Tracker issue for success and failure reports with the new launch GUI
cpw opened this issue · 130 comments
Please attach gists as comments to this issue. Success reports or failure reports. Please be CLEAR if it was a success (the game loaded to the main screen and looked like normal) or failure (what it looked like, how/if it crashed).
_Please indicate if you have optifine or fastcraft installed_
_GISTS please, not pastebins. Please indicate success or failure as well._
There is a special crash report in the fml-client-latest.log file- look for the text "Loading screen debug info". Grab that crash report and put it in a gist.
Looks like this:
[00:13:04] [Client thread/INFO] [STDOUT/]: [cpw.mods.fml.client.SplashProgress:start:89]: ---- Minecraft Crash Report ----
// Surprise! Haha. Well, this is awkward.
Time: 21/04/15 12:13 AM
Description: Loading screen debug info
java.lang.Throwable
at cpw.mods.fml.client.SplashProgress.start(SplashProgress.java:88)
at cpw.mods.fml.client.FMLClientHandler.beginMinecraftLoading(FMLClientHandler.java:190)
at net.minecraft.client.Minecraft.func_71384_a(Minecraft.java:480)
at net.minecraft.client.Minecraft.func_99999_d(Minecraft.java:867)
at net.minecraft.client.main.Main.main(SourceFile:148)
at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
at java.lang.reflect.Method.invoke(Method.java:483)
at net.minecraft.launchwrapper.Launch.launch(Launch.java:135)
at net.minecraft.launchwrapper.Launch.main(Launch.java:28)
A detailed walkthrough of the error, its code path and all known details is as follows:
---------------------------------------------------------------------------------------
-- System Details --
Details:
Minecraft Version: 1.7.10
Operating System: Linux (amd64) version 3.16.0-4-amd64
Java Version: 1.8.0_31, Oracle Corporation
Java VM Version: Java HotSpot(TM) 64-Bit Server VM (mixed mode), Oracle Corporation
Memory: 2557808120 bytes (2439 MB) / 4080271360 bytes (3891 MB) up to 4080271360 bytes (3891 MB)
JVM Flags: 7 total; -Xmx4g -Xms4g -XX:+UnlockCommercialFeatures -XX:+FlightRecorder -Xmn2G -XX:+UseConcMarkSweepGC -XX:CMSInitiatingOccupancyFracti$
AABB Pool Size: 0 (0 bytes; 0 MB) allocated, 0 (0 bytes; 0 MB) used
IntCache: cache: 0, tcache: 0, allocated: 0, tallocated: 0
FML:
GL info: ' Vendor: 'NVIDIA Corporation' Version: '4.4.0 NVIDIA 340.76' Renderer: 'GeForce GTX 460/PCIe/SSE2'
Works as well as could be expected for me
Works as expected on Infinity.
A crash report, a suggestion, and a few questions:
- Crash report: https://gist.github.com/luacs1998/81f6996715892896fff2 but it otherwise works as advertised.
- Question 1: Towards the end, after postinit, the FML progress bar disappears on the final resource load (in my case, I have LiteLoader installed, so LiteLoader's displays instead). Is it possible to show a "Finishing up..." or something like that instead?
- Question 2: Is it possible for modpacks to replace the Forge logo in the bottom corner with custom GIF or MC animation format loading indicators instead? And does the current logo support being replaced with a resourcepack?
- Suggestion: Can FML show a "Reloading resources" or something on the progress bar when a resource load is taking place? Also, a percentage indicator would be nice...
Crash: It loads up near the end and when its about to finish it crashes.
https://gist.github.com/SJ1133/c8f15959b2f7f0b78a48
EDIT: Should Probably Mention i am using optifine and fastcraft. Will edit with result of No optifine
Halvors,luacs I would like gists, not pastebins. Include the section I highlight and any subsequent crash.
Ah, sorry. Fixed (see original comment)
@luacs1998 In answer to your questions, this is ALL a test harness at the minute - we are validating mostly that having a threaded display context is not GL-breaking, so we don't need to do the hacky interleaving that Liteloader is doing.
We will be working with @Mumfrey to standardize the API and how information is displayed, which is why the whole "API" in the code is deprecated, we don't want mods using it.
Another Failure without optifine: https://gist.github.com/SJ1133/b2ea5be12ffce62259c9
The Progress shows up and crashes when its about to finish. I do have Fastcraft Installed.
Successfully loaded with 83 mods installed, including fastcraft but no optifine.
Success! https://www.youtube.com/watch?v=9efYGbVTI98
(No optifine or fastcraft, AMD graphics card)
@amadornes can you please extract the fake "crash report" and post it as a gist, thanks!
Wrong link * facepalm * https://gist.github.com/amadornes/ddf6dc8684c89e39ad33
It was also related to this but yeah... :P
Laptop w/ just my mod (HEE)
Success - https://gist.github.com/chylex/65f2531da79185bdf199
Desktop w/ a lot of mods, including both Optifine and Fastcraft
Crash - https://gist.github.com/chylex/787116eb1631456672c0 (both debug info & crash report included)
Desktop w/ two mods (HEE + YUNoMakeGoodMap)
Success - debug info is identical with the previous one, with the exception of used memory
Works just fine!
With Fastcraft: https://gist.github.com/Shinisuryu/e34df67bab99c4f91836
Success - https://gist.github.com/Glassmaker/3a5c1e3c2158576d7bb5
No issues so far.
Oh, I forgot to include some notes in the post: the screen looks fine, although on smaller screens (or default window resolution), all of the mod names are partially outside the window area. I would either move everything to the left, or perhaps better, add another text line that would have the mod name separately under the current event.
Success: https://gist.github.com/fnuecke/b19525bf637dc6579b66
160 mods, including Fastcraft present. No issues. Even tried stressing it out a bit (resize/maximize/minimize window, lock and resume session while loading).
Success: https://gist.github.com/Guichaguri/7658767b94a10f64807f
With almost no mods.
Crash: https://gist.github.com/Guichaguri/f0ac2356e89783821b13
With Optifine 1.7.10 HD U B6
Success: https://gist.github.com/Guichaguri/73578711e1a6b30f6662
With Fastcraft 1.21
Success: https://gist.github.com/Guichaguri/57ada39d80283c42243f
With FPSPlus 1.7.10
I don't know if the success logs are the same, so I uploaded all of them.
Success without any mods : https://gist.github.com/Dayrien/1ff090406ed90a5eebec
Success with Fastcraft : https://gist.github.com/Dayrien/a8d2be2b3ecb7c85ed6d
Success with a lod of mods, Fastcraft and Liteloader : https://gist.github.com/Dayrien/e451d71e425c3511463b
Everything works fine, no flickering like I saw in a preview video.
With Liteloader installed i saw Liteloaders loading bar appearing at step 4/5 a few times for a short moment.
Sounds to me like there is some issues with Optifine, would explain why this isn't working for me as well 👍
Success: https://gist.github.com/kevin8082/18daf877a69e25d0e403
Without any mods.
Success: https://gist.github.com/kevin8082/40038d0b4d5a9e0f6c0d
With BetterFPS(new mod indev by my friend).
Success, dev environment with CodeChickenCore and NEI loaded
https://gist.github.com/VapourDrive/ae38fea6e1d3881c49b0
Two separate successes (iGPU and dGPU AMD Enduro), as I have heard there have been problems, but I'm glad it worked both times. https://gist.github.com/Alklaine/ec7edd2d5f1b6c6b1ceb
What don't people get? Just post the actual logs.,
YOU WILL ALWAYS GET the first crash report because cpw forces it to be printed as it has needed information {video card info}
Post the full log, always.
Just posting that first one and not the other doesn't tell us what is breaking, IF anything is breaking, please guys be more descriptive.
Failure - Crashes when reaching menu, while loading there is a point where the loading screen will stop functioning (stops at a mod) and the log will continue, loading up as usual. Only mod being loaded is MrCrayFishesFurnitureMod. Discovered while trying to load up modpack on latest version of forge.
Full log - https://gist.github.com/slikrick98/b4fb6390db70b11b6b9b
Debug info - https://gist.github.com/slikrick98/e19279b49536460ddf3f
Got a crash report. https://gist.github.com/Kaelten/cf987dc3c42cac7c4b83
I got this crash: https://gist.github.com/erkro1/dee637e1ef69ac738480
I have Optifine, Fastcraft and FPSPlus installed. (all latest 1.7.10 versions)
Failure - https://gist.github.com/slikrick98/60731e4f5d6d8224a9a5
Using a japanese mod (glasspipe) seems to cause a crash when the text tries to be rendered, this causes the screen to stop at whatever mod it last loaded, then crash when it tries to load to the main menu. No optifine or fastcraft, just buildcraft and glasspipes.
Success with FastCraft v1.21 - https://gist.github.com/Arrabal-/afa556867b99be2513e3
FML v7.10.130.1395
Forge 10.13.3.1395
166 mods loaded
First attempt: success (game starts without crashing, can play without immediately crashing)
Clean install of forge, 0 mods added.
https://gist.github.com/ray10k/cf60f171013e990e592e
Second attempt: success (same as before)
Forge plus fastcraft, no other mods added.
https://gist.github.com/ray10k/7eebf4eed60fb01d813c
Third attempt: fail (Crash to launcher before the main menu is reached)
Forge, fastcraft and a load of other mods, about 80 total.
https://gist.github.com/ray10k/a20b689bb4d27cddab06
Additional report, from the crash-reports directory:
https://gist.github.com/ray10k/c199fe5ea965d622a637
During this trail, by the way, the screen freezes towards the end of the "constructing mods" phase, though the process continues until crashing about a minute later.
@cpw
With that : http://www.minecraftforum.net/forums/mapping-and-modding/minecraft-mods/1272953-1-8-3-optifine-hd-b1-fps-boost-hd-textures-aa-af?comment=46824 , Optifine B7 works with Forge #1295
Mods Installed:
Minecraft Forge v10.13.3.1395
FML v7.10.130.1395
OptiFine_1.7.10_HD_U_B7
fastcraft-1.21
Tested it and this is the crash report in console:
https://gist.github.com/Xee1/b23fd21f92f01ddf919f
The game starts and works perfectly. Success.
@ray10k Any chance we can get you to work on narrowing down that list of mods? remove 1/2 see if works, and so on.. Would be nice to know which causes the error.
@LexManos Done. Turns out that Mr Crayfish' furniture mod caused the crash, I'll get the log in a little while.
edit
Test 4: Several mods, minus furniture but including fastcraft. Successful launch
https://gist.github.com/ray10k/1ad8c28a33755ee1a3e7
Thanks @ray10k, now to find out what crayfish's mod is doing to cause that. Its the first time we've heard of a mod that can crash the new loading screen..
@cpw Glad I could help. On a side note, running with ONLY the furniture mod also causes the crash. Haven't tried redownloading the latest version yet, so maybe the issue is resolved on Mr Crayfish' side. I was using v 3.4.7 for minecraft 1.7.10.
Perfect. Hopefully we can see what he's doing that triggers it
Hello
I am using the forge installer-win 10.13.3.1395 on a clean 1.7.10 minecraft installation.
The start up stops on the mojang splash with:
"Loading - Initializing Minecraft Engine"
"Rendering Setup - Render Global instance".
If I close the launcher it is stuck on:
"[Sound Library Loader/INFO]: Sound engine started".
So I tried around if changing the sound devices does anything and find out that without
any active sound devices MC starts just fine. Even a heavily modded version.
Prior to this version I used FML 10.13.2.1291 without the sound related problem.
Hmm, start successful.
No Optifine, No Fastcraft, No CoFHTwaks
Forge 1307 and LtieLoader 1710-04 game start ~ 1min.
Forge 1343 and LtieLoader 1710-04 start ~ 1min.
Forge 1395 and LtieLoader 1710-04 start ~ 12min. wat? Slow, slow... is conflict load screen? Load screen changed old version and back repeatedly. Slow process run in Intitialezed MC Engine and after. LL re-inject output info old version load screen?
Load screen debug log(1395) - https://gist.github.com/Bogdan-G/c381f55e0b95e9cb1008
game process pause and re-pause cycle(time between 02:40:46 & 02:44:03) O_o -> https://gist.github.com/Bogdan-G/09648d8c909dc29a31ba
LL strange doing his job, his every action is accompanied by pauses that Forge in 1343 do not appear at all.
full fml log(forge 1395) - https://gist.github.com/Bogdan-G/412b00ce523d6b311aff
full fml log(forge 1343) - https://gist.github.com/Bogdan-G/bcd5d359ecdb28ee1665
full fml log(forge 1307) - https://gist.github.com/Bogdan-G/7716e9c2b9b3b46d0609
The same operation but strongly different runtime.
@Bogdan-G that looks like liteloader having a really hard time. Try latest forge and disable the screen. It should work as normal then.
@Mazranache you'll need to recreate and include BOTH the fml log and the latest.log files after the crash.
Works like it should! - fastcraft installed or.. not?!?
Forge 1395 = Fine!
https://gist.github.com/KristjanKr/6e59608dc46e3fc62fd1
Forge 1399 = Crash
https://gist.github.com/KristjanKr/837787501e76351338b4
It worked with 1395 but now it crashes with 1399, something with Font renderer. (I have Optifine and Fastcraft installed)
https://gist.github.com/erkro1/19b6f2ae64bd95bae21f
@cpw thanks.
Test 1. forge 1399 and only LL,
i edit LL config:
https://gist.github.com/Bogdan-G/a8a743e47b4adc58016f , disable loadingbar and modInfoScreen,
but no run and no loading debug screen,
full fml log -https://gist.github.com/Bogdan-G/cd9e6a69eba1025403d0
Im LL delete and normal start:
- full fml log - https://gist.github.com/Bogdan-G/cf3ef1e4937685ae7969
- loading debug screen info log - https://gist.github.com/Bogdan-G/08952d7a1ddbe02d2974
Heh, LL happy!
Test 2. Forge 1395 and only LL, edit config LL - https://gist.github.com/Bogdan-G/8c0845cbdf418bd3901a
succesfully game start ~4min, full fml log - https://gist.github.com/Bogdan-G/57b173c30f0d28b8ecd1
Other result 12 min, really progress xD
Test 3. Forge 1399, Optifine, Fastcraft, CCC, DragonAPI, LL del
Loading screen debug - not found
full fml log - https://gist.github.com/Bogdan-G/5fc1ee02c5300be19a12
succesfully game start ~1.5min
Crash before it even reach the menu screen of Minecraft
@UnrealParadox that's a context free error report. Do you have optifine? Fastcraft? Other things?
Yes fastcraft is in the modpack
Here the full list of mods for it:
http://bit.ly/1Jlmhh4
@UnrealParadox there is a bug in crayfish's mod that causes the new loading screen to crash. See above you.
My local test crayfish's mod v3.4.7 and normal start and work in game. What am i doing wrong?
http://i.imgur.com/StrwzdM.jpg I http://i.imgur.com/aqWiIBt.jpg
Crash while loading Minecraft
java.lang.IllegalStateException: Splash thread
Forge build 1395
Only mod installed OptiFine_1.7.10_HD_U_B7
Update/Remove Optifine, this has been stated many times before.
No luck updating to latest Optifine.
Updating the instance to build 1395 with a lot of mods + fastcraft and removing Optifine does work.
@cpw New issue:
With:
Minecraft Forge v10.13.3.1401
FML v7.10.140.1401
OptiFine_1.7.10_HD_U_B7
And approximately 200 mods
In splash.prop :
https://gist.github.com/Xee1/c03d211cffbd56540fec
Everything works without crash, but the splash screen doesn't appear on loading.
Without optifine, same settings, splash screen works.
(In forge #1395 splash screen works with Optifine B7, see my other report above)
@xee1 That's not an issue it's intentional, there are problems with Optifine and the new loading screen so if Optifine gets detected it doesn't show loading screen (at least for now until Optifine fixes it on their end)
@erkro1 Optifine fixed this issue, read this: www.minecraftforum.net/forums/mapping-and-modding/minecraft-mods/1272953-1-8-3-optifine-hd-b1-fps-boost-hd-textures-aa-af?comment=46824
@xee1 Yeah but in 1401 it doesn't show the loading screen with Optifine (they disable it if they detect Optifine), 1401 was released yesterday so Optifine hasn't had the time to fix that.
Yeah, optifine dev didn't wait for us to push our compatibility indicator and hacks our code instead.. He uses the compatibility indicator and we'll all be happy..
Oh, I understand. So, I've reported to optifine dev ( http://www.minecraftforum.net/forums/mapping-and-modding/minecraft-mods/1272953-1-8-3-optifine-hd-b1-fps-boost-hd-textures-aa-af?comment=46907 )
@cpw so we have to wait a little bit until the splash screens returns with optifine?
Optifine has to update to set the flag that tells FML that it's compatible with the loading screen. Without asking for that flag there is no way to tell that the game is loading the old incompatible version or the new compatible version so it just opts to not show the screen to be safe and stop crashes.
b2650a0#diff-1b82bd3139ce0104a9f4fa9a1c96cff1R126
I'm getting some kind of font rendering error after the client has been started. The game starts up fine and the loading screen displays properly but the menus of the game are completely unreadable.
I didn't have any mods installed other than Forge. I installed Forge via the installer and I tested this on the official Minecraft launcher and on MultiMC. I'm running Mac OS X 10.10.3. Looking at the logs didn't show any exceptions that I noticed but I would be more than willing to pastebin them if desired.
Also I'm using Forge version: 10.13.3.1403
Back date one version at a time and tell us when it stops happening. And provide logs.
Problem seems to have come from version 10.13.3.1389. Here are the logs:
latest.log
fml-client-latest.log
Mystery white squares where text should be? All the textures look odd? This is the weird mostly macosx race condition in the gl handling code. Seems to only happen on a specific subset of macs that we've seen. There's a forge issue tracking this as well..
If you need any more information from me about the Mac I'm using just let me know.
crash with forge 1403, no mods installed (also, it seems it tries to create first the file, only after the config folder ....)
Failure: https://gist.github.com/jotac0/c3f5449f5b0294f20d11
linux 64bit, java 1.7.0_75, forge 1403, the same crash as @TheSchwa has. 1388 works fine.
[xcb] Unknown request in queue while dequeuing
[xcb] Most likely this is a multi-threaded client and XInitThreads has not been called
[xcb] Aborting, sorry about that.
java: ../../src/xcb_io.c:179: dequeue_pending_request: Assertion `!xcb_xlib_unknown_req_in_deq' failed.
Failure with Forge 10.13.3.1403
https://gist.github.com/TjoMas-A/9ccadf7caa6d7be79026
Everything starts up fine, but once the loading bars are all finished, and the game looks as though it is going to the menu, it will crash.
I'm on Mac, and have had the issue any time I have tried using this version.
@TjoMas-A yes, there seems to be a persistent issue with the loading screen on certain types of Mac OSX. Disable the screen through the config file. We think it's some sort of race with the desktop renderer, somehow, but it's impossible to reproduce without the right hardware.
Two crashes on launch, Mac OS X & Linux mint, without Optifine, with Fastcraft:
Same xcb error as @pftbest. For me, it does not crash without mods. But if I add any mod it won't start.
The xcb error line is not present in the fml log file.
Forge 1403
Ubuntu 15.04 64bit, Oracle java 1.8.0_45
Logs:
- fml-client-latest: https://gist.github.com/xypista/d4f366eca704aff31a70
- Launcher output: https://gist.github.com/xypista/dbedc7afd9cf45ed0310
No crash log, the actual minecraft window most of the time just freeze or disappear. No files created in the crash-reports folder.
@xypista, full jvm folder and full MC folder rights 777? I have seen similar to older versions forge a bug with the game window(in Linux distributions) (forge 1286, 1291, 1307).
@Bogdan-G, I have it on an NTFS volume, mounted with rw and exec options (although its fuse and I don't know if it's affected, but I do know that I can execute scripts on it + read and write it). I played this way many times before.
But the strange thing is, it works now. The only thing has changed is that I played the game on Windows (the game directory is shared). No updates, no changes regarding the mount options.
(I have to mention that I have Win 8.1 fast boot/hybrid shutdown or whatever turned off, so it's not because of that)
The jvm was installed from WebUpd8's PPA, No 777 on that one. But why is it working suddenly now?
BTW, I wrote here because of this topic on the forum (I'm not the one who started it) and @pftbest had something similar
I'm guessing the gallium driver isn't compatible. Disable and verify it works normally.. Then file an issue with the gallium guys..
i have intel mesa, not gallium
@TheSchwa is also using intel
@xypista, the behavior of the software can not be always the same when performing the same operation. Either the something has changed and this escapes the eye.
tried to run on nvidia gpu (instead of intel) on the same machine using optifine and it worked fine,
no crashes so far.
but it would be nice to get it it working on the integrated graphics.
my nvidia card is overheating badly, so i cant use it for more than half an hour.
@pftbest as you can clearly see, this is a driver related issue, NOT a forge issue. We can't fix that the driver doesn't like semi-concurrent GL access, that is _NOT_ possible. File a bug with your graphics card driver provider.. Meanwhile, disable the loading screen in splash.properties and carry on with your day.
As an addendum, on Linux, the MESA code is the core of both Intel and Gallium 3d drivers. This error is coming from MESA. It _IS_ the same problem.
a lot of people is using (and will use) this drivers, and will get crash. if you cant fix it, than at least show a warning or something, (on the web site for example)
P.S. is the rest of minecraft single threaded? i thought this is java bug for not calling XInitThreads properly.
@pftbest given it doesn't happen to, e.g. myself (running Linux) I seriously doubt that's the cause. Where exactly would we show the warning in-game? It's exceptionally unrealistic to assume that anyone reads any website. The main alternative is to simply revert the whole thing, which I think would be a massive disappointment for the significantly less than 1% who should be able to understand how to handle this.
https://gist.github.com/AbrarSyed/a966887a75d2b536e7a6
Just FML, no forge. Yes this is the dev env, specifically the FML dev env.
I am running with the MESA drivers for intel integrated graphics.
Linux of course.
Even if I disable the splash screen, I get this:
https://gist.github.com/AbrarSyed/b665bb158654a062ebce
did some more research.. its not the GFX driver apperantly, but the JVM or lwjgl..
http://theindiestone.com/forums/index.php/topic/1044-damn-you-linuxjava/
Abrar, the second error is just a derp in the splash merge I done. No idea how to fix the JVM issue. That's why we allow disabling it..
the forum thread I posted has instructions on a really ugly hack that should fix the issue. Involves compiling a separate C library and invoking it VIA JNI before doing any rendering or something crazy like that.. not sure if its an approach we might want to go with, but it apparently exists...
specifically this post: http://theindiestone.com/forums/index.php/topic/1044-damn-you-linuxjava/?p=15829
Question Abrar: this is NOT universal or am I secretly running Winderp and have been for the last 6 years and every test of everything I've ever done is just incorrect?
@AbrarSyed and other MESA users. There's two choices, either I force disable on MESA always and forever or you can be clever users and turn it off in splash.properties. Choose. Fixing it is not possible - this looks like a religious style debate with MESA saying thou must do this (XInitThreads) whereas 1: it seems it never used to need this (tested ancient laptop w/MESA from 2006 and no bug - not surprising, XCB didn't exist then) and 2: no one else on linux needs this.
@cpw you can print the message to a log file at the end of splash screen task. something like "if your log ends here, try to disable splash screen animation, known bug, etc..."
We've been discussing this, and trying to think of a solution that informs the end user/catches it.
One Idea is to have a multi-step thing.
First Load: Normal loading, right before Main menu load a screen with something along the lines of:
"Forge has detected that this is the first time running forge with the new Loading screen option. Do you wish to enable this feature? It does not work on all machines {Approximately 0.1% of users have issues}. So if you crash on next load DO NOT ask for support just try again, Forge will automatically disable the loading screen. Yes / No /Never ask again"
If yes set TEST_LOAD_SCREEN, if Never Ask again set DISABLE_LOAD_SCREEN in Preferneces: http://docs.oracle.com/javase/8/docs/technotes/guides/preferences/index.html {Doing it in there stores it in the registry which is system wide and doesn't get shipped with mod packs and the like, as this is PURELY hardware issue storing here is the perfect choice}
Second Load when Yes:
Do normal loading screen, right before showing the main menu popup a 'Loading Screen test has finished. Do you wish to keep the loading screen? Keep / Disable'
If Enable: Set LOAD_SCREEN_TEST_PASSED
If Disable: Set DISABLE_LOAD_SCREEN
Future runs:
If DISABLE_LOAD_SCREEN, run normal without loading screen.
If LOAD_SCREEN_TEST_PASSED, YAY loading screen!
This is the only thing I can come up with the warns the 0.1% that have issues what will potentially happen and what to do in that case {Simply restart minecraft} And prevents the issue of modpacks 'Forcing the loadscreen'
@AbrarSyed XInitThreads issue isn't that simple - oracle says not to use it: http://bugs.java.com/bugdatabase/view_bug.do?bug_id=6293929; also, every Xlib call in LWJGL is protected with a global lock, from what I could tell, so thread synchronization shouldn't be the issue.
@LexManos there probably also should be a way for users to change the value later, if they, for example, update their GPU - maybe a config option to retry the test.
Yes the idea for that was to have a 'Reset Loading Screen' or config in the normal config section. That'll be taken care of it's mainly just thinking and getting the test bed implemented.
Is it even possible to store such a setting in the registry? Not to mention Macs and Linux machines don't have a registry...
http://docs.oracle.com/javase/8/docs/technotes/guides/preferences/index.html Please read what I linked. The specifics of the place this is stored is up to the implementation. It is however accessible globally outside the directory. I was told on windows it's in the registry, Linux and mac it's probably in some configs folder. Point is its not in the .minecraft folder and thus not shipped with mod packs and the 'Minecraft on USB' type distributions.
Profiling action:
During loading AG2 1.0.41:
https://dl.dropboxusercontent.com/u/3828392/Screenshot%202015-05-16%2015.03.08.png
After waiting an eternity and finished loading:
https://dl.dropboxusercontent.com/u/3828392/Screenshot%202015-05-16%2015.04.13.png
And for completes sake:
With enabled=false:
https://dl.dropboxusercontent.com/u/3828392/Screenshot%202015-05-16%2015.19.54.png
@PhoenixtheII thank you very much. Can you gist your environment info as everyone else has done. There's clearly a problem with the native buffer swap on your machine (as we suspected), there's probably little we can do to fix this in Forge I suspect, but having knowledge of what environments have the problem would be helpful..
OK. I think I know what's going on here. The machines with the loading times, have slow graphics cards. There is a "forced" sync point between the graphics thread and the loading thread, because the loading thread is calling "processMessages" on a regular basis, and both lock on the LWJGL global mutex. I'll think of a way to uncouple these two things. Tricky, because processMessages needs to be in the main thread or windows can complain about an unresponsive screen, I think.
@Phoenixthell can you confirm that your in-game FPS is less than say 30 fps regularly?
I wouldnt call a GTX 690 slow :)
Specs:
Core i7 970
24 GB DDR3 1600mhz
GTX 690
Average FPS is 60, (Nvidia adaptive vsync 'on')
Running windowed mode as you can see too :)
I'm also on IRC..., Phoenix_the_II @ esper
Can you try turning adaptive vsync off?
UPDATE: those who have slow loading, do you have adaptive vsync enabled for minecraft somehow? If so, try turning it off - it should see you resume normal speeds. Thanks @PhoenixtheII for verifying. As an addendum, this would be any kind of GPU based vsync - forced or adaptive..