[imbaczek]

use cmake, scons has been partially deprecated.

RAI crash is a result of signaling nans which are disabled by default in cmake, because they are meant to be a debugging help. you should've fixed the build options (somewhere in rts/build/scons) instead of patching the source. signaling nans can also crash your gfx drivers, so beware.

[Ionic]

Ok. I will try compile code with cmake.

Two building systems (with scons seemingly not working) are a bit redundant. Maybe it is time to remove scons at all from git?

[imbaczek]

it would've happened if make -j wasn't broken on windows.

[reivanen]

Was archivemover always hardcoded? Weird since it moved the maps and mods to my then non-default spring install dir users/username/home/spring

And about the spring launch delay, contrary to my last post, it actually takes about 8 seconds from when i click spring.exe / a multiplayer game is launched to when the normal loading starts. During the waiting time (ie. immediately after launching) the spring.exe process starts, it uses 50% cpu (dualcore) and about 15-16 megs ram.

edit:

actually the same kind of delay happens with springsettings also and springlobby so it might be something related to unitsync? or something.

edit2:

"initializing unitsync" stage takes also alot longer now with the .51 version of tasclient and 0,8x spring then what it used to take before when tasclient worked out-of-the-box (pre 0,79)

[DoR]

I need to redownload spring but I can't find a link at
http://springrts.com/wiki/Download
for the 0.80.2 version of spring for windows. Can anyone tell me where I should look:?:

[hoijui]

it is there; spring is the engine.
or search for the first "0.80.2" on that page