Updates 1.24.41 and 1.24.5854 are live

ForumNinjaForumNinja Key Master, Head Chef, Member, PRO Posts: 554
edited January 2016 in GameSalad Kitchen

Mac Creator 1.24.41 Fixes:

  • General improvements

Windows Creator 1.24.5854 Fixes:

  • General improvements

Engine 1.24 Fixes:

  • Fixed the issue with center of mass and collisions
  • Updated AdMob (7.6.0), RevMob (9.0.6), Chartboost (6.1.0) sdk's
  • Slight rework done on how the engine handles time (this should fix some of the timing/animation issues that have cropped up)
  • Some sweet load time / spawning optimizations as well! :)

Comments

  • Braydon_SFXBraydon_SFX Member, Sous Chef, Bowlboy Sidekick Posts: 9,273

    Nice work team!

  • AlchimiaStudiosAlchimiaStudios Member Posts: 1,069

    Initial tests looking good for time/animation fixes in preview. Gonna do some device tests tonight. Thanks GS team! Looking forward to 1.25 now o:)

    Follow us: Twitter - Website

  • RabidParrotRabidParrot Formally RabidParrot. Member Posts: 956

    @ForumNinja said:
    Mac Creator 1.24.41 Fixes:

    • General improvements

    Windows Creator 1.24.5854 Fixes:

    • General improvements

    Engine 1.24 Fixes:

    • Fixed the issue with center of mass and collisions
    • Updated AdMob, RevMob, Chartboost sdk's
    • Slight rework done on how the engine handles time (this should fix some of the timing/animation issues that have cropped up)
    • Some sweet load time / spawning optimizations as well! :)

    What is the Chartboost SDK version?

  • freneticzfreneticz SwedenMember, PRO Posts: 774

    Better with the slo-mo in preview but still pretty much umbrella showing in creator

  • AlchimiaStudiosAlchimiaStudios Member Posts: 1,069

    Animation issue is resolved on device now. But performance is the worst then i've ever seen on my project. This same project was fine several builds ago.

    I don't know if it is because of the changes to the Garbage Collector where images are being kept in RAM when they are called next scene, but memory usage is up higher in most situations in my game according to Xcode's Activity Monitor.

    On scene changes sometimes i'll get memory warnings now, whereas before it would drop the images fully and load up the scene and progress without issue. Very occasionally the GC would memory dump in the middle of scene and cause a lag spike, but for my project that is better than memory usage constantly being high.

    However that being said, as a result of those changes, load times are down. But it just isn't worth the trade off for me.

    I also lost about 5-15 fps in certain situations somehow. I'm still looking for the source of the drops in FPS, but collisions and spawning are the most likely culprits from what I can tell so far (that's whats happening on screen during most spikes).

    I know it's been asked before, but Is there anyway we can get some sort of behaviour to run a full GC cycle on demand (or at least on scene change as an option)?

    Anyway... back to tracking down those FPS issues.

    Follow us: Twitter - Website

  • drizzlingdrizzling Member, PRO Posts: 101

    Hi, after I update 1.24.41, I couldn't open previous game salad project! Im using Macbook Pro, OSX EL Captain v10.11.2

  • ArmellineArmelline Member, PRO Posts: 5,354

    @drizzling said:
    Hi, after I update 1.24.41, I couldn't open previous game salad project! Im using Macbook Pro, OSX EL Captain v10.11.2

    Click the chat bubble and let the GS team know.

  • patoghpatogh Member Posts: 23
    edited January 2016

    Hi, The game with " Interstitial Chartboost " hanging up in this version and other version GS since yesterday "iPhone with ios 7.1.2" with other ios I don't check it. I think with Updated AdMob, RevMob, Chartboost sdk's released this problem.

  • BitvorkBitvork Member, PRO Posts: 119

    I am now testing 1.24.41. Thanks for the update and your hard work guys! ;-)

    Although I have no motivation to report anything via Bugzilla (I have reported 3 bugs 11 months ago and everything is still without any comment, reply or something - 725, 726, 727), I have prepared two new bugs about corrupted “random” function and self/game.Time feature. Hope it helps this time…

    http://bugs.gamesalad.com/show_bug.cgi?id=1544
    http://bugs.gamesalad.com/show_bug.cgi?id=1545

  • HopscotchHopscotch Member, PRO Posts: 2,782
    edited January 2016

    @ForumNinja

    I can confirm the RANDOM() bug and see it quite serious
    http://bugs.gamesalad.com/show_bug.cgi?id=1544
    The first value returned by the random() function is always the same.

    @Bitvork
    To your other bug
    http://bugs.gamesalad.com/show_bug.cgi?id=1545
    Maybe I am not understanding your problem, but to me this still works fine. Timers always worked in real numbers, and still do. Doing a delta time in fractions of a second like you describe in your bug still seems to work.

    However, timers seem to only have a precision of 100s of a second now.

    Where before game.Time would give values like 1.317263558 to 9 decimal places,

    you now (v.41) get
    1.25 or
    1.3166666666667 or
    1.0833333333333
    to 13 decimal places, but only an accuracy of 2. The decimal places after 100s of a second, are ALWAYS either 0 or a repetition of 3s or 6s due to some internal calculation.

    Not sure what the ramifications of this will be yet.

  • MentalDonkeyGamesMentalDonkeyGames Member Posts: 1,276

    Thanks for the hard work!

    Anyone noticed creator loading a LOT with this release? Especially when opening/backing out from actors with lots of rules.

    Mental Donkey Games
    Website - Facebook - Twitter

  • BlackCloakGSBlackCloakGS Member, PRO Posts: 2,250

    updated post with ad sdk versions

  • freneticzfreneticz SwedenMember, PRO Posts: 774

    @NipaGames said:
    Thanks for the hard work!

    Anyone noticed creator loading a LOT with this release? Especially when opening/backing out from actors with lots of rules.

    Yes its really slowing down really fast :(

  • SocksSocks London, UK.Member Posts: 12,822

    @Hopscotch said:
    you now (v.41) get
    1.25 or
    1.3166666666667 or
    1.0833333333333
    to 13 decimal places, but only an accuracy of 2. The decimal places after 100s of a second, are ALWAYS either 0 or a repetition of 3s or 6s due to some internal calculation.

    Not sure what the ramifications of this will be yet.

    Looks like the timers have followed the animation behaviour in that now they are quantised to the nearest 1/60th/sec ?

  • stevejstevej Member, PRO, Chef Emeritus Posts: 435

    It appears that the random number issue is a bug in Lua. We'll look at replacing the Lua function with a different one.

  • ArmellineArmelline Member, PRO Posts: 5,354

    @stevej said:
    It appears that the random number issue is a bug in Lua. We'll look at replacing the Lua function with a different one.

    If you're doing that, any chance we can have the ability to specify a seed too? (The 30th or so thing I've asked for today...)

  • HopscotchHopscotch Member, PRO Posts: 2,782

    @Armelline said:
    If you're doing that, any chance we can have the ability to specify a seed too?

    I was just typing this. A BIG please from me to.

  • BitvorkBitvork Member, PRO Posts: 119

    @Bitvork
    To your other bug
    http://bugs.gamesalad.com/show_bug.cgi?id=1545
    Maybe I am not understanding your problem, but to me this still works fine. Timers always worked in real numbers, and still do. Doing a delta time in fractions of a second like you describe in your bug still seems to work.

    However, timers seem to only have a precision of 100s of a second now.

    Where before game.Time would give values like 1.317263558 to 9 decimal places,

    you now (v.41) get
    1.25 or
    1.3166666666667 or
    1.0833333333333
    to 13 decimal places, but only an accuracy of 2. The decimal places after 100s of a second, are ALWAYS either 0 or a repetition of 3s or 6s due to some internal calculation.

    Not sure what the ramifications of this will be yet.

    @Hopscotch
    It is hard to explain. :) But I have updated that bug (1545), take a look. ;-) This is definitely not any kind of critical issue. I just noticed the difference between 1.24.35 and newer versions. If someone is dependent on EXACT real time during any session (lots of code or actors occurred) he can get into trouble because game.Time could be slower in some situations than normal. It should be constant.

  • patoghpatogh Member Posts: 23

    Chartboost does not working and the game hanging up (useless). Test device ios 7.1.2 with GS 1.24.40 & 1.24.41 but with older GS version 0.13.41 working very well and smooth.

  • drizzlingdrizzling Member, PRO Posts: 101

    @Armelline Update with regards to my file cant open in 1.24.41.

    Have asked and sent a file to the game salad team. And they solved it for me right away!

    Just to share with the rest incase any also have this Issue.

    Cause: an empty/missing actor tag name that cause my file to be broken!

    Solution:
    1) Use an older game salad creator version 1.24.40 and open up my project.
    2) Just add any name to the empty actor's tag. Save.

    And my file is able to open in 1.24.41!!

    Thank Gamesalad team for the solution within 24 hours! Thumbs up =)

  • AlchimiaStudiosAlchimiaStudios Member Posts: 1,069

    So I did some more testing. Spawning and collisions are def the issue for me now.

    I created a testing/troubleshooting version of my project and ran it through instruments again. This time I created a few settings to turn off my melee collisions (when attacking it spawns a collider to determine if hit) and to turn off my enemies combat behaviours (taking dmg involves spawning custom combat text numbers).

    Just turning off my collision actor that spawns for melee increased avrg fps by 15.

    Turning off my enemies but leaving collisions on for combat also increased fps by 15.

    Combined it was, as you might guess, a 30+ fps increase.

    This wasn't an issue in the past for me at all, and very little has changed in the project since last time it was stable.

    I know spawning isn't always optimal, but this unworkable in some cases. I guess i'll rewrite some systems to not use spawning and do some other type of detection for melee collision now. Maybe this info can help the GS team though (also going to contact support tomorrow).

    Follow us: Twitter - Website

  • 8bitninja8bitninja Member Posts: 367

    I'm having an issue with publishing. It's getting hung up while generating. Just putting that out there.

  • drizzlingdrizzling Member, PRO Posts: 101
    edited January 2016

    I realise that the new version cant seem to save/load game boolean attribute!! Anyone having same issue? Game can load/save the other attribute! Only cant save/load game boolean attribute!!

  • epicmanepicman Games Developer Crewe, United KingdomMember, PRO Posts: 21

    @drizzling said:
    I realise that the new version cant seem to save/load game boolean attribute!! Anyone having same issue? Game can load/save the other attribute! Only cant save/load game boolean attribute!!

    I am actually having the exact same issue, this needs fixing because many games save / load for the user :/

  • crestwoodgamescrestwoodgames Member Posts: 191

    I am also having to roll back to 1.24.40. Performance in this latest update is down. Things seem a bit more choppy.

  • vinceclennonvinceclennon Member, PRO Posts: 5

    @epicman said:

    @drizzling said:
    I realise that the new version cant seem to save/load game boolean attribute!! Anyone having same issue? Game can load/save the other attribute! Only cant save/load game boolean attribute!!

    I am actually having the exact same issue, this needs fixing because many games save / load for the user :/

    Same here

  • The_Gamesalad_GuruThe_Gamesalad_Guru Member Posts: 9,922

    I was never one for booleans. I always us an integer as it gives me flexibility going forward in my code as I can always add another option just by adding a number.

Sign In or Register to comment.