3.0 Alpha Testing

[retroben]

Yeah I get the same thing,but I can access the video settings.
My no frameskip cheat won't enable,and worst of all,when I select restart game,it auto-resumes on me instead.

Edit: Whenever realtime cheat access becomes available,I thought of a great idea on how to fix some codes' response time by utilizing savestate buffering to refresh the RDRAM like when loading a savestate normally,but instead hastily saving one and loading it back immediately to refresh said RDRAM and trigger the code as enabled.
I could always add two codes for enabling and disabling one address as such for my B-T "Fly Anywhere" code which lacks proper activator functionality on Mupen64Plus AE's recompiler due to crippling refresh lag.

Interpreter allows the code to work properly,if there is a way to port whatever instruction that makes it work to the recompiler,could someone please do it when they get the chance?

[retroben]

BAD news,Banjo-Tooie keeps exiting randomly when I try to use the split-up pads at HailFire Peaks' entrance the instant I press the "A" button. That really sucks.
The app itself is not crashing.
I found the settings menu,and could turn on my normally okay "no frameskip" code,but that can't be the cause of it.

Having no way to avoid auto-resume,I can't restart like I want to for normal testing.

Edit:Cheats are even non-functional from the settings page,I still get the jagged 30fps instead of the enhanced 40-60fps from the cheat.
Other split-up pads are working fine,it may just be the status of the area I get auto-resumed in because the Icy Side's pads worked.

Edit2: Menu navigation layering is bugged when navigating with hardware controls,as it highlights games when the sidebar is active if you hit right.

[retroben]

Conker looks fantastic with Glide now!
It even has the speech bubble's shadow!
I wonder if Banjo-Tooie is also working any better.
I deleted all of my auto-saves so I can at least TRY to test some more stuff.

Edit: My mistake for the video settings I can't change a game's selected plugin profile.
The global menu worked,but a game's menu never lets you access anything directly despite the old menu stuff being in the sidebars "settings" button which is the only one that changes menus.

Damn it! I can't test Banjo-Tooie with Glide because I have Rice set to it with no way of changing to test it.

Edit2: Just tried Banjo-Kazooie with Glide and not looking so good.
Random bits and pieces are missing from Banjo and friends during the intro and the honey barns look even worse with the whole body portion going missing at some angles for the one at the first note door area in Gruntilda's Lair.
I had to use 1st person view to look at it directly because of the forced camera angle.

[retroben]

Anybody gonna post? (disregarding littleguy and some others due to GlideN64)

I tried the Audio SLES build and it sounds fantastic! (yes,I changed the plugin in global settings to it)
It has the new UI,but without the sidebar options,so I was able to change settings and enable cheats.

Ran Conker again like I did with the other one,but it even had one issue.
His legs are blue on the low resolution model. (easier to notice when hovering)
I also tried Banjo-Tooie,which had corrupted bits of text. (mostly the letter "I")

Finally,the worst offender,Super Smash Bros.which has many missing polygons,and now some misplaced textures.
Just look at Fox's tail and the Fan item colors,also Link's sword was brown in 1P Mode.
I will try to upload an image or two,if I really need to.

[BonzaiThePenguin]

BAD news,Banjo-Tooie keeps exiting randomly when I try to use the split-up pads at HailFire Peaks' entrance the instant I press the "A" button. That really sucks.

Random crashes in Banjo-Tooie is unfortunately a known problem. There's almost certainly a bug in the dynarec somewhere but the debugging tools available seemed too limited for a newbie like me to bother giving it a try. Temporarily switching to one of the interpreters to get past the crashing part almost always worked fine for me.

[BonzaiThePenguin]

It's beautiful. Only issues I encountered that changing settings from the sidebar doesn't appear to be functional yet (cheats, video plugin etc). And I encountered a crash when exiting a game after the first launch, but I couldn't get a useful log, and it's not happening again.

Yeah, the sidebar is only partially implemented at the moment, especially the cheats system and controller settings. IIRC you can still access those features through the Settings option, which takes you to the old settings panel.

[retroben]

Those menus no longer affect anything in your build.

Gillou's audio-sles build uses the new UI without the extra sidebar parts,and works flawlessly.
I kinda like it better like this.

That Banjo-Tooie issue was caused by the broken game start logic forcing it to auto resume,even if I selected restart.
Something is deeply going wrong with your build internally,sorry.

The Glide64 issue is in both builds where textures are wrong in Smash Bros. and others like Banjo-Tooie,yet Conker and DK64 look perfect.
What catastrophe caused those other games to get so bad in these new builds?

[littleguy]

I am in the process of refactoring some front-end stuff so that it integrates nicer with BonzaiThePenguin's new UI.  When that is finished we can rebase the game-sidebar branch onto master and those regressions should disappear.  Fingers crossed, my schedule might be opening up just a bit to allow me to get back into mental sync with the new UI.

There is a known texture regression in one of the upstream modules (I forget which).  It can be seen in Kirby.  That might be related to the texture issues you're seeing retroben.  You can probably dig through the github issues or the google mailing list for the discussion (I forget where I saw it).

[xperia64]

In terms of the texture I noticed it also seems to occur with banjo kazooie's font. Some letters are either the wrong letter or a completely out of place texture. A squished letter 'K' is the most common

[retroben]

For me with Banjo-Tooie,the letter 'I' is affected most.
So is upstream is responsible for all of the freaky business with Glide64 or a large chunk related to GLES issues?

Would this also explain the broken Rice screen drawing for Smash Bros. too?
Using another type of the Rice plugin (Mudlord 6.1.4) on PJ64 results in Smash Bros. with a clean view.
But that's a different emulator.

Still find it strange that my x86 tablet's Intel has identical graphic issues to the Qualcomm on FireTV,they look exactly the same despite different architectures and processors.
Maybe the GLES implementation is somehow identical regardless of the architecture or processor,excluding Qualcomm's bad ES 3.0 flaws.

[littleguy]

Correct, graphical bugs in most cases won't depend on cpu architecture.  Most differences created by arm vs x86 will be emulation/core related stuff (i.e. dynarec).

We are using completely standard upstream plugins for
 - mupen64plus-core
 - mupen64plus-rsp-hle
 - mupen64plus-audio-sdl
 - mupen64plus-video-glide64mk2
 - mupen64plus-video-rice
 - mupen64plus-ui-console

So any issues related to those are best directed to the upstream google groups forum.

The plugins that android edition is responsible for are
 - mupen64plus-video-gln64
 - mupen64plus-input-android
 - mupen64plus-audio-sles
 - SDL2

So any regressions related to those are best directed here on the paulscode forum.

[Mikhail]

Probably already mentioned but just incase.
http://filthypants.blogspot.co.uk/2014/12/n64-3-point-texture-filtering-in.html

[littleguy]

@Mikhail - You're better off posting on the upstream list, since we are using standard upstream plugins now, and there are many more devs there who are in a position to weigh in.  Gillou may be the only active downstream developer with the skills for it, and he follows the upstream discussions too.
https://groups.google.com/forum/#!forum/mupen64plus

[gdark100]

On the lastest Alpha, everytime I try to re-scan ROMs at the root of my sd card, the emulator crashes. If I try to scan inside my ROMs directory, it works. Below is the log inside the "CrashLogs" folder.

Spoiler: show

Code: [Select]

Uncaught exception in package org.mupen64plusae.v3.alpha

****MESSAGE****
An error occured while executing doInBackground()

****STACK TRACE****
java.lang.RuntimeException: An error occured while executing doInBackground()
at android.os.AsyncTask$3.done(AsyncTask.java:299)
at java.util.concurrent.FutureTask$Sync.innerSetException(FutureTask.java:273)
at java.util.concurrent.FutureTask.setException(FutureTask.java:124)
at java.util.concurrent.FutureTask$Sync.innerRun(FutureTask.java:307)
at java.util.concurrent.FutureTask.run(FutureTask.java:137)
at android.os.AsyncTask$SerialExecutor$1.run(AsyncTask.java:230)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1076)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:569)
at java.lang.Thread.run(Thread.java:856)
Caused by: java.lang.NullPointerException
at paulscode.android.mupen64plusae.task.CacheRomInfoTask.getAllFiles(CacheRomInfoTask.java:215)
at paulscode.android.mupen64plusae.task.CacheRomInfoTask.getAllFiles(CacheRomInfoTask.java:218)
at paulscode.android.mupen64plusae.task.CacheRomInfoTask.doInBackground(CacheRomInfoTask.java:118)
at paulscode.android.mupen64plusae.task.CacheRomInfoTask.doInBackground(CacheRomInfoTask.java:52)
at android.os.AsyncTask$2.call(AsyncTask.java:287)
at java.util.concurrent.FutureTask$Sync.innerRun(FutureTask.java:305)
... 5 more

****LOGCAT****


****DEVICE****
Processor : ARMv7 Processor rev 1 (v7l)
processor : 0
BogoMIPS : 4.80

processor : 1
BogoMIPS : 4.80

Features : swp half thumb fastmult vfp edsp thumbee neon vfpv3 tls
CPU implementer : 0x41
CPU architecture: 7
CPU variant : 0x2
CPU part : 0xc09
CPU revision : 1

Hardware : SAMSUNG JANICE
Revision : 000e
Serial : XXXX

Board: montblanc
Brand: samsung
Device: GT-I9070
Display: JZO54K.I9070VJLPD
Host: SEP-119
ID: JZO54K
Manufacturer: samsung
Model: GT-I9070
Product: GT-I9070


****PERIPHERALS****

Device: Virtual
Id: -1
Descriptor: a718a782d34bc767f4689c232d64d527998ea7fd
Class: BUTTON

Device: sec_jack
Id: 1
Descriptor: bda7e7374aba7c80dee4f7f0d239c94d5609e45a
Class: BUTTON

Device: gpio-keys
Id: 7
Descriptor: 485d69228e24f5e46da1598745890b214130dbc4
Class: BUTTON

Device: sec_touchkey
Id: 8
Descriptor: af4d26ea4cdc857cc0f1ed1ed51996db77be1e4d
Class: BUTTON

Device: AB8500 POn(PowerOn) Key
Id: 11
Descriptor: 6ecc9411d15f2f29cc10f7afdeaaacc1773c3bfc
Class: BUTTON


[littleguy]

I just noticed that last night too... thanks for the info.