OFFICIAL 2.3.3 Beta Release

Status
Not open for further replies.

Zhanghua

Member
Beta 2 is now rolling out - IDE v23.1.1.300 Runtime v23.1.1.297

Mac IDE users, there is an issue with our webserver updating the feeds properly, so you won't get an update notification just now (we will fix this soon, but it may be tomorrow before you would see a prompt). If you do want to jump on the new version asap, then please just download it from https://gms.yoyogames.com/GameMakerStudio2-23.1.1.300.pkg or the release notes page.

A small stack of fixes have been made (and the Inspector for Objects was indeed added), so please see the release notes for these. The English manual is currently uploading, so this should finally fix the search being broken.

We have confirmed the Android Studio 4.2 issue, and so this is being fixed soon. However, for now our advice remains to use Android Studio 4.1, as you would need this version for 2.3.2 anyway.
nice done!
 

drandula

Member
Please check if Beta 2 has fixed it and, if not (and If you haven't already) please submit a ticket with that project (or a cut-down sample which still shows the issue) and we will look into this. Thanks.
I think original problem is fixed, atleast I cannot find similiar error message. I thought following error messages related to first problem, but they persisted in new update.
1623861401190.png


As there were now fewer error messages, I could do more sensible testing, and found out that YYC doesn't like if you put undefined or function as default argument:
1623861443647.png


edit. I'll make bug report too
 
Last edited:

kburkhart84

Firehammer Games
As there were now fewer error messages, I could do more sensible testing, and found out that YYC doesn't like if you put undefined or function as default argument:
Oooh, I don't like that one. I currently have some functions with undefined values as defaults for optional arguments(and I even check against that to possibly do different behavior). I haven't converted or messed with this beta yet, but I really hope that one gets fixed because undefined is a nice value to have for these certain specific use cases.
 

drandula

Member
Here is a thing which works in Windows VM & YYC, but doesn't work with HTML5 (stable or beta):
1624011286707.png

Here global methods are initialized in script-file, which is automatically executed at the start of game. Game starts normally, but later when you try call global.gBar, this error happens:
1624011557992.png

Problem happens, if function is after the method initialization or stored in different script-file.

There is no problem if all functions are in same script-file and before method initialization.
1624011321418.png

Or there is no problem if globals store only function references.
1624011658424.png

I have tested with Beta that there is no such problem with Windows VM or YYC.
 

Spasman

Member
any insider wanna help me cheat and let me know if a fix for this memory leak I reported will be out any time soon? Request #184671
It's no rush if it'll be a while, I just don't want to refactor if the fix will be out like.. next week.
For context, the memory leak occurs if you set a previously created array to a non-array value, the array is destroyed but the memory allocated is never freed
 

malonso

Member
I hope that mean progress for Linux IDE too ;) I've got a feeling that if OpenTK got some issues on macOS, then it might cause same issues on Linux, as they are both closer together than Windows.
Linux support is needed, the main reason for using Windows is GameMaker Studio.
 

COWCAT

Member
I can no longer import several frames at once by using the "Import" button on a sprite.
I used to select multiple files and it would insert them all, but now it only seems to import one of them.
Now I have to Edit Image, THEN Image => Import Frames which is less practical.

Another minor bug, my cursor seems to disappear when I'm on a field such as asset browser search or input fields in the settings.

Also one bug that was introduced in a previous version and I was hoping for it to be fixed but it isn't.
When I do a search/filter in the asset browser and I duplicate an asset (with CTRL + D) it collapses everything instead of doing a refreshed search. So I have to change my search and look for my duplicated asset. I do this all the time and it's very annoying.
 
Last edited:

ScottD

GameMaker Staff
GameMaker Dev.
Beta 3 is currently going out just now: IDE v23.1.1.306 Runtime v23.1.1.305

Contains some minor fixes for the IDE and a major fix for YYC in the runtimes (along with some not so majors).

As always please use our bug reporting service if you encounter any bugs with this beta.


Enjoy!
 

gnysek

Member
Beta 3 is currently going out just now: IDE v23.1.1.306 Runtime v23.1.1.305
Contains some minor fixes for the IDE and a major fix for YYC in the runtimes (along with some not so majors).
details:
Changes/Fixes since IDE 300
  • Please note this release is mostly about the runtime fix, so ensure your runtime updates, but we did add a couple of very small IDE fixes
  • Inspector: Fixed the content for a tooltip on an Object's Events tab and also made it disappear properly once an event has been added
  • Preferences: Interacting with Keystore in any way when using Android Studio 4.2 results in a "JDK not installed correctly" error
  • Undo: A small fix for undoing a change after closing its editor didn't always open the editor properly
Changes/Fixes since runtime 297
  • Made a fix for YYC not working correctly with dot notation and some in-built variables, which might have caused issues with keyboard_string not clearing, room persistance not being remembered, view_camera values not being accessible, and a couple of other things reported as bugs overnight (thanks!)
  • Building Projects: YYC will fail to compile if using undefined or a function as a default parameter
 

8BitWarrior

Member
I have found a bug with YYC when setting arrays belonging to a global struct. Example below reproduces the issue:
GML:
// YYC PRODUCES WRONG RESULTS
global.struct = {}; // Create global struct
global.struct.array = array_create(3); // Assign new array to global struct
global.struct.array[0] = 100;
global.struct.array[1] = 200;
global.struct.array[2] = 300;
show_message(global.struct.array); // Show values -- YYC reports wrong values

// Using "@" accessor corrects problem for YYC
global.struct.array = array_create(3);
global.struct.array[@ 0] = 100;
global.struct.array[@ 1] = 200;
global.struct.array[@ 2] = 300;
show_message(global.struct.array);
Update:
I have submitted a bug report.
 
Last edited:

gnysek

Member
I've reported a bug about Inspector animating even if IDE animations are disabled. Nothing important, as it's not breaking anything, it's just inconsistency.

Also I've got idea about some optional feature, that collapses other groups in inspector as new one is expanded, however I have no idea how to add it not removing current behaviour, as lot of users may like it. Maybe a button for quick expand/collapse every group?
 
Last edited:

mino_dev

Member
I can no longer import several frames at once by using the "Import" button on a sprite.
I used to select multiple files and it would insert them all, but now it only seems to import one of them.
Now I have to Edit Image, THEN Image => Import Frames which is less practical.
Additionally, doing Image => Import Image(s) seems to import the images in a random order; rather than in the order they're selected in.
(Edit: submitted a bug report)
 
Last edited:

gnysek

Member
In next beta, there will be update, described as:
in the manual we have added some new pages for recent additions and lots of typo/formatting fixes for various older pages
And it's always up to user to find out differences. Sometimes those important, as it was not only a typo, but something was wrongly described, or even code example was wrong, but people may not find it out as they remembers previous description, and finding 2-3 changes among 1000 pages isn't easy. Would be good, to put a list of every new page, or updated page (except it was really word typo/formatting only).

Btw. we're in Q3 2021 since today, so this update seems to be little late... ;)
 
Last edited:

rIKmAN

Member
Btw. we're in Q3 2021 since today, so this update seems to be little late...
1 day into Q3 is a little late?
The day isn't even over it's 12:15 here in the UK - lunchtime and that means bacon sandwiches all round! :banana:
 

rui.r6o

Member
1 day into Q3 is a little late?
The day isn't even over it's 12:15 here in the UK - lunchtime and that means bacon sandwiches all round! :banana:
I believe @gnysek mentioned it is a little late because according to the roadmap version 2.3.3 should have been released during Q2. Either way I'm looking forward for the stable version to come out
 

rIKmAN

Member
I believe @gnysek mentioned it is a little late because according to the roadmap version 2.3.3 should have been released during Q2. Either way I'm looking forward for the stable version to come out
Yeah but it's only 4hrs into the working day - give them a full day at least! :)
I do often forget people are waiting for Stable versions though as I only use the Beta's.
 

Kezarus

Endless Game Maker
Has it been reported? If so what is the ticket number? We don't track forum posts...

Several issues have been fixed - there should be another Beta very soon so you should check that.

Russell
Yep, it was indeed reported. =]

Unfortunately I didn't kept the e-mail that I receive the ticket number and I can't search tickets posted on the Bug Report, can I?

Keep up the good work, mates!


Cheers
 

Dan

GameMaker Staff
GameMaker Dev.
As already spotted above, Beta 4 is now rolling out - IDE v23.1.1.312 Runtime v23.1.1.313 (I was editing the required SDKs FAQ, so a "slow" announcement post here / you were all too quick ;) )

I have just spotted I missed a "G" on the second point in the release notes, so apologies there, and be advised whilst they are not in the notes there are actually a few additional attempted fixes for source control issues when you have conflicts in your project. However, we have found today that these fixes caused other issues. So, for now, if you do get conflicts in your projects, please close GMS2 and fix the issue externally, rather than attempting to fix inside GMS2, as this can cause assets to be lost and warnings to pop up. We have all this bugged already, so there is no need to report it, but feel free to use source control beyond this one caveat and if you have other issues there, please do report them as per normal.
 

chamaeleon

Member
As Russell says, although I did just go and find your ticket and confirmed the issue, so I will write it up now. Thanks!
Got email notifications just now about the progress, thanks.

For me at least, a workaround appears to be to use a YYC build. There is no observed leak when I try that. The drawback is of course that you have a suffer with a faster execution speed.
 

Kezarus

Endless Game Maker
As Russell says, although I did just go and find your ticket and confirmed the issue, so I will write it up now. Thanks!
Hmmm, wait. I sent this bug on Dec 23th. I'm positive about it. I don't keep tickets e-mails. Can't you search for a ticket posted from someone?

I will make sure to keep this kind of e-mails in the future, but I find this very odd.
 

JeffJ

Member
Would be good, to put a list of every new page, or updated page (except it was really word typo/formatting only).
I have tried to advocate for this before, too. I don't see why it would be a big bother if the changes are tracked and noted - just put it into the release notes. Same with when changelog says something like "a bunch of functions were changed"... Eh, mind elaborating? Would love to see some more transparency on this and not just leave the user to guess and look around manually.
 

Dan

GameMaker Staff
GameMaker Dev.
Hmmm, wait. I sent this bug on Dec 23th. I'm positive about it. I don't keep tickets e-mails. Can't you search for a ticket posted from someone?

I will make sure to keep this kind of e-mails in the future, but I find this very odd.
I have found your ticket now, and actually I did confirm the bug and sent you a reply back that same day (we got your ticket a few minutes into the morning of Dec 24th and I replied lunchtime that same day). I will go hunt tomorrow as to why the bug report I filed was not actually then fixed, but that's now an internal matter and you can rest assured we did handle your ticket. Thanks again, and as Russell says, we will now finally get this fixed.
 

Kezarus

Endless Game Maker
I have found your ticket now, and actually I did confirm the bug and sent you a reply back that same day (we got your ticket a few minutes into the morning of Dec 24th and I replied lunchtime that same day). I will go hunt tomorrow as to why the bug report I filed was not actually then fixed, but that's now an internal matter and you can rest assured we did handle your ticket. Thanks again, and as Russell says, we will now finally get this fixed.
Oh ow, thanks for replying. =]

It was on X-Mas Eve, I was doing a lot of things here, imagine the hassle for you guys.

Glad this bug have its days numbered now. Thanks!
 
Last edited:

gnysek

Member
I also need to get back for my old tickets, as I'm pretty sure some are reported as "added to bug database" but I didn't seen them fixed, but I will do that after beta, as I also want to make some UWP tests, cause I saw some issues with iterating over arrays and ds_maps which is really, really slow (I'm getting 60 to 40 FPS drop cause of using return map[? "string]; on 200-records ds-map, about 15 times in step... - checked that on debugger profiler). It's especially annoying on Xbox Series X, which is theoretically 2x better than my PC, but on UWP it's 10x worse :p That requires some more testing, so I'm leaving it for my holidays in second half of July, but I'm pointing it out cause maybe someone have feeling that he found similar issue, but wasn't sure about it.
 

FrostyCat

Redemption Seeker
I appreciate the work that has gone into fixing YYC array/struct chaining bugs over the past few betas, but can the dev team please make time for one more before releasing 2.3.3? The ticket is #184475 and it has been open since June 10 without response.

For those interested, it has to do with using the dot syntax on argumentN variables in extensions (background note: GML extensions still take arguments the old way). If you do something like argument0.doSomething() or reference something like argument0.value, it crashes with a seemingly arbitrary error (e.g. "variable not set before used" in beta 1, "I32 argument is unset" in beta 4). It can be worked around by routing the value through a local variable, but I have 4 libraries affected by this that would need a workaround release if the beta is released to stable as-is.
 

Spasman

Member
As already spotted above, Beta 4 is now rolling out - IDE v23.1.1.312 Runtime v23.1.1.313 (I was editing the required SDKs FAQ, so a "slow" announcement post here / you were all too quick ;) )

I have just spotted I missed a "G" on the second point in the release notes, so apologies there, and be advised whilst they are not in the notes there are actually a few additional attempted fixes for source control issues when you have conflicts in your project. However, we have found today that these fixes caused other issues. So, for now, if you do get conflicts in your projects, please close GMS2 and fix the issue externally, rather than attempting to fix inside GMS2, as this can cause assets to be lost and warnings to pop up. We have all this bugged already, so there is no need to report it, but feel free to use source control beyond this one caveat and if you have other issues there, please do report them as per normal.
OMG precisely the patch I needed, thank you

Windows Defender was being set off by Igor when installing the new runtime, after looking it up online it seems to be a new false positive from Defender
If I had to guess from looking at other threat actions, it might be being set off by random name folder creation in Temp

igor.png
 

gnysek

Member
It might happen when you download update early, and checksum for that EXE file isn't reported enough times as "safe". Same happens for Smartscreen sometimes.
 

Zhanghua

Member
Fixed on newest beta!! efficient!


Dan
rwkay


I had submitted this multi-images import order issue few days ago, and it can't be report as a project style, so it's pending now.
It should import ordered by name A-Z automatically.
Please check out these problem, thank you!

bug.png


bug2.png


BUG3.png





----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
 
Last edited:

Dan

GameMaker Staff
GameMaker Dev.
OMG precisely the patch I needed, thank you

Windows Defender was being set off by Igor when installing the new runtime, after looking it up online it seems to be a new false positive from Defender
If I had to guess from looking at other threat actions, it might be being set off by random name folder creation in Temp

View attachment 41043
Yeah, we saw this as well. There was a Windows Defender definitions update which came out mid last month and started saying this. However, Microsoft fixed it with a new definitions update about 48 hours later, so please ensure your Windows Defender is up-to-date and then you will be fine.
 

rwkay

GameMaker Staff
GameMaker Dev.
We have done a new Beta Update - please check the release notes for changes most notable were the fixes for the bugs mentioned above - memory leak on strings used on multi dimensional arrays AND @FrostyCat extension bug.

We are treating this as a Stable Release Candidate - if you find any major issues please make sure you let us know ASAP.

Russell
 

nsobject

Member
...if you find any major issues please make sure you let us know ASAP.
Russell

If this is not a major issue, then what is a major issue? (To summarize: core collision functions broken on macOS)
 

FrostyCat

Redemption Seeker
The extension bug isn't over yet. This time it has to do with scripts taking arbitrary arguments being struct-unfriendly. Quick reproduction:
GML:
function biggerThan() {
    return argument[0].value > argument[1].value;
}
show_message_async(biggerThan({ value: 5 }, { value: 3 }));
It works fine on stable, but whines about argument not being set on beta 4. The ticket number is #185225.

I discovered this after noticing that my Extended Collision Functions test suite still fails on beta 4, and it has code similar to the above as test fixtures.
 

Dan

GameMaker Staff
GameMaker Dev.
Beta 6 is rolling out now to fix just the above two issues - IDE v23.1.1.314 Runtime v23.1.1.318

Also just went back and filled-in the first post in the thread properly, so now all the releases are represented again.
 
Last edited:

FrostyCat

Redemption Seeker
Sorry, but Beta 6 has a new bug that is completely detrimental to anything using argument_count. Case in point:
GML:
function argumentCount() {
    return argument_count;
}
show_message_async(argumentCount()); //Shows 1???
This is a deal-breaker for at least 3 of my libraries. There is no way I would let this build out into stable, when something this basic isn't working.

Update: The ticket number is 185262.
 
Last edited:
Status
Not open for further replies.
Top