Oups! Wanted to delete this post..
Printable View
Oups! Wanted to delete this post..
open a thread in forum for this GDPR so to only post here 292 related issues
@AND :( Apple Mode
Android does the same, but it does not tell you just send you an email or blocking any new application build with lower api than 26 when you are about to upload.
@AND can you send me a final xcode project for me to test here and make some changes for you to try in xcode 9.4
Rien que ça c'est plus que ce que j'avais osé imaginé! Excellent patch, j'ai trop hâte de pouvoir tester Fusion 2.5 ++ !!!Quote:
- General: new Global Event editor button in toolbar.
- Frame Editor: "ghost" objects are automatically removed when the MFA is loaded.
- Event editor: holding CTRL while mouse wheel scrolling scrolls horizontally.
- Debugger: the Add Object button is no longer grayed when you click on Rewind or Play.
- Debugger: now takes less CPU time.
J'ai eu quelques problèmes d'adaptation avec mon projet, mais après quelques réglages dans les nouveaux paramètres, ça semble OK - Pour l'option "Forcer l'option Chargement à la demande", j'ai dû mettre ["Oui" (forcer)] pour éviter une perte de framerate au bout de quelques secondes quand je teste mon projet. Mais je ne comprends pas bien les différences entre les 3 possibilités il faudra que je demande.
Et il semble que certains de mes fichiers sonores joués aient vu leur volume mis à 0 lors de la migration, mais je n'en avais pas implémenté encore beaucoup donc ça allait pour corriger (Option "Changer le volume d'un échantillon")
Encore bravo et merci pour cet excellent build!
This is very interesting. I have already been building with Android Studio and Gradle for a while now, with some tricks and importing. If i install the update, can i still edit with Android Studio after exporting?
By the way, how did Objects not work with API 24+ before? i've been using API 26 without problems.
I wonder how you were editing the project but yes there is no change, that should still work.
Fernando says that "Objects now work with API 24 and above." is about a change in the management of the "dangerous" permissions. I guess those permissions are not used by everyone. I'll remove this line as there is already another line about those permissions.
Bizarre le problème de frame rate, il aurait fallu plutôt mettre "Oui (ignorer)". Cette option fonctionne de la façon suivante:Quote:
Originally Posted by Windlake
- Non: l'option Chargement à la demande des objets est prise en compte
- Oui (forcer): l'option Chargement à la demande des objets est forcée, ils seront tous en mode "Chargement à la demande"
- Oui (ignore): l'option Chargement à la demande des objets est ignorée, aucun ne sera en mode "Chargement à la demande", ils seront tous chargés au départ
Un petit exemple peut-être? Ca ne semble pas normal.Quote:
Originally Posted by Windlake
Salut Yves,
Merci pour les éclaircissements, je comprends mieux :)
Il y a beaucoup de fastloops et de foreach qui génèrent des particules en permanence (avec transparence + shader custom plutôt basique), il est possible que je doive optimiser tout ça et/ou que je ne doive pas en abuser.
J'avais remarqué qu'en debug, l'option "Oui (forcer)" consommait bien moins de mémoire (350 M0 au lieu de 500+ dans la scène en question), mais je testerai la fonction "Oui (ignorer)" :)
Pour le bug sonore, rien de compliqué, j'ai un système d'affichage de texte custom et il y avait un bruit d'affichage de caractère qui était lancé. Bon j'avais fait ça un peu à la volée, ça fonctionnait mais je n'ai pas poussé plus loin. Je joins une image de la ligne (j'utilisais l'objet "Sons" par défaut, j'ai juste modifié l'icône dans le thème)
Attachment 26432
Le son était à 0 quand j'ai testé sous le build 292, alors qu'il était à 75 auparavant.
(A y regarder, le timer "tous les" dans la boucle est plutôt... moche. Un comptage des caractères couplé à un modulo serait plus propre pour créer un intervalle dans le lancement du son.)
>> Le son était à 0 quand j'ai testé sous le build 292, alors qu'il était à 75 auparavant.
Tu veux dire qu'il a été modifié dans les événements? O_o Je ne vois pas comment ça pourrait être possible. Tu as le MFA original et tu peux reproduire le problème?
Les for each sont optimisés dans le 2.5+ qui va sortir incessamment sous peu, ça devrait booster un peu.
L'option "Oui (forcer)" consomme moins de mémoire effectivement au départ car les objets sont chargés au fur et à mesure. Dans 2.5+ (toujours) ça sera moins un problème car les images utilisent moins de mémoire.
@Fernando I did not see the new GPG yet in the clickstore for the new android versions
+ Gradle automatically downloads the libraries and build tools it needs, however it doesn't ask you for accepting licenses. If you get an error message that says you have to accept licenses, open an explorer window, go to the Android SDK folder (usually C:\Users\YourName\AppData\Local\Android\Sdk), open the tools/bin sub-directory, open a command prompt window in this directory and type "sdkmanager --licenses" (or sdkmanager --update for old versions), this will allow you to review and accept the licenses of the downloaded files. You can also simply open the SDK Manager and update the build API/Tools, this may work.
It compiled fine for me, but I do not have that folder
What I do have is this
C:\Users\User name\AppData\Local\Android\android-sdk\build-tools\28.0.3
Under C:\Users\User name\AppData\Local\Android\android-sdk is an "sdk manager.exe" but once I use that with the --licenses I do not get any licenses to accept.
As far the Gradle files, not sure what I am looking for. Are those files in the folder somewhere to check ?
edit2: found all the updates under my downloads....
if you can compile no problem the license folder should be in another place
@Fernando
Faulty Flap Xcode project - https://drive.google.com/file/d/1QTe...ew?usp=sharing
I built the Faulty Flap final Xcode project in the Mac Editor of CTF2.5 beta 292.0
I'm running the latest MacOS Mojave 10.14.4 and the latest Xcode version 10.2
Faulty Flap runs fine on my iPad 5th gen. running the latest iOS 12.2, but has the problem I mentioned earlier on my old iPhone 5 which has iOS 10.3.3 - the most updated iOS it can handle. It's also the same problem for the iPad 2 which has iOS 9.3.5 - the most updated iOS for this device.
I set the deployment target in Xcode to version 8.0 as I did previously with my own apps
I also checked building the mfa in CTF2.5 R291.6, then building it Xcode with the same details above, and with 291.6 Faulty Flap works normally on the iPhone 5 and iPad 2
AND, I've just sent you (by PM) a new version of the iOS runtime with a fix made by Fernando yesterday. Let us know if it fixes your problem.
Excellent :) Thanks!
Just in case you missed my post, UserGoogleAccount$( "Android object" ) seems to not work.
T'as pas de problème de résonnance du son si l'action est répété avant la fin de lecture du son? perso j'ajoute un "Stop" du son avant de le lire pour éviter des effets de résonnance vraiment désagréables.
@Yves
Oui c'est ça, les 2 lignes où se trouvait le son à 75 ont vu le volume de l'échantillon sonore tomber à 0. Je suis sûr de ne pas avoir touché ça (et même si j'avais eu à le faire, j'aurais simplement désactivé les lignes comme d'habitude, plus simple et sans modification). Peut être y'a-t-il une manip dasn Fusion qui met automatiquement le volume à 0 pour certains fichiers sonores et j'aurais fait un clic accidentel sans m'en rendre compte? C'est possible aussi.
Le son est vraiment très court (bruit de machine à écrire pendant l'affichage du texte), là j'ai changé les conditions en ajoutant un intervalle de temps pour jouer le son toutes les 5 lettres affichées (le "Tous les 00"04 fonctionnait aussi mais je ne trouvais pas ça très précis), c'est maintenant vraiment très modulable. L'astuce du "Stop", c'est sûr, je vais l'utiliser à un moment ou a un autre quand je m'attaquerai sérieusement au sound design et tout ce qui va avec :)
@Yves
Je viens de check mes mfa sur tout le mois dernier (sauvegardes différentes) et l'échantillon sous le build 292 est à 0 dans tous mes MFA. C'est trop bizarre. Ce n'est clairement pas de mon fait, j'en suis sûr. Mais ce n'est pas très grave dans mon cas, c'était juste pour laisser une trace du "bug" si ça arrive à d'autres.
Je t'ai envoyé un PM.
I have a problem on iOS, my background ingame are all black (i use active system box for parallax effect).
On the main menu i can see the background Moving, but if i click on a menu button (i move the X position of the screen to show another part of the menu, but the background setting "follow the frame" is unchecked so it must be Always visible)the whole background is black but i can see a little part Moving.
please post a small mfa that reproduce this issue for me to check here
I don't know why, I can't create a new theme, I can only reply here.
How can I get the layers of each object?
Attachment 26434
Here is a MFA simulating my two problems with the active System box object, pretty explicit:
Attachment 26438
Update: the build 292.1 is available, it fixes the 2 iOS issues reported in this thread, as well as Android building issues. This build is going to be uploaded to Steam too, as beta version.
Thank you again for testing!
The "Ios Single Line object" (i Don't know if it's the same for the ios multi line object) action "show" does Nothing since the beginning and was never fixed in any build.
i will take a look
Hello Clickteam, your skins (dark, girly, glossy) have a spelling error in the info.ini, it says -> [EX_Editor] - fontName = Courrier.
Since we are talking about these little errors from earlier Builds already, could you also take a look at these two minor things please?
Attachment 26441
About Android and gradle:
Is Fusion attempting to download a specific gradle release if we build an app for the first time?
I don't really like this solution, since it requires internet connection if the gradle files are missing
and I'm also afraid that the automatic solution doesn't work safely on every Windows OS for all people.
It did work on my Windows 10 for example, but on my machine with Windows 7 and Windows 8.1 it always seems to get stuck in the process.
But generally - if this is only about getting gradle, could you additionally provide a solution for us to download and install gradle manually and add the option to turn this automatic behaviour off?
Otherwise we have this only option and if something goes wrong, Fusion will attempt to download again and leaves no chance to fix the problem manually.
If there is no way to improve this or provide another more comfortable solution, I would even prefer a new Path setting for gradle installation in General Preferences, because
you have to go there anyways to set the path for Android SDK directory, so there is no extra hassle.
AFAIK yes, Fusion needs version 3.3.1 of Gradle. I'll discuss with Fernando about it.
I have a problem here. Font was changed in Settings for Expression Editor but every time Expression Editor shows - I need to change font size again here... to load my font of choice:
Attachment 26443
Fernando i have big problems here with 292.1
I build i with api 28 target android 9 and the save screenshot function saves the screenshot with glitches ,the colors are different and the alpha transparency of active items on screen appears as solid black in the screenshots..
Example an active object that has an image that contains semitransparent pixels with alpha of 30 appears as solid black around the frames of the sprite
Original image as it is seen when app is running
https://pasteboard.co/I7J6NBQ.png
Image after screen save function on android with cf 292.1
https://pasteboard.co/I7J72F2.png
please post a small mfa that reproduce this issue among the devices model, specs, thanks in advance
The debugger now shows 52 AltVals. Does that mean that they'll be nameable in 2.5+? If not, is it necessary to show them? It clutters up the debugger to have such a long list of AltVals, especially if they're unused.
Actually, either way, I think it'd be great to have something like a "show named alterable values only" checkbox for the debugger, or some other method of hiding unused (or at least probably unused) values from the debugger. That would be particularly great if combined with an option to reveal alterable values immediately (so we don't have to click that tiny + icon every time)
Just to clarify my above comment. The checkbox I suggest would force the debugger to only display those alterable values that have had their name manually changed by the user, whether that happens to be 0, 3, or 26.
Of course, the purest solution to displaying only values the user is interested in would be for the debugger to just display only those altVals that are actually used in the events, and ignore ones that are never referenced. But I imagine that that would be difficult/impossible to implement, since altVals could be modified dynamically in expressions without explicit references to their name or number.
But I think a "named variable only" checkbox would offer a reasonable approximation.
Actually, Perhaps a simpler and more elegant solution would be if it only showed altVals that were accessible in the properties window (if I've only clicked "new" 5 times to reveal an object's altVals A-F, then only show altVals A-F in the debugger for that object). At a guess, that sounds like it might be relatively simple to implement. And then you could make this the default behaviour (I think most users would be satisfied with it) , but add a "show all alterable values" checkbox in application preferences to override it with the old behaviour.
Fernando i am now in vacation far away from my development pc so i cannot provide mfa right now , i will post an mfa in a week time , thanks