Such a shame my project doesn't even start with this one. It dies and returnsIDE v23.1.1.261 Runtime v23.1.1.264 is rolling out now. Changes are shown at the top of the release notes and at the top of this thread, but you can see we have added even more IDE and manual translations in this version (plus added the remaining stuff for Chinese), so please do give these a try also if you're familiar with these languages.
One request for CocoaPods users who had issues compiling projects with the "pod install" command saying it could not be found - can you please confirm this is now fixed on your machines, please? We did find an issue with the OS permission level GMS2 was trying to use, which could have caused this to fail in some cases.
X://windows/x64/Runner.exe exited with non-zero status (-2147483645)
Give it a little while and then refresh your feed inside Preferences / restart GMS2, as the release was only just going out as I posted my last message, so quite possible some web cache between you and us is causing a delay there.Updated but my installation is not picking up the v23.1.1.264 runtime.
function Base() constructor {
static printMe = function() {
show_debug_message("I'm Base!");
}
}
function ChildA() : Base() constructor {
static printMe_Base = printMe;
static printMe = function() {
show_debug_message("I'm ChildA!");
}
}
function ChildB() : ChildA() constructor {
}
function ChildC() : ChildB() constructor {
}
var first = new ChildA(); //WORKS
var first = new ChildB(); // Crash: <unknown_object>.printMe(100980, -2147483648) not set before reading it.
var first = new ChildA();
var second = new ChildB(); // DOESN'T CRASH - because I instantiated "ChildA" first.
Use the static carefully. So I haven't use the static long time ago.Well I found some problem which I already filed... for people experiencing the same here it goes:
[BUG1]Here are the tests for this code:GML:function Base() constructor { static printMe = function() { show_debug_message("I'm Base!"); } } function ChildA() : Base() constructor { static printMe_Base = printMe; static printMe = function() { show_debug_message("I'm ChildA!"); } } function ChildB() : ChildA() constructor { } function ChildC() : ChildB() constructor { }
[TEST1]
[TEST2]GML:var first = new ChildA(); //WORKS
[TEST3]GML:var first = new ChildB(); // Crash: <unknown_object>.printMe(100980, -2147483648) not set before reading it.
GML:var first = new ChildA(); var second = new ChildB(); // DOESN'T CRASH - because I instantiated "ChildA" first.
[BUG2]This bug was introduced in 2.3.2 (beta) it wasn't there in last stable!GML:function A() constructor { static print = function() { show_debug_message("A"); } } function B() : A() constructor { } function C() : B() constructor { print(); static print = function() { show_debug_message("C"); } } var instance = new C(); // this prints "C", this is WRONG it should print "A";
[NOTES] On the code above if:
1) the code inside **C** moves to **B**
2) you create an instance of **B**
then there will be NO bug!
[BUG3]the print is being declared after the call!!GML:function B() constructor { } function C() : B() constructor { print(); static print = function() { show_debug_message("C"); } } var instance = new C(); // this should crash, but prints "C"
So there is no print method declaration when the call is made (in stable this would throw an error)
[IMPLICATIONS]
this can lead to a lot of problems as we are suppose to overwrite methods using:
if we have such a bug how can one assure that print (1) is from the parent (0) and not the print (2) (declared below)??GML:function A() constructor { static print = function() { // (0) // this is the declaration } } function B() : A() constructor { static print_A = print; // (1) static print = function() { // (2) // this is the redeclaration } }
These bugs reproduce 100% of the time for both YYC and VM (with the exception of BUG1 which only happens in VM).
[FINAL NOTES]
/// @description Create Event
objVar1 = 50
myStruct = {
structVar1 : objVar1, // This works OK.
structVar2 : max(objVar1, 200) // This causes an error: Variable struct.objVar1(100003, -2147483648) not set before reading it.
}
Generally, there's a bug if you use any function inside struct and want to pass variables outside of struct as arguments. Are you already reported it, or I should do it too?There is a runtime crash at structVar2's initialization
Already reported. Thanks.Generally, there's a bug if you use any function inside struct and want to pass variables outside of struct as arguments. Are you already reported it, or I should do it too?
In Windows Defender, you will need to verify device security in kernel isolation; if enabled this locks some parts in CMD along with folders.This topic is for the discussion of issues relating to the update of GameMaker Studio 2. However, this does not replace our normal bug reporting system and you should report all bugs as normal using https://accounts.yoyogames.com/contact-us
As ever, PLEASE check the Required SDKs FAQ before/during updating and ensure your antivirus/OS permissions are correct after you do the new install - User Permissions and Internet Access Required by GMS2 - before you submit any bug reports or post issues here.
You should post here if:
When reporting issues to the Helpdesk, please try to add as much information as possible, for example:
- Something has changed between the previous version and this one, breaking your game (i.e.: code that worked previously now does not)
- Something reported as fixed in the release notes is not actually fixed
- You are having IDE or system issues related to something new in this update
- You are having issues with new features introduced in this version
- You are having issues updating to this version
- Everything works and you want to tell us we are great (seriously, let us know if it all works correctly for you)
Again: posting a comment in these release threads is not a replacement for filing a bug!
- If the issue is with the IDE, then please give a screenshot if you can to help illustrate the problem
- If there are error windows, then a link to the "ui.log" file found via the "Open Log in ..." Help menu command inside GMS2
- If it is a code issue, then link to or attach a YYZ showing the most basic project possible which exhibits the behaviour
- If you receive any error dialogues, please screenshot them
- If it is a system issue (like compiling to a specific target) then please provide details of the PC being used as well as the software versions involved
Release Notes
Runtime Release Notes
Current Release IDE v23.1.1.261 Runtime v23.1.1.264 (Mar 12, 2021)
Changes/fixes are shown at the top of the release notes and at the top of this thread, but you can see we have added even more IDE and manual translations in this version (plus added the remaining stuff for Chinese), so please do give these a try also if you're familiar with these languages.
One request for CocoaPods users who had issues compiling projects with the "pod install" command saying it could not be found - can you please confirm this is now fixed on your machines, please? We did find an issue with the OS permission level GMS2 was trying to use, which could have caused this to fail in some cases.
Known Issues (in no priority order):
- Games launched in fullscreen, but then forced out of fullscreen via some method other than Alt+Enter or window_set_fullscreen() go into a borderless mode and mouse input is incorrect.
Beta 2 - IDE v23.1.1.255 Runtime v23.1.1.260 (Feb 24, 2021)
Quite a lot of bug fixes for the IDE and runtime, so I won't list them all here, but you can see them all at the top of each of the release notes pages linked-to just above.
Known Issues (in no priority order):
- The manual doesn't download the correct one - it gets you 2.3.1's manual.
- The Chinese IDE translation is still work-in-progress and incomplete - there is also the typo mentioned below that the Preference calls it "English" but using Chinese characters.
- Calling ord(integer) silently crashes YYC, but works on VM.
- Games launched in fullscreen, but then forced out of fullscreen via some method other than Alt+Enter or window_set_fullscreen() go into a borderless mode and mouse input is incorrect.
- On VM targets, appending multi-dimensional arrays using += and not using array accessors can cause silent crashes or "Index out of bounds" errors, calling gc_collect() can prevent this for some reason.
Initial Release IDE v23.1.1.249 Runtime v23.1.1.254 (Feb 11, 2021) - PLEASE UPDATE!
You will see from the notes above that there is also a companion post What's New in 2.3.2? which goes over the new features in a lot more detail and has a couple of small "demo" projects you can download. Hope you like it.
// Example
for(var i = 0; i < array_length(mlp); i++) {
mlp[i].Forward(input);
}
" In-Game: [HTML5] Debug runs give error "unable to call function undefined typeof=undefined" when calling a function in an array "
It was me who reported it, a couple of hours after Beta release and it is a bug and it is being solved for next beta I believe!I think I found something that broke in Runtime 23.1.1.264. Example:
There is a runtime crash at structVar2's initialization. The error message indicates that objVar1 isn't set before reading it, and that it appears to be expecting objVar1 to be a member of a struct, which it isn't.Code:/// @description Create Event objVar1 = 50 myStruct = { structVar1 : objVar1, // This works OK. structVar2 : max(objVar1, 200) // This causes an error: Variable struct.objVar1(100003, -2147483648) not set before reading it. }
This code worked fine in Runtime 23.1.1.260
Edit: This bug has been reported to YoYo and has been added to their bug database.
Thanks - have passed this information on now.nice for the Chinese manual!!!!
although has some mistake description, such as:
We have reproduced the first part and added that to the bug database now. The second line has been added to the suggestion feature backlog.GameMaker still copies included files to all of the targets ignoring included file settings (all platforms unchecked).
There's no way to unselect directories either.
I will get that one re-verified in case I made a mistake with the release notes or the fix got broken. If you could send us a ticket with small sample, just in case there is a related second issue which gives the same message, then that would further clarify things.Is this same error which was mentioned in Beta runtime 254:
" In-Game: [HTML5] Debug runs give error "unable to call function undefined typeof=undefined" when calling a function in an array "
I think you mean IDE v23.1.1.263 Runtime v23.1.1.265.IDE v23.1.1.261 Runtime v23.1.1.264 are rolling out now.
It seems I cannot replicate it in manner I could give small sample. I'll try again later.If you could send us a ticket with small sample, just in case there is a related second issue which gives the same message, then that would further clarify things.
Yeah... My silent crash got confirmed but still happens in v23.1.1.265. I'm very worried next "stable" will still have the problem forcing my important project to get stuck in an older runtime.I see that new tutorials page is up (was down yet yesterday), and whole YYG website got some overhaul, so stable should be out very soon (except QA will find some bug in last moment)
In my case still causing issues while I try to compile the game "Pod install result: /bin/bash: pod: command not found" I'm using the Beta IDE v23.1.1.263 macOS Catalina 10.15.7 Macbook Air 11-inch Mid 2012IDE v23.1.1.261 Runtime v23.1.1.264 is rolling out now.
Changes are shown at the top of the release notes and at the top of this thread, but you can see we have added even more IDE and manual translations in this version (plus added the remaining stuff for Chinese), so please do give these a try also if you're familiar with these languages.
One request for CocoaPods users who had issues compiling projects with the "pod install" command saying it could not be found - can you confirm this is now fixed on your machines, please? We did find an issue with the OS permission level GMS2 was trying to use, which could have caused this to fail in some cases.
I will check on that one - it might have just missed the Beta 4 cut-off, but I did think it was in...Yeah... My silent crash got confirmed but still happens in v23.1.1.265. I'm very worried next "stable" will still have the problem forcing my important project to get stuck in an older runtime.
Assuming you have followed https://help.yoyogames.com/hc/en-us/articles/360008958858-iOS-and-tvOS-Using-Cocoa-Pods already, and also cleaned your compiler cache at least once since installing this new Beta, could I ask you to send us a Helpdesk ticket with a full compiler log and the project .yyz which fails to compile, please?In my case still causing issues while I try to compile the game "Pod install result: /bin/bash: pod: command not found" I'm using the Beta IDE v23.1.1.263 macOS Catalina 10.15.7 Macbook Air 11-inch Mid 2012
Hey @rwkay, I'm running the right runtime (23.1.1.265) and the silent crash I'm referring to still happens on it. @Dan told me he managed to reproduce the crash (ticket 180965) so I guess the fix didn't arrive in time. Please let me know if this bug should have been fixed in 23.1.1.265... Maybe there's another issue we have to investigate and isolate.@Ricardo and @DroidGames Studio can you please check that you are on the correct Runtime should be v23.1.1.265 as both these bug fixes would be in the relevant runtime not the IDE
Russell
Yes I can confirm that my current runtine is the v23.1.1.265@Ricardo and @DroidGames Studio can you please check that you are on the correct Runtime should be v23.1.1.265 as both these bug fixes would be in the relevant runtime not the IDE
Russell
Test = {}
Test.T = undefined
Str = json_stringify(Test)
Parse = json_parse(Str)
Result = Parse.T
Yup, should be visible in the next 2.3.2 update actuallyand the documentation will be getting updated to reflect this @matharoo (just checking you are aware)
No, it should not be fixed in that one, as I can still reproduce the issue in a later version and it's still in our bug DB. Think that was some confusion on my part, to be honest, so apologies for that.Hey @rwkay, I'm running the right runtime (23.1.1.265) and the silent crash I'm referring to still happens on it. @Dan told me he managed to reproduce the crash (ticket 180965) so I guess the fix didn't arrive in time. Please let me know if this bug should have been fixed in 23.1.1.265... Maybe there's another issue we have to investigate and isolate.
Yay! Silent crash is goneIDE v23.1.1.266 Runtime v23.1.1.268 is now rolling out
Another small release, largely because CocoaPods should now be fixed in the Mac IDE (it certainly works for our Macs now), so please do give that a go and confirm that it's fixed for you also.
Plus also some more silent game crash fixes.
As per usual, you can find the detail for these changes since Beta 4 at the top of the release notes pages.
event_perform(ev_cleanup,0);
My example work well on current Stable, so... There's something there.I think it's by design, cause on instance create you also aren't going into newly created instance "create" event (while it's executed before next line of code, so it enters there).
But it would be nice if it would work this way, so you can report it as suggestion.
I get that Json doesn't support undefined and just outputs it as the string "null". In the current stable release json_parse converts a "null" back to <undefined>, but in the current beta, json_parse converts "null", from what I can tell to a pointer of <0>? Is that intended?@Dwarfaparte - undefined handling in JSON (according to the spec) is, well; undefined and Javascript filters out undefined completely from JSON handling to allow for cross platform handling we have decided that any undefined value that you have is converted to a null (because we can handle that consistently across platform) so that is what you are seeing.
Russell
Yes as I stated that was the only way to keep it all working across the platforms, my preferred approach would have been to convert it to undefined but I could not get that to work correctly (and consistently) within the HTML5 target.I get that Json doesn't support undefined and just outputs it as the string "null". In the current stable release json_parse converts a "null" back to <undefined>, but in the current beta, json_parse converts "null", from what I can tell to a pointer of <0>? Is that intended?
Thanks Russell, all is clear now. Sorry for the misunderstanding! I'll have to root out all the undefined values I have in my json save game system but it's not the end of the world!Yes as I stated that was the only way to keep it all working across the platforms, my preferred approach would have been to convert it to undefined but I could not get that to work correctly (and consistently) within the HTML5 target.
Russell