RMMV Guide to Upgrade MV to PIXI v5

[ephesus]

Good work, though I swear TheUnpropro had a fully working version somewhere, odd.



The core script files for MV are MIT License.

Here you go, I assume you have probably figured things out or eventually found this on your own but it can still be helpful to many people out there it seems...
https://theunpropro.itch.io/rmmv-auto-pixi-530-updater

 

[DarkSearinox92]

I tried, out of curiosity, and went against this error using the GradientWipe script with crash

Gradient Wipe - old RM Transition effect, and more

Gradient Wipe v1.0.0 Tsukimi (cji3bp62000) Introduction There are fancy transition effects in RM 2000/XP/VX, but only fading and scaling in MV. So I decided to make this script to make it more fancier, and also add the effect to characters/pictures/message window. *** Notice !!! *** this is an...

forums.rpgmakerweb.com

"texture.castToBaseTexture"


Incredibly, however, the particles-emitter works.

However, i see that pixi 6.0 has even been released. LOL we are already overtaken XD

 

[Abitkin]

I tried, out of curiosity, and went against this error using the GradientWipe script with crash

Gradient Wipe - old RM Transition effect, and more

Gradient Wipe v1.0.0 Tsukimi (cji3bp62000) Introduction There are fancy transition effects in RM 2000/XP/VX, but only fading and scaling in MV. So I decided to make this script to make it more fancier, and also add the effect to characters/pictures/message window. *** Notice !!! *** this is an...

forums.rpgmakerweb.com

"texture.castToBaseTexture"


Incredibly, however, the particles-emitter works.

However, i see that pixi 6.0 has even been released. LOL we are already overtaken XD

6.0 is still a release candidate, not released yet. I wouldn't use it for production, nor should anyone else yet.

Once it has had time to settle, then maybe.

Also, I've updated my repo to PixiJS 5.3.7 as well as fixing JOIPlay compatibility. Someone contacted me about memory usage, I'm still looking into their changes.

 

[DarkSearinox92]

6.0 is still a release candidate, not released yet. I wouldn't use it for production, nor should anyone else yet.

Once it has had time to settle, then maybe.

Also, I've updated my repo to PixiJS 5.3.7 as well as fixing JOIPlay compatibility. Someone contacted me about memory usage, I'm still looking into their changes.

Yeah, of course... however with 5.X there are too much things to check... on Rpg maker MV core and all engine, and the scripts...

 

[Roult]

Hello and thank you for this plugin

I have performance problems with a light plugin (ocram), I plan to use the pixi update "4.5.4" to "5.xx", however I will have a question,

I just have to install the plugin and start it or do I have to follow the complete transformation guide in addition? To avoid doing the same thing twice, sorry I have a little trouble understanding
.


Thanks

 

[Abitkin]

Hello and thank you for this plugin

I have performance problems with a light plugin (ocram), I plan to use the pixi update "4.5.4" to "5.xx", however I will have a question,

I just have to install the plugin and start it or do I have to follow the complete transformation guide in addition? To avoid doing the same thing twice, sorry I have a little trouble understanding
.


Thanks

What I would do, instead of a plug in, is simply replace the files in www/js and www/js/lib with the ones I have on GitHub. You should be keeping your project in Git anyways, so backing those changes out if they don't work properly should be trivial.

 

[Restart]

What I would do, instead of a plug in, is simply replace the files in www/js and www/js/lib with the ones I have on GitHub. You should be keeping your project in Git anyways, so backing those changes out if they don't work properly should be trivial.

Tried this and it seems to have destroyed the built-in layering

Here's a minimal test project

30.29 MB file on MEGA

mega.nz

I hope it's just something obvious that I messed up.


EDIT: 5.3.0 (this version: https://fenixenginemv.gitlab.io/guides/pixi5-upgrade.html ) works for layering, though it still has the singleletter textbox bug. I think there's a bug in the 5.3.3 build on your github

 

[Kyuukon]

I'm having the same layering issue when using @Abitkin library :/

Haven't tried LTN Games' update yet. What is that "singleletter textbox bug" you are talking about?

 

[EGARPG]

First of all @Abitkin thank you for the time that you give to us. This is helping me too much on the project, but I have a problem xD I don't know how to apply the Patch I even tried to use the guide on moodle but still nothing. Do you have a GitHub or somewhere to download all the rpg_core, rpg_manager, etc updated with the patch?. Thank you so much and wish you a wonderful day╰(*°▽°*)╯

 

[Abitkin]

Okay I think I fixed up the layer issues.

As to how to apply this, assuming you have bash:

git clone https://git.kins.dev/rpgmaker-mv-pixi5
GAME_FOLDER="<path to game folder>"
cp -r "${GAME_FOLDER}/www/js" "${GAME_FOLDER}/www/js.bak"
# or use git instead to keep track of changes to your game
cd rpgmaker-mv-pixi5
cp js/rpg_*.js "${GAME_FOLDER}/www/js/."
cp js/libs/pixi-tilemap.js "${GAME_FOLDER}/www/js/libs/."
cp js/libs/pixi-picture.js "${GAME_FOLDER}/www/js/libs/."
cp js/libs/pixi.min.js "${GAME_FOLDER}/www/js/libs/pixi.js"

And it should be good to go.

 

[Restart]

Okay I think I fixed up the layer issues.

Latest patch seems to fix the layering issues, though it seems like the updated pixi causes some serious performance loss when YEP_GridFree Doodads is added as a plugin (even on maps with no doodads). I have a 144 Hz monitor so the refresh hovers around 144ish when idling in a map with nothing happening (using the f2 fps meter) however when I have the new pixi + doodads it drops down to high 70s low 80s on my full project, and high 80s/low 90s on an empty test project.

I also see corresponding performance drops on other maps (including seemingly random lag spikes) which were previously very smooth (I have one area with about 100 moving events, which before the update had no issues - after the update it was down to single digits). That's a lot harder to have quantified numbers for, so hopefully the idle fps change comes from the same underlying problem.

I'd upload a minimal test case but doodads is a paid plugin now, so here's screenshots of me adding the doodads plugin to the previous test case I uploaded. Given that doodads was the only active plugin, this should be fairly isolated as an issue.

Doodads plugin disabled:

Doodads plugin enabled:

Notably, no doodads exist anywhere on the map - merely having YEP_doodads enabled is enough for this big drop.

 

[Abitkin]

Latest patch seems to fix the layering issues, though it seems like the updated pixi causes some serious performance loss when YEP_GridFree Doodads is added as a plugin (even on maps with no doodads). I have a 144 Hz monitor so the refresh hovers around 144ish when idling in a map with nothing happening (using the f2 fps meter) however when I have the new pixi + doodads it drops down to high 70s low 80s on my full project, and high 80s/low 90s on an empty test project.

I also see corresponding performance drops on other maps (including seemingly random lag spikes) which were previously very smooth (I have one area with about 100 moving events, which before the update had no issues - after the update it was down to single digits). That's a lot harder to have quantified numbers for, so hopefully the idle fps change comes from the same underlying problem.

I'd upload a minimal test case but doodads is a paid plugin now, so here's screenshots of me adding the doodads plugin to the previous test case I uploaded. Given that doodads was the only active plugin, this should be fairly isolated as an issue.

Doodads plugin disabled:

Doodads plugin enabled:

Notably, no doodads exist anywhere on the map - merely having YEP_doodads enabled is enough for this big drop.

I happen to have a copy (1.09) from a different game.

So based on that, and assuming this is not a playtest (the playtest enables a lot of extra code that could cause problems) we would need to restrict ourselves to what APIs YEP_GFD is overloading.

It looks like the two functions are:

Tilemap.prototype._compareChildOrder = function (a, b) {};
ImageManager.isReady = function () {};

It is simply a mater of disabling each piece until you find the real cause. Note the code being overridden original was:

Tilemap.prototype._compareChildOrder = function (a, b) {
    if (a.z !== b.z) {
        return a.z - b.z;
    } else if (a.y !== b.y) {
        return a.y - b.y;
    } else {
        return a.spriteId - b.spriteId;
    }
};

And now it is:

Tilemap.prototype._compareChildOrder = function (a, b) {
    if (a.z === b.z) {
        if (a.y === b.y) {
            return a.spriteId - b.spriteId;
        }
        return a.y - b.y;
    }
    return a.z - b.z;
};

Since the shader tilemap inherits from tilemap and Yanfly is overriding that, you might need to set shader tilemap back to a normal sort.

Of course this is all based on looking at code, not running anything yet.

 

[ephesus]

I happen to have a copy (1.09) from a different game.

So based on that, and assuming this is not a playtest (the playtest enables a lot of extra code that could cause problems) we would need to restrict ourselves to what APIs YEP_GFD is overloading.

It looks like the two functions are:

Tilemap.prototype._compareChildOrder = function (a, b) {};
ImageManager.isReady = function () {};

It is simply a mater of disabling each piece until you find the real cause. Note the code being overridden original was:

Tilemap.prototype._compareChildOrder = function (a, b) {
    if (a.z !== b.z) {
        return a.z - b.z;
    } else if (a.y !== b.y) {
        return a.y - b.y;
    } else {
        return a.spriteId - b.spriteId;
    }
};

And now it is:

Tilemap.prototype._compareChildOrder = function (a, b) {
    if (a.z === b.z) {
        if (a.y === b.y) {
            return a.spriteId - b.spriteId;
        }
        return a.y - b.y;
    }
    return a.z - b.z;
};

Since the shader tilemap inherits from tilemap and Yanfly is overriding that, you might need to set shader tilemap back to a normal sort.

Of course this is all based on looking at code, not running anything yet.

I'm about to go through with this jump. Is your build the most stable and complete compared to unpropros? are they different in any significant way? if I am, in fact, using GFD's will UPP's plugin have the same issue? THANK YOU in advance!